Using Audacity to create 5.1 FLAC from multi-channel wav files.

QuadraphonicQuad

Help Support QuadraphonicQuad:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Neil Palfreyman

1K Club - QQ Shooting Star
Joined
Jun 15, 2011
Messages
1,323
Location
Wiltshire, UK
Although I have used other audio editors extensively, I'm very new to Audacity but need to use it to create a 5.1 flac from a set of wavs that I have been sent by an audio engineer for play testing prior to release.

It goes without saying that it's absolutely essential that I get as close to possible, preferably identical to what he’s hearing in his audio workstation (differing hardware aside, of course.)

I think I have figured out the process but as I couldn’t find any tutorials or manual contents on multi-channel and being new to Audacity I thought I would come here and ask for advice. The procedure below describes what I’m doing. What I’m looking for is for someone who knows how Audacity works with multi-channel flac to “peer-review” what I’m doing and tell me if it’s right, if there are any “gotchas” or problems to look out for that could result in my flac not being the same as his original in the studio.

• I have four audio (.wav) files from the engineer. A front stereo pair(LR), centre, LFE and a rear stereo pair (LR.)
• I’m using the import menu to import each in turn so that I end up with six channels across four tracks in Audacity – a stereo Front L/R pair, mono Centre, mono LFE, a stereo Rear L/R pair. Audacity telly me they are all 48k 32-bit float files.
• In preferences I have set Import/Export to allow custom mixes (I read about this in an Audacity forum post.)
• I’m then using the export dialogue, type “FLAC Files”, level 5 compression, 24 bit.
• This brings up the Advanced Mixing Options which allows me to check my channel mapping, which I have checked against the flac standard for multi-channel and once I have checked and continue it produces a flac.

The flac looks and sounds right when I stream it to my player for play testing but as I said earlier it’s essential that I do this right avoiding any problems such as channel mapping errors, phase issues, track offsets and anything else that could invalidate the play testing.

Also, are there any other settings that could impact the import or export adversely that I should be looking at?

Any comments or suggestions gratefully received, even if it’s only a “yes, that all okay.”

Many thanks in advance!
 
That all sounds legit. In my experience with Audacity, front L/R are channels 1 & 2, center is channel 3, lfe is channel 4 and rear L/R are channels 5 & 6. If it sounds right on playback you probably already got that right (and if you always import in the order you specified you can map straight across at export).
 
I cannot comment on Audacity. What I can say is, you are on a roll with all this testing :) Great work Neil. Especially when it involves artists new to 5.1!

I've been doing it for quite a while but right now is the most I have ever done in a short space of time. Some exciting stuff in the pipeline! :)
 
That all sounds legit. In my experience with Audacity, front L/R are channels 1 & 2, center is channel 3, lfe is channel 4 and rear L/R are channels 5 & 6. If it sounds right on playback you probably already got that right (and if you always import in the order you specified you can map straight across at export).

Great, that's exactly what I'm doing and it sounds "right".

The engineer just confirmed to me that his files are actually 48k 24bit so I'm thinking Audacity is doing a conversion up to 32 bit float on import, which will be exact since it's just a 32 bit representation of 24 bits, and then dither back down to 24 bit on export. I'm assuming that any dither "noise" introduced will be so low as to be inaudible?
 
Great, that's exactly what I'm doing and it sounds "right".

The engineer just confirmed to me that his files are actually 48k 24bit so I'm thinking Audacity is doing a conversion up to 32 bit float on import, which will be exact since it's just a 32 bit representation of 24 bits, and then dither back down to 24 bit on export. I'm assuming that any dither "noise" introduced will be so low as to be inaudible?

I have never noticed any dither when exporting at 24-bit (FWIW). However, in my preferences, I have the default bit-rate set to 24-bit, and the sample conversions set to highest quality and dither set to "none" under the Quality tab. That may help, though I haven't found a conclusive write-up.
 
Back
Top