Originally Posted by arnyk
Love to see Amir pursuing his favorite hobby which may be inventing science. The audibility of jitter (FM distortion) is given in the scientificf literature by Zwicker and Fastl:
30 Hz modulation is far from being the lowest frequency and is also not the least likely to be detected. I chose it for one reason and one reason only which is that it showed up repeatedly in HDMI-related equipment tests, probably because it is a very commonly used video frame rate in A/V media.
I thought we covered this already Arny. 30Hz only showed up in one of my measurements. Here is the graph again:
See, only the Pioneer has a sideband at 30 Hz. You can't run off with that one frequency and say it is "showed up repeatedly." You have other measurements that you are drawing this conclusion from?
The key learning about all of these measurements is that every device is different. What bleeds into the DAC is completely system specific as the above shows.
If we zoom into those measurement with narrower band FFT in Paul Miller's data we see more problems:
Not only is there no 30 Hz component but we have a *ton of jitter* components. The graph you post is at 1 Khz tone, at a specific level of 80 db and a single modulation frequency. That is not at all a model of what is going on here.
Here is another set:
For pete's sake, how can you guys with a straight face defend the graph on the right??? I mean the question is not what jitter it has, but what it doesn't! It has a ton of specific jitter components. Then it has that widening of the skirt due to narrow-band random jitter. Then it has those ripples in there. Everything other than a single spike is distortion (you can see the single spike on the S/PDIF measurements on the left).
Here is yet another profile:
Now we have a series of spikes going pretty high level of -75 dbfs. Their sum total is not -75 dbfs. Energy is additive.
Now, I am not saying the audio experience is this ugly. I am saying we can't trivialize this topic by creating one example "jitter" which only I and one other person ran and say, "oh, here are the results of jitter audibility." Or throw around snippets from a book that are not representative of the measurements we are discussing.
Originally Posted by arnyk
Furthermore I have offered to provide similar test files based on other jitter frequencies if people would agree to ABX them using any reasonable ABX comparator.
Thank you for the kind offer. I have provided the above examples. Please create the files with the completely profile of any of the above AVRs. Not just one frequency but the full profile.