This was the subject of much discussions on standards for television when the need arose to set speed limits on the audio highways.
After the Vumeters (300ms), peak meters (10ms) and digital meters (0ms), we saw sophisticated and outpriced measuring instruments appear like the Dorrough 40.
YouTube
To reflect on old embittered, 'It was better before" style, I note that the musical quality of the products of the industry of the same name, seems to go in a direction inversely proportional to that of technical progress.
This endless discussion is a fine example of the immense confusion of spirits that seems to be the hallmark of our time. From the start, two camps have opposed each other, arguing deaf on two completely different subjects: The validity of Nyquist theories and the way our auditory perception works.
May I suggest an answer to these two questions:
1- Yes
2- Chose your poison.
The second will give rise to long battles over whether or not to read the labels on the bottles of poison, thanks to evenharmonics ;-)
any reference to a real decibel cut in four would be pure coincidence.
YouTube
Last edited:
I quoted you yourself from an earlier post by means of cut and paste. This is how you address people often. And it certainly applies to you, but I am no historian.
If new tech allows you to run closer to the limits someone will do it. it is not necessarily a good thing. In this digital age many seem to want to cosy up to 0dBFS as much as possible.
As you can see from the results Hans provided of the waveform I said to try, it is very clear I know what I am talking about, so in this case, no. It does not apply.
That said, you have very elegantly put me in my place regarding verbage.
Thank you
John
No, it is not the droop, that would fight amplitude which it clearly did.
It does not pay to simply ignore the contents of a room behind a door simply because you believe it to be empty. Hans already inadvertently missed the frequency shift because it was assumed identical.
W/R to presentation, had he overlayed the two files exactly as Scott did, the change would have screamed out to him. As it were, I almost missed it as well, and I am looking for a difference. The spectra should also be displayed exactly the same as well, the frequency shift again would have screamed.
Not that there will always be a difference, just that it is always important to test and verify.
The carrier frequency in this case, has been altered. Clear, clean, easy to see.
As to the modulating entity, you play semantics. And the form of the modulation envelope is not the concern, only the fact that it causes information in two side bands, one of which is above the breakpoint of the filter.
Since the waveform and the techniques are testable, I hope Scott can take the time to duplicate the modulated signal, as well as an FM run. It would be nice to confirm this on another filter with more or less taps, as it is necessary to confirm that this is not just some trivial software bug.
Ps.. Also of importance, if this carrier frequency shift is verified with a filter of roughly the same number of taps, will a higher tap count do the same, or as it gets larger, will it then have enough information to attack the envelope instead. Recall my concerns with the marginally sampled 22 kHz sine, where it took 880 samples to show the full beat envelope(the fish).
I am taking this to some of the really smart guys at work, it will be interesting.
Jn
Last edited:
Thanks, interesting. Do they ever provide details of the setup? There are always going to be compromises as you are aware. Have you ever been tempted to do what SL did here? AS_creation
"I have used this microphone and a similar setup for my own head, Figure 4, to make recordings of events where I also then had a direct memory of what I heard. This gave me material to evaluate my loudspeaker designs."
Perhaps this would also give a better rendering of accurate than Richard's test?
Did you notice that, during the analog period, we agreed on the average level, and we tried to do the best possible on the crests, without precise limits, and that with digital, we have a concrete ceiling, and we try to do the best we can on the medium level, by cutting off the heads. ;-)
Mr. digitalplus came, sayin: "Eh guys, I offer you 24 meters under the ceiling, but there are cops here who come, with their radar, trying to sticking us a ticket for speeding at more than 16 meters ?
I might but I have a bunch of stuff to do before the weekend. I was thinking of using a Blackman time window.
It's interesting that you say this. I've yet to see much (any non-obvious?) correlation between measurements on the electrical side and perception.
Last edited:
Yes John, but seems to me you have not followed the decades of development on the issues you raised. I'm no good in this matter at all, but sounds like you are reanimating a dead horse. As elektroj pointed out, a few months (years?) ago some manufacturers commissioned him on upgrading the digital filters. The industry is obviously aware that there are some issues on which better solutions have been made available. As Hans pointed out, an upsampled 16/44k distribution is indistiguishable to him to higher rates. I suspect reasonable quality current offerings are very close to SOA already. To be of interest, discussion on improvement should take current state body of knowledge into consideration.
There are three issues.
1. Some here clearly state that they are able to hear an audible difference in the rate, claiming that 44.1 is somehow deficient without understanding why. They indeed focussed on a percussive instrument, the cymbal, as a test vehicle. I am NOT one of those individuals who claim they hear a difference, as I am indeed in the same boat as Hans, I also find CD quality issues indistinguishable.
However, the primary argument fostered against them is one of information derived by the use of steady state sine measurements of all types as well as noise floor.
I cannot speak for anyone else, in that I do not spend any time listening to pure unvarying sine tones for entertainment. My focus is on the tone as well as how the artist modulated those tones to form the music.
2. One of the fundamentals of signal theory is that when a sine is amplitude modulated or frequency modulated, the resultant spectra has sidebands. Since the entire purpose of an artist is to modulate the frequencies, it is important to consider how that modulation affects the spectra. If one chooses a sampling rate that requires filtering which removes the upper sideband, the resultant is not a high fidelity reproduction of the origional.
I have never seen anybody address this concern, rather it has been ignored or dismissed.
3. When advancing the current state, it is always important to consider all the fundamentals. Ignoring any can lead to sub optimal solutions, the CD at 44.1 being one of those sub optimal solutions.
Electroj, understanding what I have brought to the forefront, has an additional tool at his disposal now, another viable argument for maintaining an higher sampling rate as well as an understanding that rate headroom is required because of the nature of the content being recorded, that being music.
jn
Just as Philips wrote off the entire cost of their PASC (later to become MP1) development team via the DCC (Digital Compact Cassette) which was stillborn.
Howie
There is a distinct possibility that I have indeed identified the cause.
I have never seen any paper that does consider the amplitude modulation/sideband issue with respect to 44.1 and real music. If you are aware of such a paper or papers, I would love to read them.
I do not consider raising the rate because people think "something is amiss" is a real engineering approach. Yes, it will certainly raise the margin such that modulation sidebands are no longer an issue, but it is not an elegant approach, just a band aid style fix.
My approach shows that any attempt to bottleneck the data down needs understanding. I can see some saying it's only the D/A side, I can see some saying it's the A/D side, but it is truly a fundamental issue with modulation of audio frequencies and the need to consider the actual bandwidth of the music, not just the range of sines we can hear.
jn
Last edited:
The issue in my mind is not - is the newest CD better or improved over older CD. It is how close to real is the sound of 16/44.1
And, what would be some ways to get closer to real? Keep tweeking 16/44? It seems to me and many others that one way to get audibly closer is a higher sampling rate.
I am wondering out loud if 20KHz hard limit is holding us back.... 24/96 and higher is still made to have a path compatible with CD. That makes sense for commercial reasons. In this point in time with internet streaming etal it may not make any more sense to hold onto CD compatibility.
If we were to start with a fresh sheet of paper, what would we do that would be more realistic in sound? What would that look like?
THx-RNMarsh
And, what would be some ways to get closer to real? Keep tweeking 16/44? It seems to me and many others that one way to get audibly closer is a higher sampling rate.
I am wondering out loud if 20KHz hard limit is holding us back.... 24/96 and higher is still made to have a path compatible with CD. That makes sense for commercial reasons. In this point in time with internet streaming etal it may not make any more sense to hold onto CD compatibility.
If we were to start with a fresh sheet of paper, what would we do that would be more realistic in sound? What would that look like?
THx-RNMarsh
Last edited:
As Hans pointed, upsample the 16/44 material to 24/192 on playback. Perhaps you also want to try multibit DAC as an alternative.
This is precisely why I mentioned the bottleneck two posts ago.
Once a data stream has been sampled at 44.1, any sideband content above the sampling rate is removed by filtering, and is not recoverable.
Upsampling 44.1 source material to 24/192 is putting lipstick on a pig.
jn
No, you are good. Stuck with 973 albums where 785 is completely musical irrelevant.. Happy music life - but it will sound good I'm sure!
;-)
//
- Status
- Not open for further replies.