I think higher frequency sounds, including noise, also can contribute feelings of spaciousness and envelopment. Griesinger has interesting discussion in this paper
http://www.davidgriesinger.com/spac4.pdf
see page 3
Last edited:
So, with respect to LP sound, yes LF noise could well intensify feelings of spaciousness or "naturalness" as it augments the perception of both foreground and background streams.
What you mention might all be correct, but in no way does it explain that after digitizing the (originally analog recorded) LP sound and converting this back to analog, both steps being performed with 24/192, the perceived spaciousness and naturalness are negatively effected? See postings 975 and 989.
In that way I fully agree with Pano, we are just plowing ahead and continue the tired old debates about why vinyl sounds so good.
Hans
Last edited:
Let just say that the needle is a musical instrument.
it vibrates to the groove shape and that makes your whole music shaking with harmonics with no boundaries on the complexity of the resulting sound which fortunately for us our imperfect ears likes.
I was told that dollars for dollars a turntable will sound better than the digital processors.
it vibrates to the groove shape and that makes your whole music shaking with harmonics with no boundaries on the complexity of the resulting sound which fortunately for us our imperfect ears likes.
I was told that dollars for dollars a turntable will sound better than the digital processors.
There is an error in the conversion A/D->D/A in the example you have?
Or, levels not matched?
It's a puzzle, I agree.
There is absolutely no error in the A/D-D/A conversion and levels are perfectly matched. I'm using an expensive Bel Canto DAC 3.5VB with a matched integrated A/D converter.
And to repeat what was already mentioned, sound is almost exactly like from the stereoplay CD, that was directly recorded from the same LP, giving thus an extra confirmation from an independent source..
Picture below represents a 1kHz square wave in red entered at the A/D and captured again in blue after D/A.
What you see is the ringing before and after the leading pulses, caused by the steep digital filtering.
The frequency of this ringing is at 96Khz, while conversion up to digital and down to analogue is at 24/192 kHz.
Doing the same with sinus shaped signals shows a very very low distortion.
Hans
I've re-read your posts and I think you're right. The a/d-d/a system has filtered out some part of the original LP analogue signal.
So, empirically, what was filtered away?
So, empirically, what was filtered away?
Filtered away or added for that matter.
Although the pre and post-ringing happens at a supersonic frequency, many sound elements are mingled together through this phenomenon, possible creating a new layer of micro dynamics.
Some might think that DSD would offer better performance.
Read this to get more insight in the matter.
https://benchmarkmedia.com/blogs/ap...-dsd-provides-a-direct-stream-from-a-d-to-d-a
And indeed, a very extensive test at the Detmold Music University in Germany performed with top class equipment, could not discern between 24/176.4 kHz and DSD, confirming that DSD has no benefit over PCM.
http://old.hfm-detmold.de/eti/projekte/diplomarbeiten/dsdvspcm/aes_paper_6086.pdf
So to conclude, at this very moment all I can think off is that the steep filtering in a/d and d/a processing is affecting the sound in the audio band. Nyquist has proven that sampling does not remove information, as long as 2 times the highest frequency in the signal is below the sample rate, so the only thing that remains as a possible cause is filtering.
I do not have a DAC with different filter options, but this could well be an interesting addition to the test.
Hans
What do you mean by "mingled"? What do you mean by "creating"? Are you saying that the filter is non-linear? If so, that need fixing. If not, what are you saying? Put several signals through a linear filter and, whatever the filter frequency response (and hence time domain response), they will not come out 'mingled'.Hans Polak said:
I'm sure you are aware (but others might not be) that a filter cannot be 'brick wall' in both time and frequency domains; you have to accept wiggles in one of them, so most people choose brick-wall in frequency and put up with wiggles in time. These wiggles are more a consequence of missing frequencies rather than anything added to the original signal. However, it is easy to frighten newbies by playing a square wave through such a filter. Note that the time-wiggles you see coming out of a good DAC fed a genuine signal from a digital source are the result of the anti-aliasing filter in the original source; it would be an error to eliminate them, as then you are no longer reproducing the signal at the output of the anti-aliasing filter - which is what digital is supposed to be able to do.
A question from my side first: Do you mean a linear phase filter with linear filter ?
Not all Fir filters used in Dac's are linear phase filters.
And with mingling I mean that all (unnatural) pre-ringing products are added together, in this specific case all having the same ringing frequency of 96kHz (so this is a non linear process), which might possibly result in the creation of amplitude modulation of this 96kHz and IM products folding back in the audio range.
I am all familiar with this.
And I fully agree that filters in the analogue source line before digitizing can also produce ringing, but when sampling is done at a frequency high enough, a very conservative analogue filter can do the job not having any pre ringing.
Only in case of some digital filters, pre ringing occurs, which I think may add something to the signal that is unnatural.
I don't understand what you mean with removing the Wiggles as you call them from the original source, I wouldn't know how to do this and as just mentioned I see no need for such an operation.
I understand even less what this has to with the reconstruction of the signal in the DAC, because quite different processes are taking place here like upsampling, the synchronisation of different frequencies all to the same frequency before converting it to analogue and and removing the Sinc amplitude distortion to name a few.
The reason that I posted the square wave was to prove 3 things.
1) The A/D and D/A are working properly
2) They are perfectly matched, and
3) A/D and D/A are really taking place at 192kHz.
In my experiment I used an LP completely recorded in analogue technique, having the cleanest possible source without any digital by-products.
I appreciate all your questions, and it is vital to check and check over again, but in this case we have only two simple complementary operations added in series with the analogue signal, resulting in sampling and filtering.
To my opinion sampling is a transparent process when preventing frequencies above 1/2 sample rate, and having enough signal to keep the LSB's busy, which is definitely the case when recording the output of an LP with 24 bits.
Hans
I presume we are still trying to prove digital CD or HD based systems are inferior to the good old LP....
If your question is because we seem to be losing ourselves in technicalities, the answer is still a firm yes, although "proving" should be substituted by "finding a reason why we prefer".
And from the above postings it might be clear that I tend to blame the pre-ringing from digital filtering as the most likely cause.
In real life no instrument will ever start generating sound from itself before it is touched.
Hans
I am not.
I am assuming Hans's system and room give a high quality listening experience and he hears what he says he hears. I've read his posts on other threads and think my assumptions are reasonable, given the quality of his contributions elsewhere.
What interests me is that his
L/P->A/D->D/A->AMP does not sound the same as L/P-> AMP, because it should. That it does not puzzles me.
Hans I think the pre-ringing has to be seen in context. It is not a-causal. The main signal peak is delayed wrt the input signal; the pre-ringing starts as soon as the signal arrives and the peak (or edge in case of square wave) starts a bit later. But nothing happens before the signal arrives at the input.
You probably are fully aware of this but the term 'pre ringing' might put some people on the wrong foot. It is 'pre' the signal peak or edge but it is not 'pre' the input signal.
Jan
BTW I attended Doug Self's presentation of the Devinylizer this afternoon at the AES convention, and it was well received with quite some interest from other 'analog aficionados' .
Jan
Jan
Jan,
There are so called apodizing filters, where the ringing follows the transient, with no pre-ringing at all like most Fir filters do.
I have no such device at my disposal, so I cannot compare.
But it would be a very interesting test.
Hans
- Status
- Not open for further replies.