FFT audio analysis flaws?

Orb,
Thank you - I will email Paul Miller - that FFT plot you mention on post 22 is from him.

I only found this so far from PM - http://www.referenceaudio.se/ImageUpload/MusicalFidelity/vlinktest.pdf - a test on the MF V-Link USB-SPDIF converter that references Jim LeSurf's IQ-test & prompts PM to zoom in on his FFT of the DACMAgic. It now shows "My test yielded a discrete & very obvious +/- 0.9Hz, +/- 1.8Hz,etc jitter from a 12KHz signa lat just -30dB. If this were a turntable it would equate to a W&F of just 0.0005% but treated as jitter it comes pout as a massive 1680nS"

So I guess, PM is confirming the same results & jitter calculations as Jim LeSurf did for the DACMagic

As an aside, one of the reasons this jitter is not suppressed by the DACMagic's ASRC (& most other ASRC's) is probably because of it's cut-off bandwidth i.e any jitter below about 3Hz gets through un-attenuated. It might be one of the reasons why lower jitter sources still sound better than higher jitter sources into DACs that use ASRC on their inputs?

Thanks for the info on the new AR DAC measurements.

I note also that Demian Martin from Product Design Services also uses a very high resolution FFT to analyse exactly this close-in noise or modulation - 4M or 16M point FFTs with 10 averages. So it can possibly be done with FFTs but it takes a bit of time to run & is not often done.

Orb said:

Definitely an interesting subject and thanks for a great thread.

Click to expand...

Thanks Orb
Do you have a link to Paul Miller's follow up letters with Jim - I can't find them?

Yes I think you are right regarding cut-off bandwidth, which may be due to limitation of the DAC chip; was one surmisation this was a potential cause.

Click to expand...

What I meant was that all ASRC's have a corner frequency below which they do not attenuate any jitter, they just pass it through. The lowest corner frequency I've seen stated for ASRC's is 3Hz so any jitter below this is untouched & gets through to the DAC.