Sample Formats

  • Thread starter Thread starter RhythmRmixd
  • Start date Start date
RhythmRmixd

RhythmRmixd

New member
Quick question - If I am starting a new project and composing it entirely of samples, some are at 16 bit and others are at 24 bit, would it be smarter to set the global project sample format at 16 or 24? Meaning, do I just want to try to keep everything at 16 bits from the very start, and downsample anything I'm bringing into the project at 24, or can I upsample 16 bits to 24? I'm not entirely sure how upsampling works or if its even recommended. Obviously everything will eventually have to come back down to 16 for a CD, so I wasn't sure.
 
Last edited:
Best practice is to work in 24 bits until the last possible moment before down-converting to 16 bits. The conversion from 16 to 24 bits is lossless, but once you've thrown away the resolution of 24-bit samples, it's gone forever.
 
Would upsampling from 16 to 24 bit offer any advantages when processing the 24 bit sample through effects, compression, etc., as opposed to leaving the sample at 16 bit and processing it as is?
 
Yes. Any processing will degrade the signal to some degree. With 24-bit samples, the degradation is much less significant. Even though you don't gain anything by upsampling to 24 bits, you give your processing chain more resolution to work in.

Lately I've been doing everything in a 32-bit floating point format, and I've seen some very surprising results. I took a stereo track, copied it, and then inverted the copy. Mixed together, it summed to zero -- completely inaudible. I took the mixed, seemingly silent, track, and applied about 96dB of gain to it. To my surprise, the result was virtually identical to the original track! I didn't compare it bit by bit, but it sure sounded better than I expected it to.

So my advice is to use all the bits you can get your grubby little hands on. :)

Don
 
You must log in or register to reply here.
Back
Top