Topic: R8Brain and workflow

[Brian Skalinder]

I've used Samplitude SE v8.3 (SAM) for a while now and really like the s/w and overall workflow.  I recently did a couple comps of SAM at its highest sample rate conversion (SRC) quality and R8Brain Pro.  While I think the overall quality of SAM's SRC and dither is damn good, turns out my ears prefer R8Brain Pro.  So I'm trying to figure out how to incorporate R8Brain Pro into my workflow.  My current workflow in SAM alone:

• Edit
• Resample (during Track Bouncing)
• Dither (during Track Bouncing)
• Generate Tracks (File | Export WAV)

Ideal workflow:

• Edit
• Track bounce to 32/48 (no SRC in SAM)
• Resample to 32/44 with R8Brain
• Dither to 16/44 in SAM
• Generate Tracks

The problem into which I'm running:  when I bounce my recordings to 32/48, SAM only outputs files <= 2 GB, and uses the WAV "continue chunk" to append the additional data above the 2 GB threshold.  So for a bounced recording at 32/48 that's, say, 3.5 GB, SAM creates an initial 2 GB WAV and then creates a 2nd file (linked to the first via the continue chunk) with a file extension of .W01.  Unfortunately, R8Brain Pro doesn't utilize the "continue chunk", so will only recognize the first 2 GB file of the 3.5 GB fileset.

I tried cutting the 3.5 GB WAV into smaller chunks, running them in batch through R8Brain, and then importing back into SAM for dithering and final tracking.  Unfortunately, this produces audible clicks at the point I re-join the WAVs.

So, for those of you using R8Brain (or those who are not)...any ideas for how to incorporate R8Brain into my workflow when dealing with WAVs > 2 GB?

(Note:  At this time, I'm not willing to replace SAM with another editor.)

[trajhip2000]

not an elegant solution by any means, but if you can use crossfades to reassemble the smaller files after R8Brain Pro then you will avoid clicks/pops (but you probably know that).

it would certainly be helpful if there was a DirectX or VST version of R8Brain...

[datbrad]

Brian, Off topic, but how do you feel the highest quality setting SRC in Audacity stacks up?

[Brian Skalinder]

Brian, Off topic, but how do you feel the highest quality setting SRC in Audacity stacks up?

I've not done a head to head comp between Audacity and either SAM or R8Brain, but overall I found Audacity's quality just so-so.  After editing / SRC / dithering, my recordings always sounded a bit dull to my ears.  Nothing horrible, mind you, and it surely serves its purpose quite nicely as a free, good sounding, decent-featured, easy to use audio editor.  But after I checked out a SAM trial, the $50 was a no brainer for for broader functionality, better workflow, and higher quality SRC/dither.

[j5brock]

What if you recorded to mono files.. that would keep your file size under 2GB.. pan as normal with your edit but keep the files mono until your last dither in SAM - then choose stereo.. I think that would solve the 2Gb problem - I think I saw you used a 722.. so the 2 mono files would be an option.

Jeff

[macdaddy]

cant you set track markers and then split the file into tracks..? the track splits should be seemless, and the 2gb limit shouldnt be an issue...

i agree, samplitude is great - i love the workflow, too. nice to see others doing the same kind of thing

+T

[Brian Skalinder]

cant you set track markers and then split the file into tracks..? the track splits should be seemless, and the 2gb limit shouldnt be an issue...

Tried that, though perhaps I wasn't clear about it:

I tried cutting the 3.5 GB WAV into smaller chunks, running them in batch through R8Brain, and then importing back into SAM for dithering and final tracking.  Unfortunately, this produces audible clicks at the point I re-join the WAVs.

[macdaddy]

jeez - sorry i didnt read that closer where do you get this r8brain..? i could try it out, and see if i can figure something out (plus if you like it, it would be worth checking out, anyway )...

edit: for the record, i was suggesting something different - using that seperate each cd track into WAV file option when exporting the whole enchilada. i do this all the time, and it never results in gaps. (and maybe we are talking about the same thing, just in a different way)...

[Brian Skalinder]

You may find R8Brain Free and R8Brain Pro at Voxengo.  I'm using R8Brain Pro, though I suspect the same problem arises in the Free version;  Pro simply has higher quality SRC and more options.

Not sure what you mean by "separate each cd track into WAV file option"?  Do you mean File | Export Audio | Wave, and then in the Export Sample window's Files section, select from the dropdown 2. Each CD-Track in a file (from track to track marker in a file)?

Not sure how that's different from splitting into more manageable file sizes when Track Bouncing, but just in case Export Audio does something different, I tried it.  Same issue with clicks/pops when I try to re-assemble the files in SAM for dithering.

The only solution with which I can come up at the moment is to split the stereo track into two mono tracks, and bounce each separately, run them individually through R8Brain, and then bring them back into SAM for final dither / mixdown to a stereo file.  (My concern is how SAM handles mixing down two mono tracks, each panned fully L/R...does it simply flag each with a L/R channel flag, or does it do additional processing that would be better performed at 32/48?  Though I probably wouldn't notice a difference either way.)  I think as long as I record 24/48, and import mono files into SAM, that'll give me enough time before I reach the 2 GB threshold so that SAM won't need to use the continue chunk when Track Bouncing to 32/48.  Now, if I run 24/96, then I'm much more likely to exceed the 2 GB threshold when Track Bouncing to 32/96, so in that instance I think I'm SOL at the moment.

In poking around the R8Brain forum and SAM site, it seems future versions of R8Brain will support WAV64 to address the file size issue, while new versions of SAM already support RIFF64.  Unfortunately, R8Brain has no intention of supporting RIFF64, and I've not seen any indication that SAM supports (or will support) WAV64.  So even in future releases, it's still going to be an issue, unless I'm missing something important.  It ought not be this difficult.  Hmph.

<sigh>

I may just revert back to using SAM for everything.  Would keep my workflow simple, and it sounds mighty fine even though I preferred R8Brain SRC a bit.

[live2496]

In your original files, cut your audio into 2gb segments but do so at zero crossings where the audio is changing from positive to negative.

Sample rate convert those segments and reassemble into a new project for the converted audio.

Cutting at zero crossings should avoid any pop.

Gordon

[Brian Skalinder]

Cutting at zero crossings should avoid any pop.

Beauty!  Should've thought of that, as editing portions of waveforms in the past by splitting on zero crossings has solved other issues in the past.  Thanks, Gordon, I'll give it a go.

[Althea1979]

Excuse my ignorance, but what does "Track bounce to 32/48 (no SRC in SAM)" mean?

Thanks,
Chris

[Brian Skalinder]

Excuse my ignorance, but what does "Track bounce to 32/48 (no SRC in SAM)" mean?

In Samplitude (SAM), you create a Virtual Project (VIP) in which to perform work.  The VIP may contain one or more tracks.  It's in the VIP that you make virtual (i.e. non-destructive) edits to the recording:  compression, EQ, volume envelope, etc.  When you create the virtual edits in a VIP, SAM doesn't actually perform the edits at the time of creation.  It simply logs them.  SAM only applies the edits when Track bouncing (though there are a couple other methods of applying the edits, but Track bouncing provides the greatest flexibility, I've found).  Track bouncing provides multiple functions.  It <a> applies all the virtual edits, <b> mixes all the tracks together into a stereo output track (if working on a multi-channel VIP), <c> provides the option to perform sample rate conversion (SRC, for example from 48 kHz to 44.1 kHz), and <d> provides options to apply word length reduction and dither (e.g. from 24-bit to 16-bit, or 32bfp to 24-bit or 16-bit).

I typically start with a source file that's 24-bit word length, and 48 kHz sample rate.  When I say "Track bounce to 32/48 (no SRC) in SAM", I mean (breaking it down piece by piece:

"Track bounce" = <a> apply my virtual edits, and <b> mix the tracks together into a stereo output track (if I'm working on a multichannel VIP).

"to 32/48 (no SRC)" = during Track bouncing, output a 32bfp word length file, with 48 kHz sample rate, and do NOT <c> perform sample rate conversion (SRC), or <d> apply word length reduction and dither.

"in SAM" = in Samplitude.

Hope that makes sense, felt like I was rambling a bit there.  Ask away if I've just confused things further...