FORUMS: list search recent posts

any way to edit 32-bit floating point audio in fcp without rendering?

COW Forums : Apple Final Cut Pro Legacy

<< PREVIOUS   •   FAQ   •   VIEW ALL   •   PRINT   •   NEXT >>
Douglas Glover
any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 8, 2012 at 11:05:09 pm

Hi,
I have an audio dilema. I am working on a reel and had to convert around 100 dvd's to quicktimes. The codec that I used made the audio format 32-bit floating Point and 48khz. As I'm finding now, FCP (version 7.0.3)
only allows 24-bit floating point in the sequence presets. This forces me to render the audio for each quicktime. I am debating going back and re-rendering every quick time but this will take forever. I can't seem to find a codec in compressor that just compresses the audio without taking 3 hours per quicktime to render the video as well.
Is there a way to edit with a 32-bit audio file in fcp? If not, is there a way to change the audio in compressor from 32-bit to 24 or 16 bit without having to have an immense amount of compression time for 100 quick times?
Thanks for the help

Douglas Glover
http://www.douglasglover.com


Return to posts index

Michael Gissing
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 8, 2012 at 11:08:59 pm

FCP doesn't handle floating point audio files. It has to have 16 or 24 bit fixed, 48khz. The FCP sequence can be set to 32 bit float but that is only how FCP does its internal audio processing.


Return to posts index

Jeff Meyer
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 9, 2012 at 10:08:56 pm

You should be able to build a Compressor preset that will do a video pass-through and then dial in the audio settings you want.


Return to posts index


Douglas Glover
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 9, 2012 at 10:50:33 pm

Hi,
Thanks for the advice. Is there a way to do a video pass through without altering the video so that it just alters the audio and doesn't take so long?

I've tried using the same video settings as the quicktime in compressor and just altering the audio but it still seems to take 3 hours per video. This means that it's obviously doing something to the video instead of just the audio.
Thanks

Douglas Glover
http://www.douglasglover.com


Return to posts index

Jeff Meyer
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 10, 2012 at 1:59:08 am

Going back I see a pass-through for the audio, but not for the video.

That siad, if it's taking three hours I'm hesitant to believe that your settings are the same unless you're using a long GOP codec like H.264 or mp4 or the videos are three hours long. If you aren't working in a RT codec like any member of the ProRes family or the DV family it would be wise to switch to such a codec along the way.

Alternatively, you could do a pass in Compressor with the video disabled, then reconnect to the audio re-encoded in Final Cut. This isn't an ideal situation, but it would work quickly if transcoding is taking too long.


Return to posts index

Rafael Amador
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 10, 2012 at 2:39:04 am

Does exist something like "32b FP" audio?
Floating point is a fashion of processing at 32 bits depth, but on my advice do not exist "32bFP audio files".
FC renders audio at 32b FP, but write them at 8,16 or 24b.
I think this is all a misunderstanding with 32Khz files.
Set your FC sequence to 32Khz instead of 48Khz.
rafael

http://www.nagavideo.com


Return to posts index


Michael Gissing
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 10, 2012 at 3:31:01 am

Yes Rafael, there is a 32 bit float audio file format. In a program like Audcaity, Nuendo and Samplitude it is a file record format option.


Return to posts index

Rafael Amador
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 10, 2012 at 11:42:37 am

[Michael Gissing] "Yes Rafael, there is a 32 bit float audio file format. In a program like Audcaity, Nuendo and Samplitude it is a file record format option."
Hi Michael,
Those programs can sample and process at 32bFP, but no one of them put out any kind of file/codec/format that is 32bFP.
FP is about wider dynamic range and even if possible, they make no sense in an audio track of a video clip.
In Video/graphics, FP is used to keep off-specs levels.

[Douglas Glover] " As I'm finding now, FCP (version 7.0.3)
only allows 24-bit floating point in the sequence presets. "

FC only allows 8, 16 and 24b samples at 32, 44,1, 48, 44,2 and 96 Khz. Nothing related with Floating Point here.
rafael

http://www.nagavideo.com


Return to posts index

Walter Soyka
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 10, 2012 at 1:55:30 pm

[Rafael Amador] "FP is about wider dynamic range and even if possible, they make no sense in an audio track of a video clip. In Video/graphics, FP is used to keep off-specs levels. "

Floating point has three big advantages for visuals:
  • Large dynamic range (the difference between black and white)
  • Values above and below clipping (values brighter than white and darker than black, very useful when chaining effects)
  • Very high precision (leading to reduced quantization or rounding errors, also useful when chaining effects)


With fixed point or integer math, the gap between any two adjacent numbers is constant. With floating point, the gap between any two adjacent numbers is proportional to the number itself -- about ten million times smaller than the number.

In other words, large numbers are spaced further apart, but small numbers are spaced closer together, giving you incredible precision in the lower/darker/quieter end of the range.

I'm not an audio whiz, but I imagine all these benefits would be as important with audio signal processing as they are with video.

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index


Rafael Amador
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 10, 2012 at 3:12:21 pm

As usual you nailed it down Walter.

"In Video/graphics, FP is used to keep off-specs levels. ""
When i wrote that I was thinking about what i consider the most interesting feature of FP for my workflows.
When I use NeatVideo in FC before sending to Color, thanks to the FP capability of FC, theres is not any signal clipping. In that case the 32bFP process ends up on a 10bYUV file with the highlights untouched.

[Walter Soyka] "I'm not an audio whiz, but I imagine all these benefits would be as important with audio signal processing as they are with video."
I agree, but my point was about "Floating Point" formats, or codec or files.
I understand Floating Point as way of making maths, but not an option for storing data.
However I'm not a sound expert neither.
rafael

http://www.nagavideo.com


Return to posts index

Michael Gissing
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 10, 2012 at 10:52:57 pm

Hi Rafa. I also checked another software program - Voxengo R8Brain. It is a batch sample rate and bit depth file converter. In output file options it offer 16, 24, 32 floating point and 64 floating point.

http://www.voxengo.com/product/r8brain/

I don't understand how a file can be floating point either but they are file options. Of course FCP cannot import 32 or 64 bit audio files but they do exist.


Return to posts index

Michael Gissing
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 10, 2012 at 10:55:21 pm

I should add that this software is a fee utility but Window only. I can confirm it works in Linux Mint11 using WINE


Return to posts index


Walter Soyka
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 10, 2012 at 10:57:47 pm

[Michael Gissing] "In output file options it offer 16, 24, 32 floating point and 64 floating point. "

Wow. 64-bit floating point?

A lot of visual systems cheat and use half-float for storage, because full 32-bit floating point is usually unnecessary. I'd think that double-float would be absurd overkill.

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index

Michael Gissing
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 10, 2012 at 11:02:46 pm

Yes Walter absurd overkill has been common in audio where bandwidths are low and people imagine all sorts of benefits from a file with dynamic ranges of over 300db.

Hi bit rate floating point is perfectly understandable for audio processing like EQ and mixing many hundreds of tracks together as it enable rounding errors to be very low and discarded when the final 24 bit file is produced. But electronics cannot better 21 bits due to solid state thermal noise so even the 130+ db dynamic range of 24 bit is not reproducible. But people will swear that a 32 bit file sounds better though the same 24 bit D to A.


Return to posts index

Walter Soyka
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 10, 2012 at 11:12:54 pm

[Michael Gissing] "Yes Walter absurd overkill has been common in audio where bandwidths are low and people imagine all sorts of benefits from a file with dynamic ranges of over 300db."

Can you frame that for me? Decibels are logarithmic, right? 300 dB strikes me as incomprehensibly loud.


[Michael Gissing] "Hi bit rate floating point is perfectly understandable for audio processing like EQ and mixing many hundreds of tracks together as it enable rounding errors to be very low and discarded when the final 24 bit file is produced."

Totally understand. I've saved out image sequences in floating point as intermediates destined for additional processing before final delivery.

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index


Michael Gissing
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 10, 2012 at 11:18:46 pm

Walter, 300db would kill you. It is theoretical dynamic range available through 32 bit floating point processing.


Return to posts index

Rafael Amador
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 11, 2012 at 2:22:14 pm

Very interesting and nothing to discuss.
You know that you are my audio adviser michael :-)

BTW one question Walter.
Do After Effects works in 32FP, or just 32b?
I ask because, on my experience, AE clips any peak over 100%.
rafael

http://www.nagavideo.com


Return to posts index

Walter Soyka
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 12, 2012 at 2:17:29 am

[Rafael Amador] "Very interesting and nothing to discuss.
You know that you are my audio adviser michael :-)"


Indeed. Thanks for the interesting diversion, gents.


[Rafael Amador] "BTW one question Walter.
Do After Effects works in 32FP, or just 32b?
I ask because, on my experience, AE clips any peak over 100%."


It's 32b FP.

Starting with AE CS3 (when Adobe introduced their MediaCore framework instead of relying on QuickTime for RGB/YCbCr conversion), AE can handle overbrights from certain codecs in 32bpc projects.

If you need to get overbrights back to FCP from AE, render with Trillions of Colors to the AJA 10-bit RGB format and set FCP to process maximum white as super-white.

There are only a handful of output formats that support floating point, and they're all image formats (so you'd have to use image sequences with an app that supports them): Photoshop, OpenEXR, and TIFF. I'll also note that OpenEXR only stores half-float (16b FP) by default, though it can be forced to store 32b FP if you choose.

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index


Rafael Amador
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 12, 2012 at 5:25:03 pm

[Walter Soyka] "If you need to get overbrights back to FCP from AE, render with Trillions of Colors to the AJA 10-bit RGB format and set FCP to process maximum white as super-white."
Walter, thanks for the info, but, won't ProresHQ keep the overbrights?
In Color is possible.
rafael

http://www.nagavideo.com


Return to posts index

Walter Soyka
Re: any way to edit 32-bit floating point audio in fcp without rendering?
on Feb 13, 2012 at 2:19:33 pm

[Rafael Amador] "Walter, thanks for the info, but, won't ProresHQ keep the overbrights?"

From FCP to AE? Yes. From AE back to FCP? I don't think it does (but I'm not in front of a Mac right now to check).

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
© 2018 CreativeCOW.net All Rights Reserved
[TOP]