SurroundByUs.com

This may seem like a moronic question to some and i believe I already know the answer but would like some input from others here. Lets say i have a cd quality rip (format of your choice). I wish to make this into 5.1..I have the option in plogue to leave it as 44.1 , make it 96 48 or 192 khz I also have the option to take it from 16 bit to either 24 or 32. It is my opinion at this point that the 24 bit might help out the sound a bit 32 may as well but it seems to cause issues with some players. I dont believe that increasing something that is 44.1 to 48 or 96 is really helping. If i am wrong about this the so be it. I would really like to know what others think about this and why they feel that way. Thanks in advance.

No advantage to going to 24 or 32 bit float before Plogue, as Plogue does everything at 32 bit float. Just make sure that the last step in your workflow is going down to 24 bits.

As for upsampling to 48, 88.2, 96, or 192 kbps, the old mode of thinking was that upsampling would help minimize the presence of artifacts with an extraction-based method. I feel that, over time, we've figured out that working with things like humidity settings actually does more for that than changing the sample rate. The only minor advantage I'd say may be present for some conversions with changing the sample rate is that drum soungs (especially with ArcTan) are sometimes crisper when you work with 88.2 or 96 kbps, then downsampling. I'll leave it to Zeerround to explain why that's so. Still, though, I can't someone having to do that all the time.

First we need to define your target format.

For most of us, the target format is DTS CDs, so that (and what surcode accepts) defines the end result as 24 bit 44.1 Khz.

As I want to put all my surround songs into (something like) iTunes, I'm not going to want any other format.

But to each his own. Just remember that if you are going to run SPEC with a higher sample rate, you should consider that the size of the frequency bins used to process in SPEC is:

bin size in Hz = samplerate/2/FFT Size

So if you double the sample rate, and don't also double the FFT Size your quality may actually go down. You should do your own listening tests to determine what your acceptable trade off in quality vs. cpu-load for SPEC processing is at the higher sample rates.

Some snare drum and high-hat sounds ("noisey signals"), with a long decay envelope seem to be the exception to the rule, and the one thing that Plogue's (and all phase vocoder sythisis/resynthisis) FFT/iFFT bidules don't process transparently (to me they can sound "backwards", in terms of their envelope). I have exactly one data point where upping the sample rate helped with that. However the next album spam the same artist did not benifit at all from upsampling and the intial SPEC tests we did also showed no advantage to upsampling. So, I think it remains as one possbile tool for a particular drum sound problem.

Hope that's clear.

The search goes on for example algorithms/source code (wavelets anyone?) that would be BOTH the next conversion quality improvement AND have a lower cpu-load than Plogue FFT/iFFT.