Sample rate conversion

[R.W. Grimes]

I had posted this before, but it got placed in an out-of-the-way location.

I was wondering if there is an issue of degradation by converting from the higher sampling rates mentioned down to the lower 44.1 after filtering?

I really would be interested in trying the higher rates as, like others, I have much difficulty with distortion at usefull tracking settings. But I am concerned with the possibility of causing other types of distortion by re-sampling.

[Craig Maier]

Re: Sample rate conversion

You have three options to downconvert sampling rates. If you choose the best one, it will downconvert with almost no effect on the signal, but will take more time to calculate.

[weskola]

Re: Sample rate conversion

Craig,

Are you referring to the "Conversion Quality" selection (CD, Pro, Master)? If so, what does each one do? Since it does not grey-out when selecting 44.1 as the sampling rate, I'm assuming it applies to 44.1 as well.

[Craig Maier]

Re: Sample rate conversion

Each one uses a progressively more complex interpolation algorithm. We give you the choice because the processor takes more time to perform the calculation with the most complex one, so if you have a slow computer you may opt out for just CD quality. As for the 44.1 question, I am not sure that I understand - - - do you mean converting 44.1 to something lower like 22.05?

[weskola]

Re: Sample rate conversion

I've recorded at 96 Khz. After I've done my processing and I go to resample to 44.1 Khz, I have those 3 options available. Therefore, I assume that I can use the most complex interpolation if I choose to do so.

Which leads me to ask if it is beneficial to record at 96 Khz, resample immediately to 44.1 Khz using the highest quality resample algorithm, and then perform all of my processing.

[Craig Maier]

Re: Sample rate conversion

No. The best thing to do is to record at 96 KHz. Do all your processing at 96 KHz. The last step would be to convert to 44.1 KHz. That way the algorithms can take advantage of the improved bandwidth provided by 96 KHz sampling, especially click rise time.

[weskola]

Re: Sample rate conversion

I found that out at lunchtime today. It'll be 96 Khz all the way until it's time to resample downwards.

[R.W. Grimes]

Re: Sample rate conversion

Craig,

Thanks for the info. So the next project which sounds like it will be a challenge for my filter settings, I shall record at 96k and work from that, and convert to 44.1 at the highest quality. Time is not important, the quality is. But then, maybe I should just sample all my projects at 96k.

[Craig Maier]

Re: Sample rate conversion

I guess if you have the hard drive space, and you do not mind a little more processing time, there is no reason not to use that technique.

[Dan McDonald]

Re: Sample rate conversion

OK - one more question. A friend has DAT files which came from some old open-reel tapes. The open-reels are no longer accessible. We're trying to get the best quality possible out of the end product - in CD format. He originally converted these to wav files at 44.1kHz. Would it make sense to convert at 48kHz to work on them? I believe DAT itself is at 48kHz, so my thinking is that anything above 48kHz would essentially just be redundant information and wouldn't help in any interpolation or processing, but I don't know; I thought maybe in the conversion from DAT to WAV there may be some advantage in going with higher sampling rates.

Any thoughts on this?

Dan

[Craig Maier]

Re: Sample rate conversion

Dan,

My opinion is that if the DATS are copies of reel to reel live recordings, it makes no real difference whether you use 44.1 KHz or maintain the original 48 KHz. However, if the recording is of records having a lot of impulsive noise, the impulse filter(s) will have some benefit from maintaining it in the 48 KHz rate.

[Dan McDonald]

Re: Sample rate conversion

OK -thanks, Craig. It sounds like the chief benefit to higher sampling rates will be in impulse and similar types of filtering schemes. That was what I was getting out of the discussion, but I thought it would be good to ask about the DAT.

Dan