I will tell you from the off that it’s a big improvement. Adobe have had their machines a’learning and we are grateful. It’s web only and not likely to be in Premiere Pro that soon. As it is now, we have a cutdown version of the web V1 in Premiere – still useful – but the benefit of taking the time to use the website has just gone up a notch. (The website is here: https://podcast.adobe.com/enhance by the way).

In Jason Levine’s announcement video last week he was keen to point out that this new model is more natural and less “podcast”. Personally I’m a bit on the fence with that – I agree it’s way better, but I still find sometimes it puts too much bass back in, though some of that can be fixed with a little EQ. Plus – and I guess he didn’t want to put down V1 by mentioning how often it sounds robotic – but V2 is way way better in this regard.

Pricing – it’s still free, but you can pay for extras

Adobe Enhance Speech is still free which is pretty amazing. However there are some limitations.

You have to do things one at a time, you have to export the audio from your video first to upload an audio file & the big one – no percentage slider! It’s 100% or nothing. Plus the length and size limitations you can see here. Personally I do have Premium Access, which I presume comes with my Adobe All Apps plan, but you can subscribe for it standalone for $10 per month if you need.

I think the lack of slider would be the biggest drawback, but you can blend it with your original for the same effect. There is a risk that this could cause phasing issues, but when I tested it I didn’t hear any.

Drawbacks

These are some of the possible negatives I’ve seen people write about:

• Doesn’t work that well in non-English languages
• Even putting the slider at 1% affects things quite a lot
• Can put too much bass back (I sometimes have had to EQ the bass down a bit, which doesn’t take long)

Do let me know if you see any of these.

A personal test

I thought it would be fun before I get to my main battery of tests to try it on a personal project – restoring an old VHS of my mum on Call my Bluff from the 1970s. I have to say it did a fantastic job. And I could combine this with something like Topaz Video AI to uprez the video.

Testing

I have updated the four tests I ran back in May for my review of iZotope RX 11 and I’ll give you my thoughts. Let me know if you see (hear) it differently!

Firstly you can hear the difference between V1 and V2 on all the tests. And V2 is impressive throughout.

As usual I am running the tools on higher settings than you usually would, to try to distinguish between them. The fact is they’re all quite impressive so I have to push them quite hard to hear the difference.

Test 1

On this first test – some common background noise – I still give the win to iZotope RX, as I think it sounds more natural (though bear in mind it comes in a $400 package).

Winner: iZotope RX 11 Dialogue Isolate.

Test 2

With this test of a poor phone recording, Adobe was already the winner from last time. While in this instance V2 does add in more bass, that can be fixed and the overall quality is better. But context matters – it will depend a bit on the visuals that go with your audio as to how you want it to sound.

Winner: Adobe Enhance Speech V2.

Test 3

V1 had struggled quite a bit with this tough test of a recording on a busy street, even missing some of the words. V2 fares a lot better. I still prefer the natural sound of both RX and Waves Clarity Vx though (note that I used both Vx and Vx DeReverb)

Winners: iZotope RX 11 Dialogue Isolate & Waves Clarity Vx.

Test 4

In this fourth test, I ran all the plugins on a lower setting to keep the voice less affected. Previously only the specialist Spectral Repair tool in RX (not seen here) was really able to get rid of the annoying camera focus noises though Enhance V2 could get rid of them at 100% and I will give it some credit for that, but the ideal is to keep some of the original atmos.

Enhance V2 is again clearly much better than V1 here, able to remove a lot more noise without affecting the voice (albeit it does sound a bit podcasty still). A quick highpass filter does improve it quite a bit, though the tone is still a bit unnatural. I do think RX is closer to the original, but Enhance Speech V2 sounds a bit more like I had a good boom pointing at me. So I’ll give them the draw.

Winners: iZotope RX 11 Dialogue Isolate or Enhance Speech V2 depending on context.

Conclusion

Adobe’s Enhance Speech is a big step forward and a fantastic option for editors and post audio professionals to have. It’s still the case that you want options – sometimes I’ll run through all the tools I have and one happens to work the best. Overall I would keep iZotope RX and its Dialogue Isolate tool if I could keep just one, but happily I don’t have to! Plus I’ve seen Enhance Speech V2 win the day on quite a few occasions.

Give it a try for yourself and let me know how you get on.

We’ve already covered what’s new in RX 11, like the new Streaming Preview or the Mid/Side mode. Here I’m mainly looking at the Dialogue Isolate module in RX, which now happily comes in RX Standard and now runs as a plug-in.

Pricing

In terms of pricing, Clear, dxRevive and the combo of both Clarity Vx plugins are all at the $100 mark. There are also the pro versions – $300 for dxRevive Pro and $400 for both pro Waves plugins. Adobe’s Enhance Speech continues to be free online or included in Premiere Pro.

iZotope RX 11 Standard is $400 (currently $300 on sale) and RX 11 Advanced is $1200 (currently $800). RX Standard has 33 modules in total and Advanced has 44, so you’re unlikely to buy it for noise reduction alone, but I have found as an editor I often reach for other modules like De-clip, Ambience Match or Spectral Repair.

Testing methodology

I’m testing RX 11’s Dialogue Isolate against the leading competitors as well as against RX10. If like me you are trying to decide whether to upgrade this year, I hope this will help. I’ve taken out RX Voice Denoise from my testing as it doesn’t really compete (but is still great for minor ambience reduction). I’ve also taken out Davinci Resolve’s Voice Isolation as it’s really something you’ll only use if you’re in Resolve & it’s not worth going there specifically for.

As usual I am pushing the noise reduction quite far to try to hear the differences.

For Adobe’s Enhance Speech I ran it in Premiere Pro as I think this is the most used workflow. An Adobe rep recently said that the model used online can afford to be bigger and so produces better results, so that’s an option for your trickiest audio.

With RX 10 I used both Dialogue DeReverb and Dialogue Isolate modules together (which are now combined in RX11) and I did the same with Waves Clarity Vx’s two plugins – too much reverb is just as much an issue as background noise after all. For dxRevive I tested with the newer 1.1 release which has given us a new improved model called “Studio 2” (plus an algorithm “Natural” featured only in the Pro version which is $300).

I’ll note that with RX 11 Advanced the Dialogue Isolate module includes multiband processing (similar to the Pro versions of Clarity Vx and dxRevive), but I haven’t used that here.

On with the testing – Test 1

For me, RX 11’s Dialogue Isolate does a wonderful job in this first test – it is my favourite of all the plug-ins tested in terms of keeping the original tone of the voice – and it does significantly better than RX 10. It’s a good example of why you don’t always want the enhancement of Enhance Speech or dxRevive. As you’ll see below, it wasn’t always as clear cut as this, but still I am pretty impressed.

Test 2

Here I am testing what is now a common occurrence – someone sending you some voice over recorded on their phone.

My winner is Adobe’s Enhance Speech – I feel it has taken out the noise and improved the EQ on the voice to produce the best sounding output. But after that, RX11 and Clear come in joint second and again RX11 is better than RX10 though it’s not as noticeable. Of course you would also want to EQ out some of the bass rumble.

Test 3

This is the hardest test I am throwing at the tools on display – with both heavy street traffic and lots of reverb present.

This time I only hear a very little improvement in RX 11 over RX 10 and both do a pretty good job. Supertone Clear struggled on this one, where Waves Clarity Vx was much stronger & the best of the non-enhancers for me. Enhance Speech and dxRevive were both were spoiled by artifacts. So I will give Waves the win here.

Test 4

In this fourth test there are annoying camera focus noises as well as some gentle ambience. This time I didn’t try to push the plug-ins so hard, but used lower settings on each tool to reduce the background noise slightly to try to not affect the voice at all. I noted that the camera noise was limited to a very small frequency range around 500Hz, so I also tested the Spectral Repair tool which is included in RX Standard.

As expected, going after the specific frequency brought the best results & was the only one that got rid of the camera noise completely and so iZotope RX takes the win here (whether version 10 or 11). After that, all of them did a pretty good job, except that I have to dock some points from Enhance Speech for turning the camera noise into Daft Punk style words at the end! (I did test this clip on the online version of Enhance Speech and it didn’t do this there).

Conclusion

I think it’s fair to say that in the game of noise reduction, iZotope have caught up with leading players like Supertone Clear and Waves Clarity Vx with the release of RX 11. I wouldn’t say they’ve surpassed them, but I think in my tests RX was perhaps a bit more reliable – achieving an overall lower level of artifacts.

It’s still the case that ideally you have multiple tools, because they each tend to do better in different circumstances. And the voice enhancement tools remain a bit hit and miss, but can be incredibly powerful when they work.

So do you buy? For an editor who just wants a good noise reduction plugin to use in their NLE, the much lower price will mean that Clear or Clarity Vx will be an easier choice. For audio post professionals who most likely already have RX Advanced, I think the upgrade this year is well worth it, alongside the other new or improved tools. For someone who is thinking of bringing RX into their workflow for the first time, using its multiple modules to clean up their audio – I think this is a pretty good year to do it.

But first, why do threads asking about recommendations for Windows laptops have literally half the replies saying “you should just buy a Macbook!” Frankly, because considering all things, they are better. A typical high end Windows laptop working hard will have about 2-3 hours battery life (or just be forced to curb half its power when you unplug), where the Macbook Pro can last all day, cool and quiet at full power. The design is beautiful, the build quality, the screen, the trackpad – everything really – is top notch. And you don’t have to spend endless hours looking through models to choose from – you have the peace of mind to know it “just works”.

But here are a few key points as to why you might want to buy a Windows laptop:

• You don’t care about battery as you will always be plugged in
• You want compatibility with your desktop PC
• You are familiar with the Windows 10 or 11 OS and don’t want to learn Mac OS
• You are not invested in the Apple ecosystem with your other devices
• You do work that runs much better on an Nvidia graphics card
• You find the software you use daily is more stable on Windows
• You want to game on this machine in your downtime
• You prefer OLED screens to mini-LED
• You only need a budget device (Windows/Chromebook) which can be found for much cheaper than the entry level 13″ MacBook Air.

Whatever your reasons (let me know in the comments!) – it’s pretty annoying when people don’t trust that you’ve thought it through. Well I do and here is my guide to help you choose your next shiny Windows laptop.

Where to start

Personally, I get pretty stressed doing buying research if I start filling my head up with info and don’t write it down. I like to get a spreadsheet going –  maybe you have another method, but keeping organised notes is key to help you narrow down your options. I like to have different models lined up in rows and the specs I care about in columns and to start to mark things that stand out as good or bad for my particular needs.

There are some great YouTube channels that review laptops which helps you get a feel for the consensus as to what most reviewers like and for you to see the different styles on offer. To be fair, a lot of them are geared to gamers, but there’s a lot of good info out there. There are also a number of good reviewing websites, my favourite being notebookcheck.com.

Deciding factors – what matters to you

I’ll go into some of these in more detail below, but this is a pretty comprehensive list of what you need decide:

• Budget – everything starts with your budget and that will immediately narrow down the field into lower, mid or high end tiers.
• Brand – are there brands you like the look of or trust more
• CPU – do you need Intel or can you save a bit of money with AMD (or, gasp, risk the new Apple-esque Snapdragon)
• GPU – do you need Nvidia (probably!) or can you save a bit of money with AMD or integrated graphics (which are much better these days)
• RAM – what’s your minimum? A lot of laptops only come with 16GB RAM, likely not enough for pro users
• Hard drive – pretty much all will come with an NVMe SSD (good news), but what size is enough for you
• Size and weight – how often are you going to carry this thing around and what screen size do you need
• Screen type – do you prefer OLED (deeper blacks, vibrant colours) or mini-LED (brighter, more efficient) or IPS (good viewing angles, cheaper)
• Workstation – are you willing to pay extra for the more robust build quality, pro range Nvidia GPU and upgradability that comes in a workstation range
• Ports – USB-C, USB-A, SD card reader, ethernet, etc – how many of each do you need
• Battery life – if working on the move, how important is this to you
• Other factors – design, noise and heat,  build quality, webcam quality, keyboard, screen flickering, speakers, etc.

CPU generations

After deciding on your budget, the CPU is likely the next main decision you’ll have to make. Unfortunately we’re at a slightly confusing time with Intel being in the process of releasing a new architecture, but I’m going to break it down for you. Until this year, Intel has been going each year to a new generation – 12th gen was called Alder lake, with a high end laptop having say a Core i9-12900HX. 13th gen was Raptor Lake, with a Core i9-13900HX. 14th gen was a Raptor Lake Refresh, so a pretty small tweak with a i9-14900HX.

But now Intel is in the process of releasing Meteor Lake – what you can think of as 15th gen, but as it’s a fairly different design they are starting the count again and calling it 1st generation, with a new naming convention. The i has gone, so no more i9 and the names are shorter. The upper-middle tier Core Ultra 7 155H chip is the one that features in a number of laptops out currently, with the more powerful 165H and 185H still to come (and if history repeats, the HX series will then come with a slight increase in performance).

The confusing thing is that this new 1st generation of Intel chips brings pro’s and con’s. The main pro is it features a brand new “AI” chip – the NPU (Neural Processing Unit) – designed to take the strain off the CPU and GPU for localised AI-based tasks. There’s plenty to read on this (eg here), but the bottom line is that you definitely don’t need this today – for starters most AI work is done in the cloud and then when you do need to do it locally, your GPU will do just fine. That said, it is a move towards longer battery life that many will welcome and over the next few years, more AI local tasks will be present.

The main con is that the chips sometimes benchmark worse than the previous gen chips. So now you have to decide if you really need this fancy NPU – I would say right now it’s not worth it, but others disagree. It’s going to take a year or so to really see where it’s going – for example there are reports that some of the current batch of “AI PC’s” won’t support Windows Copilot+ & we’ll have to see how much that feature will add to users.

And what about AMD? They also have an NPU in recent chips like the high end Ryzen 9 7940HS, though it is not supported as well yet. You tend to save a bit of money with a Ryzen, without much loss of productivity, unless you are dealing with video, which most of us are, in which case the QuickSync tech built into Intel’s chips is just not worth giving up, particularly for Adobe Premiere Pro users.

GPU tech

It’s a similar story for the GPU – for most of us, it’s just better to stick with Nvidia. The NLE’s work better on it as do things like 3D or CAD work. The better question is how much do you need to spend on the card – do you need a 4080 over a 4060 say – and do you need the professional “Quadro” range (they’re not actually called Quadro any more!) like the RTX 2000 Ada Generation. For basic editing, you don’t see much difference when getting a higher end card. A 4060 mobile chip with 8GB VRAM will be fine for most people, though if you have a decent budget and what to save a bit of time when rendering effects or exports or have a better time in Davinci Resolve which loves GPUs, then it may be worth it to you to get a laptop with a better card, just know the heat and noise and electricity bills all go up.

People who need the pro range tend to know who they are – 3d modellers, data scientists, CAD people and the like. That said Avid has traditionally wanted this range to fully support your machine, so that’s worth checking (though personally I’ve always run Avid fine without one). The actual cards are pretty much the same hardware as their GeForce equivalents & for the laptop range they don’t even have more VRAM – but the drivers are more robust and some apps do need them.

One other interesting thing about the newest Intel chips is their integrated graphics are more powerful than before (with the tech from their GPU range brought over) and that will mean that some on a tight budget who are just editing 1080p with proxies will likely be fine without a discreet GPU at all.

Workstations

In the arena of laptops, workstation just means they are marketing it to business and professional users who are willing to pay more for the sake of more robust build quality, more security features (like Intel vPro), longer support cycles, usually a “Quadro” card, more RAM, more port options & often the ability to upgrade the RAM/SSD/battery. Traditionally they were the top performers, but that isn’t so true these days with gamers willing to spend as much as they are. If you are moving about quite a lot and depending on this machine for your livelihood, it may well be worth it you – though a high end gamer or creator laptop will also be worth the comparison, especially as they tend to be more up to date on the latest hardware.

The big 3 laptop manufactures are all pretty good – HP ZBooks, the Dell Precision range & Lenovo ThinkPads. But there are also workstations from the likes of MSI, Gigabyte & Asus. Puget have also recently released their own very high end workstation.

High End creator or gaming laptops

Assuming you don’t need every feature of a workstation, then there are some very high performing machines tailored to gamers and creators – probably beating the workstations for performance per dollar. Mostly what gamers need is what creators need, so it partly comes down to the looks of the thing – do you want flashy LEDs all over it or plain metal.

In this arena, HP, Dell and Lenovo are there, alongside Asus, Acer, MSI, Razer, as well as Samsung & Microsoft. The video from Ben Kaiser I put above is probably the best overview I have found, but there’s also great individual reviews which you can watch for your shortlist.

What is my choice?

If you’ve read this far, I will give you my current plan and my thinking behind it – I will be buying soon, so let me know if you think I’m wrong! These choices will be different for each person, so it’s impossible to really say what machine will be best for you, but it might help to know my process.

Firstly, I have a high budget because I want a machine that I can rely on for professional work & I’m willing to pay for a workstation for two main features – spare SSD slots and build quality, though bear in mind some non-workstations still tick these boxes. I also have a very particular desire for an IPS screen as my eyes react badly to the invisible flickering inherent to OLED and mini-LEDs displays. I wanted 32GB RAM & 1TB SSD minimum and at least 1600 horizontal lines of resolution (so >2.5K) in a 16″ or 17″ screen. I didn’t need a new CPU and didn’t even mind 13th gen as the uptick to 14th was minimal.

Currently the HP ZBook Fury is my leading contender, simply because you can slot in an additional 3 NVMe drives into it! That would be a game changer for me as I have multiple NVMe drives in my desktop PC and can just take them out and slot them into the laptop when I need to travel, saving a lot of data-transfer time. The downside is the thing is an absolute brick – extremely thick and heavy! But I just need to get it on a plane to my destination, then I’ll be in an office, I won’t be carrying it every day.

Of course your priorities are likely very different from mine which is why I’ve mainly focused on the right questions to ask and not tried to give too many answers. Godspeed fellow Windows laptop buyer!

There are many quotes on how important sound is in films. Many of us, myself included, would go further than Lynch and say that audio is almost always more important than visuals in the experience of the viewer. Show someone a documentary where they can hear the dialogue clearly with blurry or shaky visuals. Then show one with pristine shots and terrible audio and see which one has the most impact.

This speaks to the importance of the audio team – from getting the right sound on the shoot, through to the mixing process in post. But it’s also a big part of the offline editor’s job – to sell the cut with good temp sound design and a good temp mix – this can make all the difference in getting things signed off. And of course at the lower budget end, many editors won’t have the money to hand over their audio to a dedicated sound editor. These are a couple of examples where the standalone app End Boost might prove useful.

What is End Boost?

The basic idea is an automatic one button audio mix. Perhaps that sounds too good to be true, but frankly it works pretty well. That’s not to say that it does everything.

There’s no automatic EQ (yet) and the noise reduction (which is in Beta) is not quite up to the standards of the leading plugins. It’s also not going to do cleanup for you – removing clicks and pops and rustle.

What it will do however is provide a custom AI mixing algorithm get the right levels of gain and compression on your dialogue with the right amount of ducking on your music and sound effects/atmos, to deliver a pretty professional sounding mix. These are things you can do yourself in your NLE of course if you know how – the point is they take time.

Née Alex Audio Butler

née      /neɪ/      adjective
“originally called; born”

End Boost is a continuation of the plugin Alex Audio Butler from Dutch developer Jerre van der Hulst at his company Unimule. Alex Audio Butler was a set of VST plugins that you’d apply to dialogue, sound effects and music in the Track Mixer and it would automatically mix the levels for you. I tried it out for a while and it was pretty good, though it had its downsides. Mainly that levels would jump around a bit when you pressed play as it did its calculations – not very conducive for editing. If you exported without a full play through, you had to do it twice – once for it to put on its thinking cap and the second to apply that to the mix. Still, the results were pretty good.

Standalone app

Nevertheless, perhaps taking it out of the NLE and into a standalone app was a good idea. It’s a cleaner and more stable approach (NLEs can be fussy about using third party VSTs) – and means it can work for anyone. It’s also better than using an online tool (think mastering sites like Landr for music) as you don’t have to send your files to anyone. And don’t worry, the little robot butler is still present.

Jerre puts it like this on his website:

“The goal was always to add more automatic audio features and presets. Think: de-noising, EQing, creative filters, and who knows what else will be possible… But with every update of NLEs and operating systems the number of “bandaids” needed to keep the “magic” working grew. Sadly it became clear that almost all programming effort went to keeping it running, with no time for adding new stuff.”

The downside is that any time you have to go outside your NLE it adds time to the equation. Thankfully though it is a very quick process, especially once you get used to what settings you want (and they are pretty good on the defaults anyway). The analysis step is much quicker than Alex Audio Butler was – no playthroughs or exports required.

Ease of Use

The first step is to export your audio tracks – separating dialogue (which would include interviews, voiceover, foreground actuality), sound effects (which includes background atmos and any foley or sound design) and music. This can be done simply by solo’ing your tracks and exporting or you can have dedicated submix tracks which can save time in bigger projects. You can either do this in three stems or just add multiple tracks (as long as your pricing tier allows it).

Then open up the End Boost app to bring them in.

And as the manual says “If a file name starts with [v], [m] or [s] End Boost will automatically set the track Source Type to Voice, Music or Sound FX respectively.” Otherwise you can set that manually once you’ve brought them.

After import, End Boost will have your stems – make sure it has correctly identified which each is. It will very quickly run its analysis based on the default settings or the ones you last used & you can hit “Play Preview”. The settings may look a bit complicated, but it’s actually fairly simple & it basically works straight away, you can just tweak it if needed. You can also bring in a video to mix to, which is a good idea as picture affects sound levels (eg if you can see the sea, it feels right to hear it).

These are the main things you need to set (and they all have tooltips over on the right):

Voice

• Set Constant Source if it’s just one voice or Varying Source(s) if multiple
• Set the denoising and compression amounts (the defaults are usually fine)

Music

• Set the target volume (the default IIII is about right, so just lower or raise it if you want to after listening)
• Set the aggressiveness of the auto-ducking (e.g. medium)

Sound FX

• Set the target volume (I find I normally turn down the default a bit, to say II)
• Set the aggressiveness of the auto-ducking (Subtle I think is a good bet)

Mix

• Mastering – basically setting how strong a limiter you want
• Delivery target
  • All-Round is about -17 LUFS which I use for internet delivery
  • EBU R128 is -23 LUFS (Europe’s broadcast spec)
  • Online is about -12 LUFS if you are playing the social media loudness wars.

You’ll notice there’s no ATSC A/85 -24 LUFS/LKFS export for the US market and I wonder if Jerre will add that in a future update. Though I don’t expect people using this tool are doing it for broadcast so I’m not sure it matter too much.

Once you are happy you hit export and there is a very handy drag and drop dialog for you to drag straight back into your NLE to align with your video.

Tests

Now down to some testing. First a simple test of some dialogue and music. Note the audio ducking in the End Boost mix and how the voice comes to the forefront of the mix.

And secondly with some background atmos (I put a little inset video to match it, otherwise it didn’t sound right!)

You can see (hear!) that it works really well, especially with noise reduction done first. The dipping is very well timed, the compression is strong without artefacts and the output levels are consistent.

Who is this for?

As I alluded to at the start, I don’t think this app is trying to compete with the world of Audio Post Production. But the busy editor sending a temp mix during the approvals process might well find it useful. And even more the lower budget content creator who does their mixing in house. Especially if they didn’t know how to achieve these results in their NLE.

I would always encourage budding editors to grow in their audio skills and doing a good temp mix is part of that, so part of me doesn’t want to recommend End Boost for that reason. But on the other hand I think for quite a few people it gives a very good end result that they wouldn’t easily be able achieve themselves. For them I say take advantage of the 7 day trial and give it a spin.

But recently I saw a post on the Blue Collar Post Collective Facebook forum by Sound Editor Andrew Macht of Minneapolis based Macht Sound sharing his extremely detailed 22 page pdf document on how to turnover from Adobe Premiere Pro to Avid Pro Tools.

I should point out right at the start that most sound studios will have their own turnover document that you should request and follow to the letter as not everyone has exactly the same requirements. For example some will prefer the older OMF format over AAF, some might allow trimmed clips with handles while others will require full clips, etc.

Bottom line: don’t try to be clever, ask for the turnover document and follow it

Still, it’s amazing to have this kind of document available and it is typical of what mixers ask for. So I’m not going to go through a full turnover guide in this article, nor am I getting into the various metadata requirements for different post workflows beyond some basics. Aside from Andrew’s document I’ll also point you to this page from Michael Kammes for some more tips, as well as some links at the bottom.

Here I am investigating one particular part of the puzzle – exactly what gain and volume information gets through from Premiere to Pro Tools and how that affects the Audio post team – assuming they are on Pro Tools which most are. Andrew and I spent a couple of hours testing as many options as we could & reached some interesting conclusions.

Gain vs Volume – they both make something louder right?

What is the difference between gain and volume?

Without wanting to do a complete history of the analog days or go on about gain staging which is written about in depth, I see three key practical differences:

• In Premiere, Clip Gain can’t be keyframed, but volume can (though you can cut clips up as much as you want)
• Gain comes before effects, volume after
• Volume can be linked to the faders on a control board

Gain is a great way to get your levels into the right ballpark & most plugins work better when you’ve done that. Volume can then used to do the fine tuning and the automation (like dipping music under dialogue). And as Andrew explained to me, if he’s working as a pure Dialogue Editor, the Mixing Engineer that comes after him will want him to work with gain alone and leave volume for them to use.

Andrew demo’d to me his faders moving automatically with the Pro Tools keyframes and vice versa, creating the keyframes by moving the faders – very cool!

Audio in Premiere Pro

Clip gain and volume automation works basically the same inside Premiere and most DAWs & NLEs for that matter.

One big difference is that volume can be removed with the “remove attributes” command, whereas gain can only be removed by using the Audio Gain command and setting gain to zero.

But my main question is what translates to Pro Tools via the AAF or OMF? And if you’re not familiar, these two file formats are the main way editors send off their audio to be mixed.

OMF

OMF, or OMFI if you’re old school, is the older format but still used by many. It has some limitations – like being limited to 2GB if you chose to embed the audio – but perhaps has more compatibility.

It supports clip volume automation but cannot send Clip Gain from Premiere to Pro Tools. Setting Clip Gain comes through as volume.

But if you use both Clip Gain and volume on a clip, it gets added together, as long as there are no keyframes. Once there are keyframes, I found that they no longer came through correctly. So if delivering in OMF, I’d advise not combining gain and volume on the same clip.

AAF

The newer AAF format can carry much more data across & doesn’t have that pesky file size limit. And it can send through both Clip Gain and Volume Automation separately.

But interestingly, Clip Gain does not always come through. This is the rule according to my tests:

Clip gain can come through, but only if there are no volume keyframes present on the clip in Premiere. As soon as there’s even a single keyframe, then the clip gain acts as a volume offset – in other words it is added to the value of the keyframe.

So for example, in this music track I have the first keyframe at +7dB and the gain is set on the clip at -10dB. The resulting keyframe that comes through on the AAF is -3db.

And a single keyframe is enough to trigger this, for example here, I have applied -30dB gain to both clips, but the music clip (on track 2) has a single keyframe added which hasn’t even been moved.

Avoid Premiere’s Track Volume Automation

It’s a bit hidden by default, but in Premiere you can also add volume keyframes to the track as opposed to the clip.

This doesn’t come through in either OMF or AAF, so should be avoided by any editor who is sending their mix off to someone else.

Dips or Cuts?

Something you do see debated a fair bit, is whether it’s better for editors to dip their music with cuts and dissolves or by use of volume keyframes. I haven’t seen a consensus on this and Andrew said for him either is fine. If you have a strong feeling on it, let me know!

Both OMF and AAF will bring through crossfades, so in theory both ways are fine.

Except that when combined with volume (AAF) or either gain or volume (OMF), the volume comes through keyframed and if you change the crossfade, it doesn’t change the volume automation – they are not linked. So that doesn’t feel ideal to me.

Metadata

I said I wouldn’t get into metadata as it’s such a big subject, but let me cover a few key points here. Metadata is very important to the audio team. It’s important that it’s entered correctly by the production sound mixer on the shoot & it’s important it gets correctly translated through to the turnover files the editor sends through. As Andrew showed me, with metadata intact, it speeds up your workflow immensely, for example switching quickly between microphones from the same take.

AAF and OMF are different here. AAF preserves track/channel names (those that are entered on set, not the track names of your edit), where OMF preserves filenames.

EDL

I’ll also mention EDLs because they are often requested too. They are very useful, both for tracking any changes (if the locked cut becomes unlocked), but also for feeding into some really great bits of software that the sound dept can use like EdiLoad or Conformalizer/Matchbox.

Conclusion

The bottom line for editors is to trust the audio department and do what they ask! We’re on the same team and there is often needless strife when editors think they know better.

Audio editors are different and some will start from scratch with mixing decisions, others will tweak the editor’s temp mix, others will keep it as a reference. Either way it’s useful for gain and volume decisions to come through where possible & hopefully this deep dive is useful for you to decide what you want to do.

Personally, I will continue to use Clip Gain to bring levels to the right ballpark and then use volume automation to dip music or do any fine tuning. But if think differently, do let me know.

Other useful links

https://helpx.adobe.com/africa/premiere-pro/using/export-aaf-files.html

https://helpx.adobe.com/africa/premiere-pro/using/exporting-omf-files-pro-tools.html

https://2pop.calarts.edu/technicalsupport/premiere-to-pro-tools/

http://leaders.terburg.com/

https://www.aatranslator.com.au/

So I’ve tested both of these new tools against most of the other main players (I could never do them all!) – in three different real world scenarios.

Here is a handy spreadsheet to keep track:

Test 1 – Background noise

I first tested with a crazy amount of background noise to put these tools through their paces.

For me, I think Accentize dxRevive is the winner here. Like Adobe Enhance Speech, it is clearly trying to do more than just remove the noise – it’s trying to emulate what the voice would have sounded like in a better recording environment. And it does quite an impressive job here – as long as that’s what you want – there are plenty of post production workflows where you might need to keep the original voice tone.

Supertone Clear / GOYO is very good at pure noise removal – on a par with the equally good Waves Clarity Vx. Davinci Resolve Studio’s relatively recent Voice Isolation function (found in the Inspector on the Edit Page) is not quite as good, but decent.

And then there’s iZotope RX, which was what I swore by for many years, but that was before the AI revolution. It is clearly not as good as the other offerings, even using the tools in the costly RX Advanced. To be fair, there are over 60 separate tools in RX Advanced and I would hate to be without them (Ambience Match is so helpful for example). Still, I am amazed that in the bread and butter of basic noise removal, iZotope is currently a long way behind the pack.

Test 2 – Phone Recording

Test 2 is also fairly typical – someone sends you a voiceover they’ve recorded on their phone.

I think probably Adobe Enhance Speech edges the win here. It really is quite an impressive job. It’s fantastic to have it present in the app now, well the Beta at least, as the workflow is so much quicker, plus it has an amount slider.

If you wanted to voice to remain as it was, Supertone Clear / GOYO takes a lead over Waves Clarity Vx here, because it is much better at removing the room reverb & in fact has that as a separate dial. That said, Waves do also have Clarity Vx DeReverb as a separate plugin & you can get both for around $80, so it’s much of a muchness.

dxRevive struggled a bit in this test & I’ve seen Adobe’s Enhance Speech sometimes do the same – it’s the downside of those two that they try for more, but don’t always make it. iZotope RX Advanced’s Dialogue De-reverb did a pretty good job here & I do think it’s a great tool and quite a bit better than the general De-reverb that you get in RX Elements. Davinci Resolve also did pretty well, removing some reverb as well as noise.

Test 3 – Clipped audio

This isn’t super common any more (and can be ruled out completely with 32bit recorders thank goodness), but clipped audio sometimes comes across my path.

To be fair, a lot of these tools aren’t claiming to be able to fix clipped audio, but Clarity Vx and Supertone Clear (GOYO) don’t do anything for this, nor does Davinci Resolve do that much. But the voice rebuilders – dxRevive & Enhance Speech – do a wonderful job of removing the clipping. iZotope RX De-clip (in RX Elements) also does a good job and I’ve used this many a time to rescue audio.

Summary

When it comes down to it, workflow is king. You have to be able to get the right tool – or more likely tools, as they have different strengths and weaknesses – to work well for you. Happily most VST/AU/AAX plugins work in most apps these days (though always check) and can sit on a dialogue track or a submix track to avoid any round trip at all.

Plugins

• Supertone Clear (GOYO) – has basically the same excellent noise removal as Waves Clarity Vx, but with much better reverb removal. Out soon for $99 ($29 for Beta testers).
• Accentize dxRevive – very impressive noise removal and voice enhancement, similar to Adobe Enhance Speech, but in a VST/AU/AAX that you can use in your favourite DAW/NLE. $99 (or $299 for the Pro version). They do also have a pure noise removal tool – VoiceGate – that you might want to check out, plus DeRoom for reverb removal.
• Waves Clarity Vx – beats Supertone Clear on price, if you don’t need the reverb removal. Normally on perma-sale for $39 (also a Pro version available normally for $249) & Clarity Vx DeReverb (normally $39.99 or $149 for Pro)
• iZotope RX Elements – very good for light noise removal work, but can’t compete past that. Includes 6 plugins which are useful (particularly De-click). $99.

Apps

• Davinci Resolve Studio – useful Voice Isolation tool, but not as strong as the plugins above.
• Adobe Premiere Pro subscribers ($21 per month) can download and install the Beta alongside the regular version to take advantage of the excellent Enhance Speech or anyone can use it on the website for free (though it’s quite slow for longer files).

There are three new cards from Nvidia and unlike the higher end cards, there are plenty of third party cards from the likes of Asus and Gigabyte at MRSP:

• RTX 4060 $299 / £289 (July)
• RTX 4060 Ti (8GB) $399 / £389 (out now)
• RTX 4060 Ti (16GB) $499 / £479 (July)

Let me give you a TL:DR summary of the RTX 4060 Ti (8GB) version I am reviewing here. This is a great upgrade for you if:

• You have an older card like a 1060 or 960 or older AMD card
• You don’t need more than 8GB VRAM
• You don’t need the extra speed/power of a more expensive card
• You need AV1 encoding

A week of editing

Swapping the 4060 Ti in for my 4080, the first thing I noticed was that the 4060 Ti is much smaller. This is quite welcome, as I can barely shut the case with the 4080 in it. Plus the power usage/cost is much lower too and it means you don’t need as powerful a PSU.

You can read all about the specs if you want, like the increased L2 cache or the reduced memory bus (compared to the 3060 Ti), but the only thing I’m interested in is how it performs in day to day editing tasks.

I used the 4060 Ti for a week of editing in Adobe Premiere Pro. I noticed no difference whatsoever in downgrading from the much more expensive 4080. This is to be expected – these cards are much more powerful than NLEs need in the day to day. In the last few years I’ve used the 2060 Super, the 4060 Ti and the 4080 – all of them gave me the same experience in day to day editing.

The RTX 4060 Ti is more than powerful enough for most editors. As you can see below, unless doing more intensive tasks, it has most of the performance of the much more expensive 4080. It’s a great card, though if your budget is smaller, you could also consider waiting for the 4060 at $100 less or getting a 20 or 30 series card – I see that in Puget’s tests the 3060 Ti outperformed the 4060 Ti in some tests. I don’t recommend going back further than the 20 series because they brought a significant increase in H.264 encoding quality & since then I’ve never bothered with software compression. And I do recommend sticking with Nvidia, they just work that bit better than AMD for editing tasks.

Test 1: Decoding H.264

These tests are slightly different from my last review and the software versions and drivers have advanced, so I tested all three cards again. This test is decoding H.264 (8bit 4:2:0) at 1080p to DNx (could also be ProRes) – which you would do when transcoding or making proxies at the start of a project.

• Specs: Windows 10 Pro, AMD Threadripper 3960X (24 core), 128GB RAM (CL-16 3200MHz), Multiple NVMe SSDs. Avid Media Composer v2023.03, Adobe Premiere Pro v23.4, Davinci Resolve v18.1.2, Topaz Video AI v3.2.7.
• I did the test at 1080p this time as opposed to UHD in my previous article, which is why all the scores are higher, though the patterns are the same.
• Avid continues to rock every thread of my Threadripper, whilst not using the GPU’s decoder at all, which shows that an Avid transcoding machine wants a nice big HEDT processor.
• Premiere continues to not access Nvidia’s hardware decoder (NVDEC) for export (it does use it for playback at least) – this has been true for many months across different versions and drivers. If you have an Intel CPU with Quick Sync you should be fine as Premiere can decode with that.
• There are decent gains to be had in the 40 series over the 20 series, though not much between the 4060 Ti and the 4080.
• Davinci Resolve is able to make the most use of hardware & in a UHD test pulls ahead of Avid by a greater margin.

Test 2: Encoding H.265

This is the opposite test – encoding H.265 from a mezzanine codec. For example you have exported a ProRes master and need to send it for review. This test is at UHD and makes use of what I learned about presets to get the most speed.

• Avid is terrible at encoding – it doesn’t use the GPU, nor use the CPU well. Best to export same as source and then use Shutter Encoder or Resolve.
• Premiere is better, but isn’t making the most of the hardware so gets no gain from the better cards.
• Resolve is able to access the dual encoders in the 4070 Ti, 4080 & 4090 for remarkable speeds, you just have to be using H.265 or AV1, the “Faster” or “Very Fast” preset & be at UHD or above.

For interest I did an H.265 export test for Resolve at 1080p which doesn’t access the dual encoders of the 4080. It still beats the 4060 Ti, but by a lot less. The 2060 Super falls further behind though.

Test 3: Decoding and Encoding H.264 to H.264

This test is for when you have H.264 media on your timeline and export to the same for review. Again I did this one at 1080p so all the figures are higher than last time, though Resolve pulls ahead remarkablty at this resolution.

• Avid’s bad encode is more powerful than its good decode here
• Premiere falls behind Resolve by a much greater margin at 1080p compared to my previous UHD test. I don’t know if it’s Nvidia or Adobe who need to sort this out, but I hope they do. Again, if using an Intel chip with Quick Sync, you’ll probably be getting much better speeds than this.
• Aside from Resolve being impressive, when doing both encoding and decoding, the more expensive cards aren’t really bringing that much. Though at UHD and above the difference is greater.

Test 4: Topaz Video AI

Strangely, with the latest version of Topaz, the card didn’t make any difference at all, though it was much faster across the board since my last test. Realistically this will probably change again with future versions and I expect if you do use Topaz a lot, a 4080 or 4090 will be great to have.

Test 5: Intensive or AI tasks in Davinci Resolve

This is the kind of work where it really pays to have a 4080 or 4090 – things like noise reduction in Resolve or even more an AI based tool like the remarkable Magic Mask.

Other considerations: 8K and VRAM

A good rule of thumb for VRAM is: Resolution +2 is safe, Resolution x2 is better. So for 1080p/2K you are fine with 4GB VRAM. With UHD/4K, 6GB VRAM is usually fine and 8GB is better. If you do use 6K or 8K you might want to wait for and pay the extra $100 for the 4060 Ti (16GB) version that’s coming out in July.

Conclusion

Graphics cards get more and more powerful each year and while the gaming community seem to complain for each and every release, I think for the post production professional, you get a lot of value out of cards such as the 4060 Ti.

However, if you’re a colourist or do graphically intensive tasks or are a Davinci Resolve power user, you can get quite a bit more out of the more expensive cards and probably want more than the 8GB VRAM on offer here.

Even though the trend these days is towards full frame, all the manufacturers at this bracket have gone with Super 35 sized sensors aside from Sony. It’s worth remembering that over the years films were mainly shot on Super 35 (or digital sensors of that size) – they just used fast and/or long lenses to get the signature look (the former reduces depth of field and the latter increases compression). It’s worth doing your research though in your market/field as to what directors and producers are asking for – both in terms of whether they care about full frame or whether they require a level above this – you might find in some documentary work for example that you need a Sony FX9 ($11,000) or Canon EOS C500 Mark II ($11,000) or possibly the C300 Mark III ($9000)- in which case it might be better to rent for a while until you can afford one.

My top tier – Sony, Blackmagic, Canon & RED

But assuming you are ready to spend $6k on a camera, here is a spreadsheet of the main contenders:

Notes:

• Pricing – is rounded to the nearest 5 to be more clear and avoid marketing
• Dynamic range & Rolling shutter – taken from the excellent lab tests at Cine D – I’m using the more realistic 2:1 signal to noise ratio figure (so it will be lower than what is quoted by the manufacturer).
• IBIS – In body image stabilization – useful for handheld shooting and vintage glass, but often only used in lower end smaller cameras
• SDI – BNC connection for SDI output (a better connector than HDMI)
• TC – Timecode – the ability to jam sync timecode from a device like a Tentacle Sync
• ND – Built in ND filters

What are the key considerations?

• Lenses – making sure your lens collection is compatible or that you are going to be investing in the right lenses. Most older lenses can be adapted to the modern mounts (e.g. Canon EF to Sony E or Canon RF), but generally EF lenses only have good continuous autofocus on either the EF or RF mount if that’s something you want to use. Newer lenses from the mirrorless ranges can’t be adapted – for example if you own Canon RF lenses you would only be looking at the C70 or the Komodo here and for people with Sony E mount lenses, they’d stick to the FX6.
• Full frame or Super 35? This is a case of pros and cons – full frame has better low light, potentially shallower depth of field (depending on the lens), but requires more expensive, larger lenses. Super 35 gives you a crop factor which you need to allow for with photography lenses. There is always the option of speedboosters for using full frame lenses on a Super 35 sensor.
• 6K or 4K? If you’re delivering 4K/UHD, it is a bonus to shoot more than that and have the leeway to crop in or stabilize, but it may not be a deal breaker for you.
• Cards/media – factor this into your budget and go through the process of what media you need for your desired codec and resolution/frame rate, I haven’t gone into detail on it here.
• Extra features? There are many other features that may or may not be important to you – I will cover key ones below and generally ignore things that all cameras have like timecode in.

In no particular order…

1. Sony FX6

I think overall Sony continues to have the most buzz in the video world and that was shown by how hard is has been to get hold of the Sony FX6 ($6000) during most of the time since launch (combined with chip shortages globally), but that seems to have settled down now. As I mentioned above, if you want full frame at this budget, this is your camera. The FX6 has the same sensor as the FX3 and A7SIII, but most agree this camera gets a bit more quality out it. It has excellent autofocus, a clever built in ND that can be automated & it outshines the others in low light capability with the full frame sensor and high usable ISOs. The increased size over the FX3 brings a more robust cooling system and a plethora of quick access buttons and dials. It can give gyro data for post stabilization which is a nice touch. Like the Canon C70, it is extremely light. The E mount brings the greater choice of lenses compared to Canon’s RF – in part due to allowing the third parties to use it (like say the Sigma range).

The FX6 does have a lower measured dynamic range than the other three cameras here. It’s also lower than the FX3/A7SIII, but that’s more to do with less baked in noise reduction which actually produces a more pleasing image. However, I don’t think that excuse can realistically be used against its competition. It is also the most limited in terms of internal codecs – the only one without internal RAW recording, should you want that. Realistically though, most aren’t going to need it. As is often the case, there were some limitations at launch that have since been updated in firmware to bring improvements.

2. Blackmagic URSA Mini Pro 4.6K G2

The URSA Mini Pro 4.6K G2 ($6000) – as we’ve come to expect from Blackmagic – packs a lot into this price point. For starters there’s the 4.6K – that extra 0.6 might not sound much, but that’s all you need to keep full resolution when stabilization or cropping in and delivering 4K. I’d like to see Sony & Canon considering it if they don’t want to go to 6K. Being able to shoot 4.6K at up to 120 fps (or 2K300) in BRAW- surely the easiest to use RAW codec – is excellent as is the ability to shoot straight to SSD. It also has a SDI in as well as out, making it great in a studio set up and Blackmagic have really great integration with all their hardware for multicamera setups – you can even have your live cut magically appear in the Resolve timeline. (You could also look at the cheaper URSA G2 Broadcast Camera 6K ($4200) for a studio set up or the other studio cameras). The URSA Mini Pro 4.6K G2 comes with an EF mount as standard (with optional PL, B4 and F lens mounts available separately), so it can be a great ENG camera too.

A possible downside is that it’s much heavier than the others, around the same weight as an Arri ALEXA Mini. Some DPs will mind that, some won’t, but it’s worth considering if you’re shooting all day, as is the BRAW or ProRes only recording or lack of continuous autofocus. And although the Mini Pro 4.6K was updated to the G2 version in 2021, it doesn’t quite get the same attention as it used to.

Stepping just outside this budget, you might also look at the URSA Mini Pro 12K ($6385) if you really need that extra resolution, but the dynamic range is less & there are reduced codec options (BRAW only).

3. Canon EOS C70

I’ve brought a fair bit of criticism to Canon at the lower price points, but things change at this price point and up. All the cameras cheaper than the C70 are hybrids and are hampered by that. The Cinema EOS team, with the freedom of making video only cameras, start to trade blows with their main rivals Sony and a case in point is the Canon EOS C70 ($5500). Its dual gain output sensor (the same as in the C300 Mark III) has the highest measured dynamic range and the footage looks great. It has good autofocus, but perhaps a little behind Sony. It has plenty of buttons, a good cooling system and a digital stabilization system designed to work in conjunction with a lens’ optical IS. It’s also a bit cheaper than the others.

Not everyone will appreciate the somewhat DSLR-esque looks of the C70 at this price, though once rigged out that may not be as noticeable and it’s an advantage for gimbal use that it’s small. It has the newer RF mount, with some great lenses, but less choice. EF lenses also adapt well to RF (in terms of continuous autofocus). It lacks SDI out (so monitors use HDMI only) and only has mini XLRs (requiring adaptors to full size). The rolling shutter is a bit worse than the FX6, but still very good. It launched with slightly limited internal recording options, but more have been added since then with firmware updates.

A big question – with the firesale of the C200 mark I – is when will the C200 mark II come out and what will they price it at? Maybe at NAB?

4. RED Komodo 6K

The RED Komodo 6K ($6000) is a different offering to the others, being a box camera and will cost a bit more to rig out, but I thought I’d include it here. It will generally appeal to DPs who rate the RED look above the other contenders or who want the 6K resolution. It’s amazing that it has a global shutter at this price point – giving you peace of mind that you’ll never see any jello-type artefacts no matter how you move the camera. RED also have a great phone app to control the Komodo 6K which could come in handy.

There are more limitations with the Komodo – it’s RAW or ProRes only and the high frame rates are less than the others. XLRs can be added, but only with add on adaptors. There’s no useable continuous autofocus and there are no built in NDs, in keeping with this being normally used with a matte box and separate NDs. It’s designed to be used in controlled conditions (narrative, studio, etc.), but that’s not to say doc DPs can’t use it, they might just struggle with obs-doc work unless they’re pretty experienced.

Best of the rest

Just outside this price point is the Panasonic EVA1 ($6500) which does have an impressive 5.7K of resolution. There are also some RED Komodo alternatives, like the Kinefinity MAVO mark 2 LF (from $6000) or the Z CAM E2-F8 ($6000) which are both full frame. You could also look for a second hand FS7 or C300.

I am again very impressed at the features that these manufacturers pack inside these camera bodies for the price. Long let the competition between them flourish!

Probably the most obvious difference here is that these cameras are all full frame, whereas most in the lower bracket top tier are Super 35. There are full frame cameras in the sub $2500 bracket, but for me they compromise in too many areas. At this price point you get full frame without those compromises. The hybrids here are very much video first and designed to be workhorses, without the overheating issues that can limit you on longer shoots.

My top tier – Sony and Panasonic

This time I have only three top tier cameras – the Sony A7SIII ($3500), the Sony FX3 ($3900) and the Panasonic Lumix S1H ($4000, often $3700). I think these cameras are set apart by their excellent dynamic range, image quality, video features and reliability. Added to this is the fact that both the Sony E mount and the Panasonic/Leica L mount both support more affordable third party lenses from the likes of Sigma, Tamron, Rokinon, Samyang, etc. – something Canon no longer offer.

Again a handy spreadsheet of some of the key data:

Notes:

• Pricing – is rounded to the nearest 5 to be more clear and avoid marketing
• Resolution – approximate – I am not worrying about C4K vs UHD here but it may be important to you – see manufacturer’s sites for full specs – usually the top spec involves a crop or limited functionality, plus they all can do more externally
• Dynamic range & Rolling shutter – taken from the excellent lab tests at Cine D or Gerald Undone – I’m using the more realistic 2:1 signal to noise ratio figure (so it will be lower than what is quoted by the manufacturer). Note that about 2 years ago Gerald Undone adjusted his method to align with Gunther’s at CineD, so I’m only using those recent figures
• IBIS – In body image stabilization – useful for handheld shooting and vintage glass, but sometimes taken out of higher end cameras
• TC – Timecode – the ability to jam sync timecode from a device like a Tentacle Sync (if you don’t have this, you can still record LTC onto an audio track which Davinci Resolve can read)
• ND – Built in ND filters
• Year – year of release

What are the key considerations?

• Do you need a hybrid? My rules is: if you need a hybrid, buy a hybrid, if you don’t, just get the best camera for your needs whether it is one or not – but bearing in mind that a hybrid can overheat as it usually doesn’t have active cooling. That said, at this price point the hybrids from Sony and Panasonic do much better.
• Will your older lenses fit? Most older lenses can be adapted to most mirrorless cameras, but you should check out the available adaptors from the likes of Metabones, Sigma or from each camera manufacturer.
• Will your newer lenses fit? Newer lenses from the mirrorless ranges can’t be adapted – for example Canon RF, Sony E, Nikon Z, Fujifilm X or M4/3 – so choose carefully before you buy into an ecosystem.
• Continuous autofocus for older lenses? If this is something you use, it doesn’t tend to work across brands via adaptors, for example Canon EF lenses on a Sony.
• Full frame or Super 35? This is a case of pros and cons – full frame means shallower depth of field which many want, but more expensive, larger lenses and less use of vintage glass. That said most full frames can be put into crop mode when you need it. Super 35 / APS-C gives you a crop factor which you need to allow for, but it might suit you if shooting far away subjects.
• 6K or 4K? If you’re delivering 4K/UHD, it is a bonus to shoot more than that and have the leeway to crop in or stabilize, but it may not be a deal breaker for you.
• Cards – I haven’t covered that particularly as I don’t think it swings things that much, unless you happen to have a good collection of say CFast or CFexpress, then you might want a camera that takes what you have. Though you can just shoot on a good SD card on any of them if you’re happy with H.264.
• Extra features? There are many other features that may or may not be important to you – I will cover key ones below

1. Sony A7SIII

The first two options – the Sony A7SIII ($3500) and the Sony FX3 ($3900) – are essentially the same camera internally, with the same sensor and processor and perform the same in most key metrics. So first I’ll cover what is common to both. Both have excellent dynamic range, one of the lowest rolling shutter times of any small camera & the full frame aesthetic. They have excellent low light / high ISO performance due to the lower megapixel dual gain sensor. They have the fantastic and very customizable autofocus, great IBIS, 10-bit 4:2:2 internal recording, full size HDMI out & USB power delivery. They shoot 4K120 with a slight crop and full frame from 4K60. Both are fantastic cameras.

So really the key difference is that the A7SIII is a hybrid – for sure it’s video-first (thus the letter S as opposed to R for the photo-first cameras), but it can still take great photos as long as you don’t need a very high megapixel count. If you want a hybrid that is photo first, but still has great video capabilities, the Sony A7R V ($3900) is a great option with its 61MP sensor and it even has some things that are better than the A7SIII for video, due to it being a newer camera. It has a brilliant tilt & flip screen, 8K recording if you needed it & some smarter autofocus options. There are some compromises on the video front, though, with various crop modes in play, worse rolling shutter, dynamic range that’s a bit lower & worse low light performance.

2. Sony FX3

The Sony FX3 ($3900) comes in at $400 more, but comes with the XLR top handle which is well designed. If you find it a bit short, it can be extended nicely with the Kondor Blue extension kit ($45). While you can get a $600 XLR adaptor for the A7SIII, it’s not as secure and there’s no handle.

The FX3 is part of Sony’s cinema line like its younger and much cheaper Super 35 sibling the FX30 ($1800). Aside from that being partly marketing (as I mentioned, it’s the same internally as the A7SIII), that means quite a few useful video additions. It has the active cooling fan (which you can have blow faster when not shooting and slow down to keep noise levels lower while filming). However the A7SIII solved pretty much all of the overheating issues that dogged the A7SII, so this isn’t as important as it could have been, but still this does make it more robust. While it does away with the EVF, it adds a bunch of useful 1/4″ threads, better buttons and tally lights as well as better firmware. That means you have access to things like Cine EI mode, user LUTS, ability to jam sync timecode and a better video menu which is very welcome. These extra video features helped it get Netflix approved, pretty impressive for a camera of this size.

3. Panasonic Lumix S1H

The Panasonic Lumix S1H ($4000) is a very similar camera to the A7SIII in many ways. It is a video first hybrid (thus the H), though it has a 24MP sensor rather than 12MP, which may suit some hybrid users better. However that means the low light performance isn’t quite as good. It has pretty similar specs to the cameras above so I’ll dive into some of the differences. It has even higher dynamic range and beautiful highlight roll off. Colour-wise, for me it’s slightly harder to grade, but that’s quite subjective. The rolling shutter is not as good, the IBIS is good, but can be warpy and there are less lenses available. It has better cooling with the active cooling fan & a more useful & higher resolution tilt/flip screen. Like the FX3, but unlike the A7SIII, it’s Netflix approved – and has the advantage of recording C4K internally, with the Sonys limited to UHD. With an external recorder like the Atomos Ninja V it can record 10 bit 4:2:2 C4K60 or RAW 12-bit 5.9K/30, with BRAW (my favourite of the RAW codecs) having been added in a firmware update.

The S1H has the advantage of shooting 6K (well 5.9K) 3:2, which is great for anamorphics or when shooting for social, so that when you do your square or vertical versions you can make the framing less tight really easily. I’ve done this with the Arri classic’s 4:3 shooting mode and it works really well.

One of the main downsides is the autofocus – it lacks phase detect and it is significantly worse at tracking than the Sony or Canon offerings, but this may not be something you plan on using. Another big issue is its age – it was released in 2019 and so will not hold its second hand value as well. And we’ve seen that the IBIS and AF were vastly improved in the S5 II over the S5, so I’d expect the same in the S1H II which I’d expect by the end of the year. Still, it remains an excellent camera.

Best of the rest

Canon EOS R5

The Canon EOS R5 ($3900, $3500 during March) is the main camera that really should be in the list above. It can produce some stunning video for sure and could be a great option for those with RF lenses. There are two major issues with it – poor dynamic range and overheating. CineD measured the dynamic range significantly lower than its competition above. It was released with a fanfare about its 8K, but people soon released it just couldn’t be used for any serious video work due to overheating and slow recovery. That said, in July 2022 with firmware 1.6, Canon managed to mostly solve this, adding in a setting that Sony cameras have to allow the camera to build up more heat, but still, how hot do you want your camera or cards to get.

Both of these issues were fixed with the Canon EOS R5C ($4800, $4400 during March), with its active cooling fan and greater dynamic range, making it a much better choice for video shooters, but at $1300 more than the A7SIII that is a hard sell. Still for RF lens owners and even EF owners (with the video AF compatibility) it is a great camera.

But the problem is, at that price, you start looking at the Canon EOS C70 ($5500) which I personally think shoots much nicer video, though it is Super 35. But by then you’re pretty close to the next price bracket which includes the Sony FX6 ($6000), RED Komodo 6K ($6000) and Blackmagic URSA Mini Pro 4.6K ($6000) and more. So I find Canon’s prices too high to recommend them. The thing is, Canon are selling their cameras like hotcakes, including the video focused ones – even issuing an apology for their lack of stock, so there clearly is demand. I would guess though that this is primarily from the photography crowd who are increasingly being asked to provide some video in their work.

Blackmagic Pocket 6K Pro

The Blackmagic Pocket 6K Pro ($2535) for me still competes here even with its lower price and will rival the cameras above if you are happy with the EF mount, lack of autofocus and other factors which you can read in my write up here.

Canon C200

I have to mention the Canon EOS C200 ($3500) due to its hefty price drop (I assume to clear final stock as it’s close to end of line) – it was originally $6500 when it came out in 2017. It’s a pretty nice camera and could impress your clients to look at, but there is a distinct lack of good codec options – only the tricky to use Cinema RAW Light (fills cards fast / hard on your computer) or 8 bit 4:2:0 and I think that should be left behind in the 2010’s (hey Facebook, why is there banding in my sky?)

Nikon Z7 II

Nikon again are not really competing on the video front at the moment, but that doesn’t mean you can’t get great results out of the Nikon Z7 II ($3000), especially for those with Nikon Z mount lenses. It’s a full frame hybrid with a 46MP sensor, IBIS and records up to 4K60 with a very small crop, but is severely lacking in only recording 4:2:0 8 bit internally and with the nasty 30 minute limit. It actually has poorer video quality than its younger sibling the Z6 II ($2000), so it is more geared to photographers who need that extra megapixel count. Like the Z6 II, it can record 4:2:2 10 bit externally, so it’s worth factoring in an external recorder into the price if you’re serious about using it for filming.

Any other options?

There are some good box cameras in this price range, from Kinefinity, Z CAM, Panasonic & Bosma, but the real price once kitted out fully will be much higher which is why I haven’t included them as proper contenders.

This sub $4000 price bracket is slightly more confusing due to both the A7SIII and the S1H being a bit old and because the lower price bracket of sub $2500 is so strong. Many will be ok with the Sony A7 IV ($2500) or the Panasonic S1 ($2500) for example if they need full frame. But the cameras at this higher price point are workhorses designed for the professional video shooter and that will suit many who need the smaller form factor or who don’t have the budget for the next step up.

The answer to which is the best video camera is really that there isn’t one. There are a number of questions you have to work through which will cross some of the options off your list and after that it will depend on what you value and your particular workflow. For example, personally I value dynamic range and highlight roll off a lot. I also hate video that’s too sharp if I can sense that it’s digital sharpness rather than true resolution. But your values may be different depending on your audience and what platform you are delivering for. So this is personal to some extent and if you think differently, do let me know in the comments!

My top tier – Blackmagic, 2 x Sony and Fujifilm

I have chosen four cameras that I think edge ahead of the pack: the Blackmagic Pocket 6K Pro ($2535 – close enough!), the Sony A7 IV ($2500), the Sony FX30 ($1800) & the Fujifilm X-H2S ($2500) – in no particular order. After that I cover their main competition.

[Update April 2023: The Sony ZV-E1 ($2200) would now be an option – it’s got the full frame sensor from the FX3 / A7S III, but is very limited in record times before it overheats & in some other ways, so you’d have to be pretty sure you were never going to need it for a proper shoot & for that reason I can’t recommend it.]

First up a handy spreadsheet of some of the key data:

Notes:

• Pricing – is rounded to the nearest 5 to be more clear and avoid marketing
• Sensor size – Super 35 and APS-C are close enough to treat them as equivalent when searching for cameras
• Resolution – approximate – I am not bothered about C4K vs UHD etc but you may be – see manufacturer’s sites for full specs – usually the top spec involves a crop or limited functionality
• Dynamic range – taken from the excellent lab tests at Cine D, here I’m using the more realistic 2:1 signal to noise ratio figure (so it will be lower than what is quoted by the manufacturer)
• IBIS – In body image stabilization – useful for handheld shooting and vintage glass
• TC – Timecode – the ability to jam sync timecode from a device like a Tentacle Sync
• ND – Built in ND filters

What are the key considerations?

• Do you need a hybrid? My rules is: if you need a hybrid, buy a hybrid, if you don’t, just get the best camera for your needs whether it is one or not – but bearing in mind that a hybrid can overheat as it usually doesn’t have active cooling.
• Will your older lenses fit? Most older lenses can be adapted to most mirrorless cameras, but you should check out the available adaptors from the likes of Metabones, Sigma or from each camera manufacturer.
• Will your newer lenses fit? Newer lenses from the mirrorless ranges can’t be adapted – for example Canon RF, Sony E, Nikon Z, Fujifilm X or M4/3 – so choose carefully before you buy into an ecosystem.
• Continuous autofocus for older lenses? If this is something you use, it doesn’t tend to work across brands via adaptors, for example Canon EF lenses on a Sony.
• Full frame or Super 35? This is a case of pros and cons – full frame means shallower depth of field which many want, but more expensive, larger lenses and less use of vintage glass. That said most full frames can be put into crop mode when you need it. Super 35 / APS-C gives you a crop factor which you need to allow for, but it might suit you if shooting far away subjects.
• 6K or 4K? If you’re delivering 4K/UHD, it is a bonus to shoot more than that and have the leeway to crop in or stabilize, but it may not be a deal breaker for you.
• Cards – I haven’t covered that particularly as I don’t think it swings things that much, unless you happen to have a good collection of say CFast or CFexpress, then you might want a camera that takes what you have. Though you can just shoot on a good SD card on any of them if you’re happy with H.264.
• Extra features? There are many other features that may or may not be important to you – I will cover key ones below

1. Blackmagic Design Pocket 6K Pro

First off, let’s look at the major outlier in the group, the Blackmagic Design Pocket 6K Pro ($2535). The no. 1 thing to consider with the Pocket 6K Pro is that you are locked into using lenses that have the Canon EF mount (or the PL mount if you don’t mind modding the camera). The thing is, a lot of people have a collection of EF glass and don’t want to buy into a new lens line-up, so for those people it can be a great option.

What else sets it apart? It is the only camera with built in ND filters (a massive bonus to not be fiddling around with a vari-ND for every lens change and impressive at this price range), the only one with built in XLRs (albeit mini-XLR which are a bit of a faff) and the only one with a decent sized LCD screen at 5 inches rather than the standard 3 inch of all the others. It’s also a lot bigger and almost twice as heavy as the others – this has pros and cons – it can look more professional to some clients to have a bigger camera & handle better, but requires a more heavy duty gimbal if shooting on one and is heavier to carry for all day shoots. It’s the only one that shoots RAW internally and BRAW is a great format – it is 12 bit log equivalent to 16 bit linear. You can also record to an external SSD which is useful and it has perhaps the most simple and user friendly menu system.

On the negative side, it is the only camera here without a useable continuous autofocus, without IBIS and without a fully articulating screen. That will rule it out for a whole group of videographers and makes it better suited to prosumers or pro DPs who are used to manual focus and using their own bodies for stabilization. The battery life is probably the worst here, though it has the most versatile power input – as well as USB-C power delivery (which all the cameras now have), you can rig up Sony NP or even V-Lock batteries to power it via the Weipu connector.

2. Sony A7 IV

If you shoot video and stills and need a hybrid, the Sony A7 IV ($2500) is a great option at this price point. But even for the video-only shooter there are some reasons why you might choose it over the FX30 which comes next. First up it’s full frame. It has incredible dynamic range and low-light capabilities – even beating the much more expensive A7S III on DR tests. Both the A7 IV and the FX30 have the industry leading autofocus and IBIS – this can make life much easier for run and gun shooting or vlogging. Both have some neat features like the focus map and lens breathing correction and gyro data.

On the other hand, not everyone loves the complicated Sony menu system which can take a bit of time to research and get used to. The rolling shutter is not great on this camera unless you’re in crop mode. It shoots 4K60, but only in Super 35 crop mode, with the full sensor starting at 4K30. Probably the biggest thing against it would be the possibility of overheating if you are shooting continously or in warmer climates.

3. Sony FX30

The Sony FX30 ($1800) is the odd one out price-wise – at a solid $700 cheaper and given that I think it competes here aside from price, it is easily the best value contender. It has the classic Sony industry leading AF and IBIS, the same colour profiles like SLOG3 or S-Cinetone, similar inputs and outputs, gyro data, etc. and the same fairly poor rolling shutter. Dynamic range is said to be good, but I haven’t seen enough proper tests yet and so far I think it’s lower than the A7 IV.

Aside from the smaller sensor, the big thing that sets it apart from the A7 IV is it being a video-first camera, part of Sony’s Cinema Line (which includes the FX3, the FX6 and FX9 and even the Sony Venice), with very limited photo shooting capabilities and no EVF. The body is much better suited for video rigs, with many 1/4″ mounts and better button layout. A big thing is that is has a built in cooling fan designed to overcome the overheating blues that hybrids can suffer from – and in the menu you can set it to run minimally while shooting and then spin up when you’re not. And it’s even got a thumb zoom lever which works with some lenses like the Sony 18-105mm f/4 which is bundled in some kits.

For an extra $400, still within our budget, you can buy it with the XLR top handle, which provides pro level audio and is handy for handling the camera (the A7 IV does also have an XLR top handle for an extra $600). The FX30 adds some pro video features which the A7 IV doesn’t have, like being able to jam sync timecode or import custom LUTs and Cine EI mode which some will find useful though it’s worth looking into it before using it. It can also record 16 bit RAW externally.

The top spec of 4K120 is only with a large crop but still can be useful, but it shoots 4K60 without a crop. It can shoot up to 240fps at 1080p which is better than the A7 IV. It currently has better firmware too, though the A7 IV is rumoured to be getting the upgrade very soon. There’s a lot going for both Sony options (here’s a good video to help you decide) and which is better will depend on your specific needs.

4. Fujifilm X-H2S

Over the last few years, Fujifilm have developed a fair amount of admirers, especially with the Fujifilm X-H2S ($2500).

The sensor has a new stacked X-Trans engine which some would argue has the best image quality of all the cameras featured here with its 14 bit sensor readout. A stacked sensor means a quicker camera with fantastic rolling shutter performance. In camera ProRes recording is very welcome and can save a lot of time in post as the majority will either transcode or make proxies from H.264/H.265 files, though it will fill up cards quicker. The autofocus and IBIS are only ok and not as good as Sony’s. But being able to shoot 6K 3:2 is great for shooting for simultaneous vertical and landscape delivery or for anamorphic lenses. It can also shoot 4K120, with a slight crop and 4K60 with none.

Though Fujifilm don’t sell an XLR top mount, Tascam do for $500. Also, although it is a hybrid, you can buy a cooling fan for it for $200, though I’m not sure it can be as cleverly controlled as with the FX30. The menu is pretty complicated and not that well suited to video, so again will suit someone who doesn’t mind putting in the time to get familiar with it.

Best of the rest

While I think these manufacturers below are a smidgeon behind in the video game at this price point, it’s really not by that much. And for anyone who is already invested in one of these ecosystems or just prefer the look or feel of these brands, it will make a lot of sense to look at one of these cameras instead.

Panasonic S5 II

Whilst you may have expected here the latest in the much loved GH series, the Panasonic Lumix GH6 ($2200) which has probably the most video friendly features of any camera (and of course is great if you have M4/3 lenses), or perhaps the Lumix S1 ($2500), I’ve gone for the newly released Panasonic Lumix S5 II ($2000) which is just coming into stock now in various stores. And even better if you can wait til the end of May will be the S5 IIX ($2200) with some significant video add-ons like internal ProRes, recording to SSD and RAW over HDMI.

The big thing is that Panasonic have finally cracked phase detect auto focus in the S5 II and that it has auto focus compatibility with EF lenses. Personally I’ve never fallen in love with Panasonic colours and image quality, but this full frame hybrid certainly packs a punch in terms of features.

Canon R6 Mark II

The R6 Mark II ($2500) is a full frame hybrid with the RF mount. You can easily adapt EF lenses to it as well, with quite a few options of adaptor – with a focal reducer, a simple pass through or a drop in filter adaptor which is very handy to use and becomes a bit like having built in NDs. And crucially, autofocus works with EF lenses, which isn’t the case for most of the cameras above.

It has good IBIS, great autofocus, decent rolling shutter and those lovely Canon colours. It also has a Tascam XLR adaptor available. It shoots 4K oversampled from the 6K sensor, and while it does overheat at 4K60, other modes are fine. It has lower dynamic range than any of my main choices, which is one of the main reasons I don’t recommend it, as well as the lack of support for third party lenses, but it’s still a great option.

Nikon Z6 II

The Nikon Z6 II ($2000) is certainly not a bad camera, it’s full frame with the Nikon Z mirrorless lens mount, but with only 8 bit 4:2:0 internal recording with no log and a 30 minute time limit, it is behind the pack. You can get 10 bit log recording externally, avoid the time limit and even shoot BRAW with a very good dynamic range, but you need to factor in the RAW upgrade ($200+shipping) and an external monitor into the cost.

Any other options?

The Sigma fp L ($2500) could be worth a look, with pretty decent lab results, as could the RED-esque box style Z CAM E2-S6 ($2200) or Panasonic Lumix BGH1 ($2000), though these are designed to be built into proper cinema rigs so the total cost will be a lot more. Or you could look into the second hand market for an older C300 or FS7 say if you are after a bigger camera and don’t need the latest tech.

One thing I’ve learned doing all this research – the field has never been this close and you can get fantastic results out of any of these cameras.