Any benefits to oversampling 44.1 material?

[Zac Kyoti]

Wondering if any savvy folks out there can answer a question about oversampling. This subject can be a bag of worms, so hopefully we can stay on track with the specific question (please). Audio interfaces allow us to set our inputs and outputs at bit rates much higher than your average 44.1khz file. Forgetting for the moment any discussion of recording in, or of upsampling processes used in mastering, are there any benefits to upsampling a 44.1k file at the interface's DAC output, say to 96k, during pure playback? IE, if a channel is being routed from Live or Traktor containing natively 44.1k material, does it make any sense at all to have the interface set to 96k?

I'm up to speed with the Nyquist theory and the theory of "no new data created where there wasn't data before", but what is the story behind the internal data handling of the two programs I mentioned? I'm under the impression that with the exception of various plug-ins, these programs handle data at the rate it was natively introduced, and only the DAC in the interface upsamples (at the output), if it is set at a higher rate than the original material. I'm looking for documentation that says "Traktor/Live processes audio internally at "X" khz, but I'm not seeing that.

Or if you have your interface set to a higher sampling rate, and process 44.1k audio, is the upsampling applied to the signal before it goes to the outputs? (In which case it makes sense to set your interface to a higher rate - as in with mastering theory where most experts agree that the upsampling during processing can help smooth the data before it's dropped back to 44.1 for consumption). I guess this is mostly a "what is the actual chain of events" discussion.

I'm only talking about sampling rate here, not bit rate. Thanks everyone!

[DJPhaidon]

I'm currious also. Been doing some searching/reading, which is about the only thing I do at work, and there would be benefits if the programs basically "upsampled" during the processing.

If I had to take a guess I would say that because of it's target audience, Ableton would process at the hardware's settings, but not so sure about Traktor.

I would also bet that if your soundcard supports outputs at 96k, it wouldn't matter what the output of the program itself was, the 96k would sound better than the 48k due to upsampling in the card itself(Looking into that one). I say this only because Xone's and Evo's both contain 96k cards and have a reputation for sounding awesome, and this is even with 44.1k inputs from CDs.


some reading herehttp://www.mlssa.com/pdf/Upsampling-theory-rev-2.pdf

[Zac Kyoti]

and there would be benefits if the programs basically "upsampled" during the processing.

Yeah, this is what I've gathered from research as well. And I'm also leaning towards agreement about Ableton, as supposedly "Live generates and processes audio at the rate specified for the soundcard." (Their words). But as to whether audio sounds better being upsampled without any processing involved, I'm personally not convinced yet that there is any audible difference. The great sound quality of the mixers you mentioned I think has more to do with their particular DACs, and/or unique components than the fact that they output at 96k. It's kinda the same reason some people think Traktor sounds lacking when played right alongside a CDJ-1000: That unit has processors that are enhancing/maximizing the signal, where as Traktor is bit transparent.

[DJPhaidon]

The great sound quality of the mixers you mentioned I think has more to do with their particular DACs, and/or unique components than the fact that they output at 96k

Completely agree. I was trying(poorly it seems) to illustrate that lower end audio devices seem to gravitate toward 44.1k and 48k, while higher end is usually 96k or above, and for a reason.

It will completely depend on the quality of the whole signal chain to have an audible sound quality difference. Improve 1 thing and you might see a difference, but for real gains, the whole audio signal has to run smoothly through all top-knotch gear.

It's diving into the whole audiophile mentality that I don't have. You could be using the best 96k DAC's and 44.1k CD inputs on great speakers and notice a huge difference, BUT... as a DJ, how often are you really going to have a perfectly tuned amplifier driving sonically balanced speakers in a acoustically friendly environment... not very often.


Ohh, and I did find that "most" of the DAC chips that are used in todays audio interfaces automatically upsample/resample to the settings required, but as you said, the sound quality is more closely based on quality of the components that are upsampling and the filters, rather than the fact that it is upsampled.

[Zac Kyoti]

Word. In my particular case, I am running 44.1 audio from Traktor jackrouted into Live, where it's run through a few plugs (no idea what sample rate they process at, or if that even matters), then sent to the output of the interface. What I really want to know is if I am benefiting from setting my interface at 88.2/96/192 in this case. Something tells me yes only because of the processing in Live, but if I were to simply route without processing, I'm not so sure. I wish I knew if Traktor's FX processing benefits from setting a higher sample rate...