Topic: Audition 24 Bit Workflow Questions

[skotdee]

I've found I prefer Adobe Audition for editing, but Im confused on the whole 32 bit float thing and wondering if Im doing things right. Here's my workflow:

1. Open 24 bit file (converts to 32bf)
2. Apply any processing (boost gain, fades etc)
3. Apply Izotope Ozone MBit+ Dither filter
4. convert sample type (resamples to 44.1k, convert to 16 bit, dither box UNchecked)

Is this right? Should I convert to 24 bit, then again to 16 bit? Or is it ok to go straight from 32bf to 16? Also the MBit+ thing throws me. Audition combines the resample to 44.1k and convert to 16 bit in one step, under convert sample type. Since I'm applying the MBit+ dither before that step, technically Im dithering before resampling.

Any help is appreciated, just wanna make sure Im doing all this the best way...

[Brian Skalinder]

Close.  A few comments...

Creating and Saving 16-bit Files

1.  You have this step correct.  FWIW, Audition doesn't actually convert a 24-bit file to 32bfp on opening.  It's still 24-bit data, but within a 32bfp "workspace".  You can confirm this by opening a 24-bit file, then running Analyze | Statistics from the main menu before you make any edits:  the bit depth should read 24-bit in the statistics.

2.  Again correct.  It's during these edits that Audition "converts" to 32bfp.  Any section you edit will be 32bfp after applying the edit, since Audition uses an internal bit-depth of 32bfp during processing.  If there are sections of the WAV you have -not- edited, these will remain 24-bit.

3.  Convert Sample Type first, to resample to 44.1 kHz before dithering.  Set the bit-depth to 32bfp (which will disable the dithering checkbox and options).

4.  Apply MBIT+ to dither down to 16-bit.  If you run Analyze | Statistics after dithering, you should find the bit depth shows as 16-bit.

5.  Convert Sample Type again, setting the sample rate to 44.1 kHz (it won't actually resample since this sample rate matches the rate of the data, now, due to step 3.) and the bit-depth to 16-bit (with Enable Ditheringither UNticked, since we already dithered in step 4.).  All you're really doing here is truncating the bits below 16-bit (which are simply padded with 0s, after dithering), since for some reason Audition won't let you save directly as a 16-bit file without doing so, first.

6.  Save As

Creating and Saving 24-bit Files

If you want to save as 24-bit, the workflow is slightly different:

Steps 1 and 2 are same as above. 

3.  You probably won't resample if you're saving as 24-bit.  But if you do, this step's the same as above:  Convert Sample Type first, to resample to your preferred sample rate before dithering.  Set the bit-depth to 32bfp (which will disable the dithering checkbox and options).

4.  Apply MBIT+ to dither down to 24-bit.

5.  Save As.  In the Save As window press the Options button, then in the Windows PCM window, Format 32-bit data as dropdown, select 24-bit packed int (type 1, 24-bit) and UNtick the Enable Dithering checkbox (since you've already dithered from 32bfp to 24-bit in step 4.).

Creating and Saving 24-bit and 16-bit Files

If you want to create both 24- and 16-bit files, save the file in 32bfp after step 2.  Open it and perform steps 3-6, from the 16-bit section above, to save as 16-bit.  When done, re-open the 32bfp file, and then perform steps 3-5, from the 24-bit section above, to save as 24-bit.

There's a bit of discussion here, from a while back when I went through the same learning curve.

To answer your specific questions:  it's okay to dither straight from 32bfp to 16-bit (or from 32bfp to 24-bit, if you want to save 24-bit files, too.) 

[skotdee]

Wow, thanks Brian for the thorough response. I had read your previous discussion, but that was a little over my head. This makes it much easier to understand.

Thanks again.

[dorrcoq]

One thing I would say is - after you do your editing, etc. save a 24 bit version, and then do your dither and downsampling.

[Brian Skalinder]

One thing I would say is - after you do your editing, etc. save a 24 bit version, and then do your dither and downsampling.

Every edit in Audition produces 32bfp data.*  So doing as you suggest would result in multiple instances of dithering:

• Edit files in Audition, producing 32bfp data.
• Dither #1 down to 24-bit.
• Save.
• Resample, again producing 32bfp data.
• Dither #2 down to 16-bit.

No point in dithering multiple times, really.

* Audition performs all processing at an internal bit-depth of 32bfp.  It's possible to configure Audition to open a 16-bit file, perform processing at 32bfp, and then automatically -- as part of the editing process -- dither back down to 16-bit before completing processing (If you perform multiple edits with this configuration, then you're needlessly dithering multiple times.)  For 24-bit data, the only option is to process at 32bfp and produce 32bfp data as a result of that processing.

Edit to add:  Perhaps I misunderstood your suggestion.

[kirk97132]

One thing I would say is - after you do your editing, etc. save a 24 bit version, and then do your dither and downsampling.

Every edit in Audition produces 32bfp data.*  So doing as you suggest would result in multiple instances of dithering:

• Edit files in Audition, producing 32bfp data.
• Dither #1 down to 24-bit.
• Save.
• Resample, again producing 32bfp data.
• Dither #2 down to 16-bit.

No point in dithering multiple times, really.

* Audition performs all processing at an internal bit-depth of 32bfp.  It's possible to configure Audition to open a 16-bit file, perform processing at 32bfp, and then automatically -- as part of the editing process -- dither back down to 16-bit before completing processing (If you perform multiple edits with this configuration, then you're needlessly dithering multiple times.)  For 24-bit data, the only option is to process at 32bfp and produce 32bfp data as a result of that processing.

Edit to add:  Perhaps I misunderstood your suggestion.

I tend to agree with you Brian.  But there are some applications that choke on 32bit so in that instance I think a 24bit  save would be warranted. Cd wave would be an example, it will not open a 32 bit save.  .   FWIW, I use Audition almost exclusively and when I save an archive copy I do not sample it down to 24 bit.  I leave it at the 32 bit rate.   Of course I also may be misunderstanding things

[Brian Skalinder]

there are some applications that choke on 32bit so in that instance I think a 24bit  save would be warranted. Cd wave would be an example, it will not open a 32 bit save.

Not sure why one would need to use these others apps, or CD-Wave specifically, with the 32bfp file -- it's just a working file, not a final file.  Since the final file's bit-depths are 24- and 16-bit, anyway, just use these files with the other apps.

[dorrcoq]

So you're saying you WOULDN'T want to save an edited 24 bit file?  I always save my raw 24/48 file, my edited 24/48 file which I then track with CD Wave, and a dithered/downsampled 16/44.1 version.   Am I wasting my time doing that?

[Brian Skalinder]

So you're saying you WOULDN'T want to save an edited 24 bit file?  I always save my raw 24/48 file, my edited 24/48 file which I then track with CD Wave, and a dithered/downsampled 16/44.1 version.   Am I wasting my time doing that?

I thought I might have misunderstood your original posting.  No, you're not wasting your time.  The issue I'm trying to address is HOW you create your final 24/48 and 16/44 files.

Let me clarify that my whole discussion to date assumes one uses Audition's Edit View workspace.  (The Multitrack View workspace functions differently.*)  At any rate, the Edit View workspace allows for linear, single-step-at-a-time editing only.  In this case, you don't want to derive your final 16/44 file from the final 24-bit file.  Both the final 24/48 and 16/44 files should be derived from a 32/48 "working file" that includes all the editing you wish to do, other than resampling and dithering.

Using Audition's Edit View workspace, the workflow "tree" should look like this:

Workflow 1

Original Raw File (24-bit / 48 kHz)
|
Perform ALL Edits, e.g. gain, compression, EQ, whatever; everything except SRC and dither (32bfp / 48 kHz)
|
Save Edited "Working File" (32bfp / 48 kHz)
|
|--Dither to 24-bit
|  |
|  Save Final 24-bit File (24-bit / 48 kHz)
|
|
|--Resample to 44.1 kHz
   |
   Dither to 16-bit
   |
   Save Final 16-bit File (16-bit / 44.1 kHz)

As I hope the Workflow 1 tree illustrates, both the final 24/48 and 16/44 files are derived from the same 32/48 Working File.  As a result, both the final 24/48 and final 16/44 files are dithered just once:  from the 32/48 Working File to the final file format.

If you derive the final 16/44 files from the final 24/48 files, like this...

Workflow 2

Original Raw File (24-bit / 48 kHz)
|
Perform ALL Edits, e.g. gain, compression, EQ, whatever; everything except SRC and dither (32bfp / 48 kHz)
|
|--Dither to 24-bit (#1)
   |
   Save Final 24-bit File (24-bit / 48 kHz)
   |
   Resample to 44.1 kHz
   |
   Dither to 16-bit (#2)
   |
   Save Final 16-bit File (16-bit / 44.1 kHz)

...note that in creating the final 16-bit file, you're dithering twice, whereas in Workflow 1 you're only dithering once.  If possible, you should dither only a single time - as the last step - in the process of creating your final files.

* Audition's Multitrack View allows one to specify multiple edits, then apply them serially or in parallel through a single mixdown, i.e. rendering.  But when I used Audition (and granted I only used up to v1.5), I didn't utilize the Multitrack View because not all the editing features I wanted to use were available within this view.  For example, I could not amplify or normalize in the Multitrack View (at least not that I recall;  perhaps I merely didn't know how to do it).

[skotdee]

Thanks Brian, again, excellent info. Now lets take it a step further:

I often record 4 channel, either 4 mics, or 2 mics + SBD, in 2x Stereo recording format. Heres how I think it should work concerning the multitrack mixdown of the 4 channels.

1. Original raw files (2), opened separately in "edit view" workspace
2. Perform ALL edits and save each, individual, "working file" (32bf/48k)
3. Mix to taste the two 32bf "working files" in multitrack view
4. Mixdown to one 32bf "working file"
5. Continue with steps listed above for working with one 32bf "working file" to dither/convert to 24/48 & 16/44 bit files

And of course if working with mono files it would go the same way, do all your edits and save each as 32bf, mixdown to a 32bf file, then resample and dither accordingly...

[kirk97132]

Thanks Brian, again, excellent info. Now lets take it a step further:

I often record 4 channel, either 4 mics, or 2 mics + SBD, in 2x Stereo recording format. Heres how I think it should work concerning the multitrack mixdown of the 4 channels.

1. Original raw files (2), opened separately in "edit view" workspace
2. Perform ALL edits and save each, individual, "working file" (32bf/48k)
3. Mix to taste the two 32bf "working files" in multitrack view
4. Mixdown to one 32bf "working file"
5. Continue with steps listed above for working with one 32bf "working file" to dither/convert to 24/48 & 16/44 bit files

And of course if working with mono files it would go the same way, do all your edits and save each as 32bf, mixdown to a 32bf file, then resample and dither accordingly...

If you just hit save at the line item in red, you will overwrite your original file.  What I normally do is to save a master copy to a DVD disc and/or another folder just so i have the original in case something happens or I want to start over.  That may or may not matter to you.  Then I do my edits.  If you load the files into the multi track view you can still switch over to the edit view to accomplish any editing you want.  Any editing done in edit view will automatically be changed in the multi track view.  This way I can listen to what I am mixing and if there is something I want to change a quick mouse click and I can do any edits and immediately listen to the change. 

And as a side note, I've found I prefer to record 4Xmono It makes it easier to adjust the panning on individual tracks if I want or just pan each all the way for stereo. 

[Brian Skalinder]

Before I start, please note that I'm not very familiar -- hardly at all! -- with Audition's Multitrack View.  The following is based on vague recollection and an old version (1.5) of Audition.  I think one of the problems I had with Audition for multitrack work is it didn't allow me to resample during mixdown.  This created additional steps in the workflow and ultimately helped push me to Samplitude SE.  At any rate, even though Audition doesn't provide an ideal workflow (for me), it's still a quality, powerful tool.  And I could be mistaken...perhaps Audition supports resampling during mixdown , after all.  Anyway, here's how I would approach it, based on my limited knowledge, vague recollection, and old version:

1.  Open original raw files (2 x stereo) in Multitrack View workspace (FWIW, this will also open them in the Edit View workspace)
2.  Perform ALL edits possible in the Multitrack View workspace (EQ, compression, source mix, etc.)*
3.  Mixdown to 32/48 "working file"
4.  If necessary, apply final edits (if any) to the "working file" within the Edit View workspace that you could not apply in Multitrack View
5.  Continue with steps from previous posts for working with one 32bf "working file" to resample/dither to 24/48 & 16/44 bit files

* You can perform most, but not all, edits in the Multitrack View workspace using the FX rack.  Editing in the Multitrack View provides 2 big advantages over the Edit View, and you may even wish to use it for editing single stereo files: 

<1>  It allows you to preview not just single edits, but also multiple edits simultaneously, while controlling how to apply those edits (e.g. serially, in parallel).  For example, instead of just previewing the results of an EQ edit, you might also apply a little compression...and preview the EQ edit + compression effects together to hear if it achieves the desired result.

<2>  It allows you to save your session for later use.  This makes it easy to re-use the session if you need or want to re-create your final 24/48 or 16/44 files.  For example, say I do an AUD + SBD multitrack mix, applying some EQ and compression.  I finish creating my 24/48 and 16/44 files and enjoy listening to them.  But 6 months later, I decide to go back and tweak my mix because I want more AUD and less SBD in the mix (or different EQ, or less compression, or different edits of whatever kind).  So I open up the old session, which includes all the edits I made previously in the session (relative AUD/SBD mix, EQ, compression, etc.).  Then I simply adjust my AUD/SBD mix, and repeat steps 3-5.  Done!  I don't have to remember what EQ or compression edits I used, or re-apply them individually in Edit View.  I just open the old session, tweak the edits within the session, and repeat steps 3-5.  Easy!  (Well, not as easy as Samplitude SE...but still pretty easy.)

Oh, and kirkd brings up a very important point:  it's a good idea to back up the original, raw files so you don't accidently overwrite the original raw files while you're editing.

Hopefully that makes sense...