ProRes 422 LT/HQ vs h265, performance in DR.

I would like to compare these formats/codecs, in terms of agility in editing/post-production and consumption of system resources... especially between Apple ProRes 422LT 10 bit vs h265 422 10 bit All-Intra.

The bandwidth and file size are very similar, the quality seems to be very similar too... however, which one performs better in editing... but so that things go evenly, let's make the comparison taking into account that we are using a processor intel with hardware accelerated decoding, for example intel i7 gen 12... Whether for laptop or desktop, both have decoding, however the laptop version has a more powerful graphics chip.

Additionally, I would like to compare it with BRAW 12:1 or similar, in terms of performance when editing/post-production.

It all depends on your hardware. If you have the necessary decoders, H.265 can be just as fast. If not, ProRes is much smoother. I doubt that ProRes LT always looks as good as HEVC in 4:2:2 10 bit, with a good encoder it can be close to ProRes 4:2:2 HQ without needing as much space.
BRAW is more compact than ProRes for the same quality, but doesn't really compare, since it's RAW (kind of).

It all depends on your hardware. If you have the necessary decoders, H.265 can be just as fast. If not, ProRes is much smoother. I doubt that ProRes LT always looks as good as HEVC in 4:2:2 10 bit, with a good encoder it can be close to ProRes 4:2:2 HQ without needing as much space.
BRAW is more compact than ProRes for the same quality, but doesn't really compare, since it's RAW (kind of).

I understand, but I suppose that, for example, on the iPhone, ProRes 422LT has more quality, compared to using HEVC, which in the specific case of the iPhone is 10bit 4:2:0.

However, for a 4:2:2 10bit HEVC all-intra file from a Sony a7sIII/A7IV, that is better than ProRes 422 LT.

In any case, I suppose that ProRes 422LT has better quality than HEVC 10 bit 422 GOP.

I understand, but I suppose that, for example, on the iPhone, ProRes 422LT has more quality, compared to using HEVC, which in the specific case of the iPhone is 10bit 4:2:0.

However, for a 4:2:2 10bit HEVC all-intra file from a Sony a7sIII/A7IV, that is better than ProRes 422 LT.

In any case, I suppose that ProRes 422LT has better quality than HEVC 10 bit 422 GOP.

For the iPhone, it all depends on the app you use. There is a secret camera app (secret for this forum) where you can adjust the data rate for HEVC. The highest setting produces even more data than a Sony Alpha. Please contact me by PM and I'll admit you to the small, elite user group

For natural images, all-intra is only better with considerably higher data rates. GOPs are more efficient, but harder to decode on older computers.

If I could choose HEVC 422 10bit on the iPhone I would be so happy.

However, the BlackMagic app does not allow this at the moment, however all the other functions of the blackmagic app are very good for the iPhone.

I am writing to you privately through the app that mentions c:

I would like to compare these formats/codecs, in terms of agility in editing/post-production and consumption of system resources... especially between Apple ProRes 422LT 10 bit vs h265 422 10 bit All-Intra.

The bandwidth and file size are very similar, the quality seems to be very similar too... however, which one performs better in editing....Additionally, I would like to compare it with BRAW 12:1 or similar, in terms of performance when editing/post-production...4:2:2 10bit HEVC all-intra file from a Sony a7sIII/A7IV, that is better than ProRes 422 LT...

Sony's "All-Intra" XAVC-SI codec used on the A7SIII does not use H.265, aka HEVC. It uses AVC (ie H.264) for intra-frame compression. You can see that by examining the files using the third-party MediaInfo tool: https://mediaarea.net/en/MediaInfo

The recording bitrate of XAVC-SI and ProRes 422LT are about the same. In theory the quality should be about the same.

Since XAVC-SI is not a "Long GOP" but All-Intra, it is generally smooth to edit. The A7SIII and similar cameras cannot record in ProRes internally, so you'd have to use an external recorder such as an Atomos Ninja V.

If you are not talking about acquisition codec but a mezzanine or intermediate codec for post, in general you don't need that in this case. You can edit 10-bit 4:2:2 XAVC-SI directly with good performance and quality, and little would be gained by transcoding that to ProRes 422LT.

For a larger post-production scenario, sometimes transcoding to ProRes helps from a commonality and procedural standpoint. IOW all the batch jobs and procedures are set up for ProRes, so the staff doesn't have to debug proprietary codecs.

OTOH some pro codecs such as Sony's XAVC-I in the MXF wrapper from the FX6 contains much better metadata than the transcoded or simultaneously-recorded ProRes 422, so in that case the proprietary codec has advantages. But the XAVC-SI format used by Sony's mirrorless cameras use the MP4 wrapper, which has more limited metadata and in some cases timecode is not handled well.

BRAW 12:1 is a different case -- it's obviously RAW so is well-suited for post production with Resolve. In general editing performance of BRAW is good on contemporary computers. Only Blackmagic hardware can record BRAW.

If I could choose HEVC 422 10bit on the iPhone I would be so happy.

However, the BlackMagic app does not allow this at the moment, however all the other functions of the blackmagic app are very good for the iPhone.

Actually, the BM Camera app can record HEVC in 4:2:2 and 10 bit. But the data rate is lousy at 14 mbps and can't be changed (yet?).

@Joe
Yes, even the Sony MP4 files contain a ton of metadata, readable with Sony's Catalyst Browse, but most of that doesn't arrive in DR. Profile settings would be nice, at least.

I would like to compare these formats/codecs, in terms of agility in editing/post-production and consumption of system resources... especially between Apple ProRes 422LT 10 bit vs h265 422 10 bit All-Intra.

The bandwidth and file size are very similar, the quality seems to be very similar too... however, which one performs better in editing....Additionally, I would like to compare it with BRAW 12:1 or similar, in terms of performance when editing/post-production...4:2:2 10bit HEVC all-intra file from a Sony a7sIII/A7IV, that is better than ProRes 422 LT...

Sony's "All-Intra" XAVC-SI codec used on the A7SIII does not use H.265, aka HEVC. It uses AVC (ie H.264) for intra-frame compression. You can see that by examining the files using the third-party MediaInfo tool: https://mediaarea.net/en/MediaInfo

The recording bitrate of XAVC-SI and ProRes 422LT are about the same. In theory the quality should be about the same.

Since XAVC-SI is not a "Long GOP" but All-Intra, it is generally smooth to edit. The A7SIII and similar cameras cannot record in ProRes internally, so you'd have to use an external recorder such as an Atomos Ninja V.

If you are not talking about acquisition codec but a mezzanine or intermediate codec for post, in general you don't need that in this case. You can edit 10-bit 4:2:2 XAVC-SI directly with good performance and quality, and little would be gained by transcoding that to ProRes 422LT.

For a larger post-production scenario, sometimes transcoding to ProRes helps from a commonality and procedural standpoint. IOW all the batch jobs and procedures are set up for ProRes, so the staff doesn't have to debug proprietary codecs.

OTOH some pro codecs such as Sony's XAVC-I in the MXF wrapper from the FX6 contains much better metadata than the transcoded or simultaneously-recorded ProRes 422, so in that case the proprietary codec has advantages. But the XAVC-SI format used by Sony's mirrorless cameras use the MP4 wrapper, which has more limited metadata and in some cases timecode is not handled well.

BRAW 12:1 is a different case -- it's obviously RAW so is well-suited for post production with Resolve. In general editing performance of BRAW is good on contemporary computers. Only Blackmagic hardware can record BRAW.

Thank you very much for the clarification, even Chat GPT did not notice the h264 issue in that mode of the Sony A7s III.

Taking into account what you say, I understand that the "All-Intra" of Sony A7SIII is very similar to the ProRes 422LT, although of course Resolve can recognize these Sony files in the "RAW" tab, although I don't know if this provides anything significantly better. Compared to the iPhone file (which can even be recorded in A-Log)... it is clear that the Sony a7sIII sensor is much superior to that of the iPhone.

Regarding BRAW, I compared it because I understand that 12:1 produces a file/bandwidth similar to Sony's h264 4:2:2 "all-intra", and to ProRes 422 LT. Which do you think is easier to process? for a PC with Davinci Resolve? I understand that BRAW is superior in quality, I am not arguing that.

However, the curious thing is that almost no graphics card has accelerated decoding in h264 10bit 422, however Intel's integrated graphics do in h265.


FONT: https://www.pugetsystems.com/labs/artic ... udio-2122/

BRAW 12:1 is a different case -- it's obviously RAW…

It is not, it does not store actual sensor photosite information pre-debayering in the sense raw capture is usually defined. It does it ”for reasons”. Why it is still called raw, ask blackmagic. By same ”extended” logic prores422 is raw too.

BRAW 12:1 is a different case -- it's obviously RAW…

It is not, it does not store actual sensor photosite information pre-debayering in the sense raw capture is usually defined. It does it ”for reasons”. Why it is still called raw, ask blackmagic. By same ”extended” logic prores422 is raw too.

Well I also have a similar point of view. I'm not too expert.

The interesting thing is that blackmagic says that BRAW eliminates noise in compression... What Sony does, let's say using h264/5.

It turns out that now the iPhone also eliminates noise in compression (ProRes/h264/h265)... This thing about eliminating noise when compressing ProRes is not done by Atomos, for example.

It is fantastic that the iPhone can eliminate noise in compression, although details are lost, in many cases this is a great advantage.

I think that with this in mind, iPhone with ProRes is not very different from BRAW... You want something light and you use 422LT, you want more information you take the ProRes XQ 4444.

BRAW 12:1 is a different case -- it's obviously RAW…

It is not, it does not store actual sensor photosite information pre-debayering in the sense raw capture is usually defined...

If you check the several thousand posts I've made in various imaging and post-production forums, you'll see I've made that exact point several times -- but in a diplomatic manner, and when that was the actual topic under discussion.

If I could choose HEVC 422 10bit on the iPhone I would be so happy.

However, the BlackMagic app does not allow this at the moment, however all the other functions of the blackmagic app are very good for the iPhone.

Actually, the BM Camera app can record HEVC in 4:2:2 and 10 bit. But the data rate is lousy at 14 mbps and can't be changed (yet?).

14 MEGABYTES or megabits per second?

... blackmagic says that BRAW eliminates noise in compression... What Sony does, let's say using h264/5...It turns out that now the iPhone also eliminates noise in compression (ProRes/h264/h265)....

I'm not sure what that means—all cameras and all formats have noise. If you mean compression artifacts, there could be tiny differences, but these will not be noticed by the final viewer after post-production and distribution and under normal viewing conditions.

....It is fantastic that the iPhone can eliminate noise in compression, although details are lost, in many cases this is a great advantage...

Again, I'm not sure what that means. I've shot multi-cam interviews with dual iPhone 15 Pro Maxs using a variety of codecs from the Blackmagic Camera App, iOS camera app and Filmic Pro. If you understand composition and lighting, you can make it look pretty good. It's not equal to a pro camera or even an A7SIII, but it can be fairly good.

The problem is that requires knowledge, skill, preparation and practice, just like any camera.

....with this in mind, iPhone with ProRes is not very different from BRAW... You want something light and you use 422LT, you want more information you take the ProRes XQ 4444.

I think you are over-thinking this. The main factors are lighting, composition, and understanding the limitations of the given camera system. If you have that knowledge, an iPhone 15 Pro Max can shoot decent interview video. With all other factors held constant, the difference between various ProRes flavors and BRAW can be relatively small. You would rarely shoot ProRes 4444XQ due to the extreme data size. It is very burdensome, especially at > 4k resolutions.

Except for alpha channel transparency and possible bit depth, there is very little difference in acquisition image quality or editing image quality if comparing ProRes 422 vs ProRes 4444XQ.

The main differences are the 4444 codec better supports 12-bit data (if available) and the higher bitrate variants better sustain generation loss if doing multiple sequential re-encodes. That implies a hand-off each time. By contrast an NLE timeline can be rendered 1000 times during the edit, and that does not degrade quality. The edits are recorded as metadata in the project file or library. Each render reads the original files, not files containing multi-generation losses.

If anyone is curious about generation loss or encoding quality loss, it is straightforward to objectively measure those using some free command-line tools.

On MacOS, Apple provides AVQT. It is designed to measure perceptual quality loss, not just technical parameters. It can be downloaded from the Apple Developer site but that requires you create a free developer account (click "Account" at top-right corner of this page: https://developer.apple.com/.)
WWDC21 talk on AVQT: https://developer.apple.com/wwdc21/10145
Youtube tutorial on AVQT:

You can also measure visual generation loss using a tool developed by NetFlix called VMAF (Video Multi-Method Assessment Fusion). Unlike AVQT, it is available on both Mac and Windows. VMAF is built into certain versions of ffmpeg. On Mac, ffmpeg-VMAF is available through the Homebrew package manager: https://brew.sh

On Windows, pre-built ffmpeg-VMAF binaries are available here: https://www.gyan.dev/ffmpeg/builds/

I mean grainy noise... which creates high isos.

I'm trying to compare formats and codecs, I'm not trying to compare lighting or framing, I totally understand that the quality of a video doesn't just depend on the codec... but that's what I'm trying to compare right now.

14 MEGABYTES or megabits per second?

Please note that I wrote with a minuscule 'b'. That's how you normally write 'bits', while a majuscule 'B' indicates bytes. So, those 14 mbps are roughly 1.75 megabytes per second…
The secret app – you know it now, since I sent you 'admission' to the small user group – can record up to 40 MB/s, which would be 320 mb/s.

Regarding noise reduction: BRAW has some (really mild), the Sonys have some more (defeatable for stills, but not for video), and the iPhone has a lot. Not defeatable, even in log.
That's the main reason why the tiny sensor can still look good in low-light. But the Sonys with a large sensor preserve far more detail at night for video.
The iPhone does impressive things at night for photos by integration of a series of stills. You can actually do astrophotography with an app called AstroShader.

14 MEGABYTES or megabits per second?

Please note that I wrote with a minuscule 'b'. That's how you normally write 'bits', while a majuscule 'B' indicates bytes. So, those 14 mbps are roughly 1.75 megabytes per second…
The secret app – you know it now, since I sent you 'admission' to the small user group – can record up to 40 MB/s, which would be 320 mb/s.

Regarding noise reduction: BRAW has some (really mild), the Sonys have some more (defeatable for stills, but not for video), and the iPhone has a lot. Not defeatable, even in log.
That's the main reason why the tiny sensor can still look good in low-light. But the Sonys with a large sensor preserve far more detail at night for video.
The iPhone does impressive things at night for photos by integration of a series of stills. You can actually do astrophotography with an app called AstroShader.

I understand the lowercase b, it's just that this is sometimes omitted, that's why I was asking to clarify it.

Because also 14Mbps, which would be 1.7MBps... clearly it is a very very small size.

The advantage of the blackmagic app is that everything is free (except the cloud thing, which I don't even care about), and it makes me feel like I'm operating a blacmagic camera. If I record in ProRes 422 LT, I guess I don't need any other app, anyway thank you very much, it is likely that many people will find it very useful.

Uli,

The secret app – you know it now, since I sent you 'admission' to the small user group – can record up to 40 MB/s, which would be 320 mb/s.

now I'm curious!

Uli,

The secret app – you know it now, since I sent you 'admission' to the small user group – can record up to 40 MB/s, which would be 320 mb/s.

now I'm curious!

We have to see how other features such as stabilization work in that app. In filmic pro this was not as good as the native app.

Filmic Pro also remembers that it overheated the iPhone 13 pro in my personal experience.
These details are what I think Blackmagic would take into account to improve.

Uli,

The secret app – you know it now, since I sent you 'admission' to the small user group – can record up to 40 MB/s, which would be 320 mb/s.

now I'm curious!

We have to see how other features such as stabilization work in that app. In filmic pro this was not as good as the native app.

Agreed. Another thing that iPhone native camera has that makes it very hard for me to switch to any other app is very good face auto-focus.

I still can't test facial focus in depth in the blackMagic APP...
If you could try it. Does it work badly?

I think it worked relatively well. And have you tried the stabilization in the black magic app?

Both AF and stabilization are things that the physic black magic camera cannot do well... however, I think that those things do not matter to them because they are more oriented to be used with follow focus and with some type of gimbal or steadycam.

I still can't test facial focus in depth in the blackMagic APP...
If you could try it. Does it work badly?

I think it worked relatively well. And have you tried the stabilization in the black magic app?

To be fair my statement is very generalized and I haven't tested all apps extensively. All I'm trying to say is that my impression so far is that in those regards native app seems to perform better and has home turf advantage.

I still can't test facial focus in depth in the blackMagic APP...
If you could try it. Does it work badly?

I think it worked relatively well. And have you tried the stabilization in the black magic app?

To be fair my statement is very generalized and I haven't tested all apps extensively. All I'm trying to say is that my impression so far is that in those regards native app seems to perform better and has home turf advantage.

Well it's quite reasonable, they know their system better than anyone...

However, the manual controls are what it does not have... But I suppose that BlackMagic is the firm that could come closest, beyond the fact that in its cameras the stabilization and autofocus work poorly, but more for a hardware issue than a software.

I still can't test facial focus in depth in the blackMagic APP...
If you could try it. Does it work badly?

I think it worked relatively well. And have you tried the stabilization in the black magic app?

To be fair my statement is very generalized and I haven't tested all apps extensively. All I'm trying to say is that my impression so far is that in those regards native app seems to perform better and has home turf advantage.

Well it's quite reasonable, they know their system better than anyone...

However, the manual controls are what it does not have...

Yes, and what I need / wish for is one that will excel at both.