Latest post Mon, Oct 14 2019 3:21 PM by jlomicka. 55 replies.
Page 4 of 4 (56 items) < Previous 1 2 3 4
Sort Posts: Previous Next
  • Fri, Oct 11 2019 7:22 PM In reply to

    • jlomicka
    • Not Ranked
    • Joined on Thu, Feb 4 2010
    • Posts 29
    • Points 425
    • Avid Developer Moderator
      Avid Employee

    Re: 96K audio?

    I don't have the ProRes codecs installed on my Mac, but starting with a UHD 23.976 sequence that contained 96Khz audio and an imported high-res photo, I exported that as Quicktime Reference with audio mixdown to 96, whcih got me a .MOV file that referenced the .wav and a couple of .MXF files in my media files folder.  (two original and one rendered effect where I put in some timecode). I was able to transcode that to DNxHR in Quicktime Pro and preserve the 96K audio, and Adobe Media Encoder claimed it would as well, and the audio came through okay, but my video was solid black - not sure what happened there.  Adobe Media Encoder was willing to do Pro Res 4444, but again, all I got was blackness.

  • Fri, Oct 11 2019 8:02 PM In reply to

    • insch
    • Top 500 Contributor
    • Joined on Thu, Oct 13 2005
    • Posts 325
    • Points 3,810

    Re: 96K audio?

    jlomicka:

    I don't have the ProRes codecs installed on my Mac, but starting with a UHD 23.976 sequence that contained 96Khz audio and an imported high-res photo, I exported that Quicktinme Reference with audio mixdown to 96, whcih got me a .MOV file that referenced the .wav and a couple of .MXF files in my media files folder.  (two original and one rendered effect where I put in some timecode). I was able to transcode that to DNxHR in Quicktime Pro and preserve the 96K audio, and Adobe Media Encoder claimed it would as well, and the audio came through okay, but my video was solid black - not sure what happened there.  Adobe Media Encoder was willing to do Pro Res 4444, but again, all I got was blackness.

    That's a bit odd. Thanks for trying to find a way. What I did in Resolve was use the exported ProRes files from MC (rather than ref. files) and then paired that with my orginal audio exports from MOTU DP. It all worked fine until I needed to export 96k audio from the app!

    I am investigating whether I can provide 48k versions of my sequences and then export 96k audio for each and hopefully they can sync up their end.

     

    Mac Pro, 3.7 GHz Quad-core, 32GB RAM, G-Tech G-RAID Studio etc. [view my complete system specs]
  • Fri, Oct 11 2019 9:53 PM In reply to

    • insch
    • Top 500 Contributor
    • Joined on Thu, Oct 13 2005
    • Posts 325
    • Points 3,810

    Re: 96K audio?

    One thing you said that I don't understand - if MC sends video and audio, based on project settings, to output stage and then that gets processed, how come when you export a ref clip and choose to export audio as 96k it doesn't end up up-resing audio that has already been down-resed due to the lower project setting of 48k? How does it keep audio at 96k throughout in this case?

    Mac Pro, 3.7 GHz Quad-core, 32GB RAM, G-Tech G-RAID Studio etc. [view my complete system specs]
  • Fri, Oct 11 2019 10:12 PM In reply to

    • Leo Baker
    • Top 500 Contributor
    • Joined on Wed, Oct 5 2011
    • London
    • Posts 485
    • Points 5,710

    Re: 96K audio?

    I think as you've imported 96khz audio into Avid Media without re-sampling it to the project maximum which is 48khz. When you do the Quicktime ref it's like a webpage where it links to parts of the video and audio. So I thin the audio is untouched as it isn't rendered? Or would you need to always do an Audio mixdown in Media Composer then do the Quicktime Ref output? 

    MacPro6.1 Late 2013 Intel Xeon 12-Core 2.7GHz 64GB RAM AMD FirePro D700 6GB OSX Mojave 10.14.6. Media Composer 2020.5, Pro Tools Ultimate 2020.5, Avid... [view my complete system specs]
  • Sat, Oct 12 2019 1:58 AM In reply to

    • gstrudler
    • Not Ranked
    • Joined on Thu, Oct 13 2005
    • Portland, OR
    • Posts 49
    • Points 625

    Re: 96K audio?

    (post removed due to inaccuracies in information)

    MacOS 10.14.6, Media Composer 2019, Pro Tools 2019, Nitris DX, trashcan Mac Pro (12-core), dual AMD D700, 64GB RAM, BCC, Sapphire, Mocha, iZotope RX, Sonnox... [view my complete system specs]
  • Sat, Oct 12 2019 6:43 AM In reply to

    • insch
    • Top 500 Contributor
    • Joined on Thu, Oct 13 2005
    • Posts 325
    • Points 3,810

    Re: 96K audio?

    gstrudler:

    If I'm understanding correctly, Media Composer processes audio in 48khz (with limited hardware exceptions), so basically you don't want MC to touch the audio besides setting in and out points (which is not really manipulating the audio in any way).

    Then you HAVE to do a QT REFERENCE export because it's just pointing back to original media in that case and not actually touching it. DO NOT do a "regular" QT export as you have been showing as that then gets handled by MC first (processes it in 48khz) and then by the Apple QT engine.

    Also, DO NOT do any kind of audio mixdown, apply any audio effects (track-based, AudioSuite, or otherwise) and I would guess you don't even want to attempt level or pan changes to be safe. You would, however, do a video mixdown of your finished sequence to ProRes 444 first before doing the final QT Reference export.

    Thank you - that's really helpful. All I have are level changes and some audio key frames but I accept that this may end up with the file being rendered at 48k by MC prior to being exported with the QT ref. The audio that has been exported has got the fades etc. I might go back to my audio workstation and do those fades and level changes there so MC has nothing to do with this audio - then I'll just use the Digital Performer files and not use audio that has been through MC.

     

    Mac Pro, 3.7 GHz Quad-core, 32GB RAM, G-Tech G-RAID Studio etc. [view my complete system specs]
  • Sat, Oct 12 2019 1:07 PM In reply to

    • jlomicka
    • Not Ranked
    • Joined on Thu, Feb 4 2010
    • Posts 29
    • Points 425
    • Avid Developer Moderator
      Avid Employee

    Re: 96K audio?

    This is NOT TRUE, and the opposite of what I was saying.  Let me make this clear:

    (1) AudioSuite effects are rendered in the SOURCE CLIP'S SAMPLE RATE always, so the 96K-ness and all audio fidelity is preserved through the render, always.  I just verified this by running Media Tool and examining the rendered media


    (2) If you export "Audio" to .WAV to 96K, including as part of a QT reference export, you are running the player at 96K.  Because automation gain, track effects, and pan are never rendered, they will be run at 96K when doing a .WAV export to that rate, and processed in full 96K fidelity.

    The one thing that IS true is that timeline menu "mixdown" will run at project rate, and down-res you to 48K, so don't mixdown. HOWEVER, you CAN apply all the audio processing you want to your 96K source media and export to .WAV 96K and preserve all fidelity. That's why I am saying you should export to QT Reference with audio mixdown to .WAV at 96K.  The QT Ref will POINT to the 96K wave for the audio, and POINT to individual original or rendered video media for the video, and my experiments show that both Quicktime Pro and Adobe Media Encoder can process the QT Ref and preserve the 96K Audio.

    I'm also saying that the approach Insch has suggested of providing the video export and the .WAV audio export (at 96K) separately is also quite valid! if the downstream recipient can accept video and audio separately merge them back together in whatever tools they are using, then go ahead an process the audio all you want.  The only thing to avoid is mixdown to timeline, or the mult-mix tool.
  • Sat, Oct 12 2019 3:39 PM In reply to

    • insch
    • Top 500 Contributor
    • Joined on Thu, Oct 13 2005
    • Posts 325
    • Points 3,810

    Re: 96K audio?

    jlomicka:

    This is NOT TRUE, and the opposite of what I was saying.  Let me make this clear:

    (1) AudioSuite effects are rendered in the SOURCE CLIP'S SAMPLE RATE always, so the 96K-ness and all audio fidelity is preserved through the render, always.  I just verified this by running Media Tool and examining the rendered media


    (2) If you export "Audio" to .WAV to 96K, including as part of a QT reference export, you are running the player at 96K.  Because automation gain, track effects, and pan are never rendered, they will be run at 96K when doing a .WAV export to that rate, and processed in full 96K fidelity.


    The one thing that IS true is that timeline menu "mixdown" will run at project rate, and down-res you to 48K, so don't mixdown. HOWEVER, you CAN apply all the audio processing you want to your 96K source media and export to .WAV 96K and preserve all fidelity. That's why I am saying you should export to QT Reference with audio mixdown to .WAV at 96K.  The QT Ref will POINT to the 96K wave for the audio, and POINT to individual original or rendered video media for the video, and my experiments show that both Quicktime Pro and Adobe Media Encoder can process the QT Ref and preserve the 96K Audio.

    I'm also saying that the approach Insch has suggested of providing the video export and the .WAV audio export (at 96K) separately is also quite valid! if the downstream recipient can accept video and audio separately merge them back together in whatever tools they are using, then go ahead an process the audio all you want.  The only thing to avoid is mixdown to timeline, or the mult-mix tool.

    Thank you for your clarification. This certainly seems a bit of a minefield. From what I understand audio keyframes and fx don't affect the integrity of the 96k files but what about adding audio dissolves in a sequence and then rendering those? What audio resolution do the rendered sections become - the original file resolution (96k) or the project resolution (48k)? The only changes I have made to these files in my sequences is a quick audio mix into them and a keyframed fade out with some adjustments to audio levels made in the Audio Mix Tool.

    It seems that the least risky way might be for me to use one programme to create and manipulate the audio files at every stage (Digitial Performer) and the other (MC) to work on the video. I wish I could be sure that MC could do both at 96k but I'm nervous that it might do an audio conversion at some point. Are you saying that as long as I avoid audio mixdowns or exporting as an embedded file I should be ok on that front?

    Mac Pro, 3.7 GHz Quad-core, 32GB RAM, G-Tech G-RAID Studio etc. [view my complete system specs]
  • Sat, Oct 12 2019 5:33 PM In reply to

    • gstrudler
    • Not Ranked
    • Joined on Thu, Oct 13 2005
    • Portland, OR
    • Posts 49
    • Points 625

    Re: 96K audio?

    gstrudler:

    If I'm understanding correctly

    jlomicka:

    This is NOT TRUE, and the opposite of what I was saying.

    Well I was clearly mistaken haha (sorry). I went back and looked at jlomicka's explanation on page 2 and realized where I think I went wrong, which was confusing certain export behavior for overall system behavior.

     

    insch:

    but what about adding audio dissolves in a sequence and then rendering those

    I'm guessing this still applies and you should be ok to do that:

    jlomicka:

    HOWEVER, you CAN apply all the audio processing you want to your 96K source media and export to .WAV 96K and preserve all fidelity.

    Considering MC will render AS effects in 96khz, I'm not sure why it wouldn't handle audio fades that way as well (although clearly I'm not an expert in this area)

    MacOS 10.14.6, Media Composer 2019, Pro Tools 2019, Nitris DX, trashcan Mac Pro (12-core), dual AMD D700, 64GB RAM, BCC, Sapphire, Mocha, iZotope RX, Sonnox... [view my complete system specs]
  • Sat, Oct 12 2019 7:53 PM In reply to

    • insch
    • Top 500 Contributor
    • Joined on Thu, Oct 13 2005
    • Posts 325
    • Points 3,810

    Re: 96K audio?

    Fingers crossed that it will be exporting uncompromised 96k audio when doing a QT ref export. That will make life easier for me - then I can export 48k embedded sequences and an identical audio track in 96k.

    Mac Pro, 3.7 GHz Quad-core, 32GB RAM, G-Tech G-RAID Studio etc. [view my complete system specs]
  • Mon, Oct 14 2019 3:21 PM In reply to

    • jlomicka
    • Not Ranked
    • Joined on Thu, Feb 4 2010
    • Posts 29
    • Points 425
    • Avid Developer Moderator
      Avid Employee

    Re: 96K audio?

    I verified that audio dissolves between two 96Khz clips results in a 96Khz precompute without any conversions.

    So the problem in Media Composer is that the inability to have a 96Khz project means 96K is unavailable to Quicktime (non-reference) export, or to mixdown.

    That, and there are likely other export formats that can ccept 96Khz material for which it is only available if you have 96K hardware.

     

Page 4 of 4 (56 items) < Previous 1 2 3 4

© Copyright 2011 Avid Technology, Inc.  Terms of Use |  Privacy Policy |  Site Map |  Find a Reseller