Compressor encoding time

Question

Level 1

33 points

Compressor encoding time

Hello all!

Compressor 4.7 on a M2 Max Mac Studio, Sonoma 14.7.

I'm converting a 14 minutes 3840 X 2160 422 Pro-res video file to HEVC 8 bits, at 50 mbps. If I leave the default setting at the "Faster" single pass setting, it takes 4 minutes to encode, which seems to be very reasonnable. If I select the "Better quality" multipass setting, the time jumps to 2 hours and 22 minutes, which seems to be way too much!

In comparison, when doing H264 encoding, the change from single pass to multipass barely double the encoding time. We're now talking 35 times slower here!

Is it a question of the hardware encoder not being used if I'm outside of the default parameters? Or something else?

I can't readily distinguish between the 2 files, but this video is mostly steady talking head style, unchallenging material.

Thanks!

Mac Studio, macOS 14.7

Posted on Sep 27, 2024 10:34 AM

Reply

Answer 1

Sep 27, 2024 11:13 AM in response to Deromax

I tested a similar video in an M1 Max, and here is what I found:

using "Faster", the CPU use was rather low, some 55% for a "Transcoder Service" plus a few VTDecoderXPCService instances, some 6% each, for a total of less than 1 core.

using "Slower", the CPU use of VTEncoderXPCService was between 300% and 400%.

This pretty much confirms my suspicion that "slower" is software based and "faster" is hardware based.

Reply

Answer 2

Deromax Author

Level 1

33 points

Sep 27, 2024 12:00 PM in response to Deromax

Me again! :) I activated another option in HandBrake (hardware decoder, that takes care or the Pro-res decoding I guess) that halved the time once more, so under 5 minutes for a multipass encode of my 14 minutes test video. That's the kind of performance I was expecting of my Mac Studio!

So it's case closed! Sad that a freeware is dramatically besting a 200$ pro software! :/

Thanks for your time!

Reply

Answer 3

Deromax Author

Level 1

33 points

Sep 30, 2024 6:04 AM in response to Luis Sequeira1

Didn't saw a difference with that material. Maybe more active picture with a busy background would show a difference. I'm also testing the end bitrate to determine how low I can go! I used to encode H264 1080P at 20 mbps. 4k is four time as much pixels, but HECV is twice the efficiency, so I guess 40 mbps would be equivalent?

Reply

Answer 4

Oct 11, 2024 2:12 AM in response to Luis Sequeira1

Luis Sequeira1 wrote:

This pretty much confirms my suspicion that "slower" is software based and "faster" is hardware based.

Have you tested if the "faster" HW (GPU?) based output is the same quality as the "slower" SW based output especially with difficult input (poor lighting, lots of noise and movement etc)?

I tested this ages ago and AFAIR SW output was better quality but I am not anymore sure about that and which macOS and FCP versions I used then.

At least ffmpeg is said to have clearly better quality with SW encoding. I have not yet tried that and there might be a difference between GPU and Intel CPU's VAAPI hardware acceleration.

Reply

Answer 5

Deromax Author

Level 1

33 points

Oct 11, 2024 6:33 AM in response to Matti Haveri

Hi! In Handbrake, I select the "Video toolbox" in the codec list, which is Apple language for their own hardware encoder. I setup the same parameters, like multipass, same bitrate, etc. I'm pretty sure they have a competent encoder. They are touting this as a big plus for pro video jobs after all!

I didn't bought a screaming fast Mac Studio for it to be only twice as fast as my previous late 2013 iMac! ;)

Maybe there is a setting in Compressor that must be set, I didn't found it.

Reply

Answer 6

Oct 11, 2024 7:07 AM in response to Matti Haveri

Matti Haveri wrote:

Luis Sequeira1 wrote:

This pretty much confirms my suspicion that "slower" is software based and "faster" is hardware based.

Have you tested if the "faster" HW (GPU?) based output is the same quality as the "slower" SW based output especially with difficult input (poor lighting, lots of noise and movement etc)?

I tested this ages ago and AFAIR SW output was better quality but I am not anymore sure about that and which macOS and FCP versions I used then.

At least ffmpeg is said to have clearly better quality with SW encoding. I have not yet tried that and there might be a difference between GPU and Intel CPU's VAAPI hardware acceleration.

No, I have not tested with difficult to encode material. I may do so but not today and probably not tomorrow, as I have some things that need to be finished. I will try when I have more time and see how it goes. I am also curious how different, if at all, the encoded results are in an M1 Max compared to an Intel mac.

Reply

Answer 7

Deromax Author

Level 1

33 points

Oct 11, 2024 9:52 AM in response to Tom Wolsky

It’s a system software that’s calling the hardware encoder! Read this:

Video Toolbox | Apple Developer Documentation

Reply

Answer 8

Tom Wolsky

Level 10

151,831 points

Sep 27, 2024 10:39 AM in response to Deromax

Better is really quite unnecessary for that type of material. Interesting tests though. Surprising.

Reply

Answer 9

Sep 27, 2024 10:56 AM in response to Deromax

Indeed, surprising... I imagine that the faster one may use the dedicated hardware encoder, and the other may bog down in software, but it's just guessing on my part.

Reply

Answer 10

Deromax Author

Level 1

33 points

Sep 27, 2024 11:07 AM in response to Deromax

I also tried HandBrake, which will do a multipass encode in 8m21s using the Video ToolBox preset, more in line with my expectations!

Reply

Answer 11

Sep 27, 2024 12:06 PM in response to Deromax

Can you notice any difference between this multipass encoded version and Compressor's single pass?

Reply

Answer 12

Tom Wolsky

Level 10

151,831 points

Sep 27, 2024 12:20 PM in response to Luis Sequeira1

His media may not be representative as a test. Better to try it with something more complex.

Reply

Answer 13

BenB

Level 6

12,559 points

Oct 10, 2024 5:52 PM in response to Deromax

Compressor may estimate 2 hours, but I highly doubt, if you actually run it, that it takes that long. Those estimates are not all that accurate.

Reply

Answer 14

Deromax Author

Level 1

33 points

Oct 11, 2024 6:27 AM in response to BenB

Hello! This is not the estimated time, it's the time it really took to do the job, as checked with the clock and in the Compressor "Completed" pane report.

Reply

Answer 15

Deromax Author

Level 1

33 points

Oct 11, 2024 8:51 AM in response to Luis Sequeira1

It would be surprising that Apple put an inferior codec in its hardware implementation! This is pro stuff! :)

Reply