Skip to main content

Greetings all,

At some point I will be sending my cubendo mix to a mixer working in Pro Tools, and will probably render all VSTis etc. to 24 bit wavs. Ideally I would like to send bounced versions of the original vocal recordings for the vocal tracks so as to have any and all processing done by the mixer at that stage. However the mixer cannot spend time tuning vocals so I will be doing this myself which adds a step of processing in the vocal tracks I will be sending out. So the question is what should I be rendering these melodyne edits to: 32 bit or 24 bit? Is there a general consensus on this?
Secondly, the levels of the original recordings vary widely in a few comped sections (an error of sorts), well beyond what you would leave compression to sort-out: Would I be best off doing any gain changes in melodyne so as to present a normal file, or would it not be best to do only the most minimal pitch/time changes and leave the engineer to sort out all levels, that would have the benefit that in parts (where not melodyned) I would still be presenting the original untouched vocal recordings. I could also set the various event levels and these may well carry over to the pro tools session. All the above in the interest of the quality of the final result. I have found various posts that are similar bit none that hit the mark.



Boswell Wed, 04/15/2015 - 07:50

I'm unclear whether it's you that is doing the mix and sending the 2-track result off to be mastered, or whether you are preparing individual tracks for sending to a mix house. Either way, I would deal with the level problems on the vocal track first. Use level enveloping to bring the track up to a consistent level throughout and then put the result through the Melodyne process.

For the Melodyne output, you should choose 24-bit fixed or 32-bit float to correspond with whatever your mix software uses for its internal tracks. Rendering to an export format will probably generate 24-bit fixed point anyway.

pcrecord Wed, 04/15/2015 - 09:47

I get the feeling you might be using the stand alone version of Melodyne. Usually I use it as a plugin in my DAW.
What I do is place it before any other processing. Althought I try to track with consistent dynamics and try to coach the artists if they are not used to work with a mic.

IF you are working with the standalone, I'd advice to do what melodyne offer the best then export to a file that fits your project and mix it in the DAW.
It's always best to mix in the sauce anyway ;)

audio creature Thu, 04/16/2015 - 05:43

Thanks for your replies. I will be preparing the individual tracks to be mixed elsewhere. The result of your suggestion (Boswell) would be (in those problematic uneven sections) gain change -> pitch processing -> render then that rendered file would receive the final processing in the mixers daw. My feeling, would have been that unless the signal is so low that melodyne would not read it correctly (not the case), that gain change is best left out. My Daw has 32 bit internal processing and I can easily export the most precious tracks (the vocal and it's doubles as a 32 bit float file and the other tracks as either 24 bit or 32 bit float. I am not worried about the other tracks as they will be straight renders from vsti's with no intermediate processing steps. The question is still whether or not intermediate processed files are best saved at 32 bit float.


pcrecord Thu, 04/16/2015 - 06:47

Maybe I not getting it but I was on the impression that 32bit or 64bit float was kind of an upscale way for softwares to process files with lower resolutions to have a better end result.
I always record at 24bit96khz and I know my daw do 64bit float. Many of the plugins I use runs in 32 or 64 float but it doesn't change that the files that are being processed are 24bit.
Now, If you'd have a audio interface capable of recording in 32bit, that would be another story !
My 2 cents, keep all your files in the same 24bit format and let the DAW process at higher resolution if it can.