What's the use of a 192 kHz sample rate?
in [DSP]
|
Prev: Is 44.1 KHz sample-rate enough to cover the entire human hearingrange?
Next: Newbie to DSP - Phase shift detection
From: Don Pearce on 4 May 2008 15:40 On Sun, 4 May 2008 12:28:34 -0700 (PDT), rickman <gnuarm(a)gmail.com> wrote: >On May 4, 12:18 pm, DigitalSignal <digitalsignal...(a)yahoo.com> wrote: >> Hi rickman, can you refer the company who can produce or analyze the >> signals with 140 dB SNR? I have never seen one. I guess they are not >> using any digital technology, or they are look at very low frequency >> signals. Right? >> >> Jameswww.go-ci.com > >Thales Communications, makers of military radios. This figure came up >in the context of digital interference with the RF sections. I was >told that the digital noise had to be below -140 dB to prevent >desensitization of the receiver. These units are *very* overbuilt and >far surpass anything commercial I have seen. One spur anywhere in a >huge range, 30 MHz to 500 MHz, IIRC and they are back to the testing >room to add more copper tape or to put more resistors in clock lines, >etc. A desensitization level is going to be measured in dBm, and -140 is a very ordinary figure. If external noise interference is present at the input of a receiver, it will add to the inherent noise already there, raising it. This is called desensitization. A common interference requirement is that a desensitization of no more than 1dB is permitted. The -140 in this instance is just a receiver input level, and nothing to do with S/N ratios, dynamic ranges or anything like that. I would be very surprised if anything military surpassed the spec of a commercial design. Far too few units are built to justify the number of trips round the design cycle that are needed to remove every spur from and maximise performance. That fact that they have to bodge units with copper tape to make them work rather bears this out. Certainly when I was designing low noise converters for domestic satellite systems I achieved a guaranteed total noise figure 0f 0.25dB at 12GHz, and the entire unit had a works cost price of 11 dollars. Nothing military has ever come close to that kind of performance or price. d -- Pearce Consulting http://www.pearce.uk.com
From: Arny Krueger on 4 May 2008 16:22 "robert bristow-johnson" <rbj(a)audioimagination.com> wrote in message news:4f0de839-fa35-4c51-93c8-751c12118f79(a)x35g2000hsb.googlegroups.com > now, long ago, when DSD was just coming out, i > participated in something like this that an audio guy > (now an author) named Bob Katz did. this was before Pro > Tools HD and they were using some expensive system from a > company called Sonic Solutions that could do 192 kHz. > they recorded at 192 kHz some test sounds (including some > high frequency percussive sounds like castenets, cabasas, > cymbals) along with synthesized bandlimited (to 96 kHz) > tones of all sorts of waveshapes at a variety of > frequencies from below 10 kHz up to 90 kHz. i don't > remember all of the test sounds, but they made sure that > most, if not all, had content well above 30 kHz. the > subject listened to those sounds in two different forms, > but both with a 192 kHz playback rate. one form was the > raw recorded sound, the other was processed through a > phase-linear FIR filter with a lot of taps (i thought it > was around 300 something taps, it was not a real-time FIR > filter but proceessed one sound file into another) that > was flat to within 0.01 dB up to 20 kHz, had a smooth > transition band from 20 to 22 kHz, and then was down by > more than 130 dB for 22 kHz to 96 kHz. we padded the > beginning an end of the original sound file with zeros, > of half the length of the FIR on both ends and the > non-real-time FIR was not "causal" and had a delay of 0 > (so it started responding before the first non-zero > samples) and the filtered sound file was lined up in time > with the original. > now this isn't what Bob Katz did, but it is what i wished > he did: called the original sound "A", and the LPFed > sound "B". so sometimes the subject hears AA, sometimes > AB, sometimes BA, and sometimes BB (all four permutations > exist in equal quantity) and for each pair, the subject > simply has to say if they think the sounds are the same > or different. every time a subject says that AA or BB > sound different, we count that as a false positive and > subtract that count from the number of times the subject > says that AB or BA sound different. likewise for false > negatives. That is called same/difference testing. It is an older form of testing, valid enough as far as it goes. The most important question in a test is not the order of presentation of variables (within reason) but rather whether the listener can determine the identity of the so-called unknowns by any means other than listening. Obviously, if one can determine the the identity of the unknows by any means other than listening, the unknowns are not unknown but known, and there is no test. > now Bob didn't do that, but he did something like it (i > think it was ABX) and there was no statistically measured > difference. At this point 100's if not 1,000s or perhaps even 10,000s of people have done this kind of test, and they obtained the same results as Katz. Please see www.pcabx.com. > people could simply not reliably tell if the > stuff above 20 kHz was removed or not. they could not > tell at all. no one could. Welcome to the well-known limitations of the human ear. > now, if they cannot tell if the content existed above 22 > kHz or not, is there a need to have it there in storage > or in transmission? No. > if there is no need to have it > there, and it is removed, what does the sampling theorem > tell us regarding sufficient sampling rate? It tells us that a well-implemented 44 KHz sample rate might even be overkill. >> I'm not trying to "prove" anything. I am presenting >> information which you can consider and believe or can >> ignore. But you can't say my statements are false unless >> you have some information to "prove" they are. Requiring absolute proof can be a sort of excluded-middle argument. IOW there are a great many things that have not been absolutely proven false, that are generally and correctly treated as being false. In fact, many many negative hypotheses, such as the idea that anything in particular is false, are difficult or impossible to prove. Just because I can't absolutely prove that pigs can't fly is not a good reason to start an airline based on flying pigs. ;-) >> Human hearing is not a microphone connected to an amplifier. Often human hearing is far less sensitive than a microphone connected to an amplifier. So what? >> It is a very complex process which even includes the >> brain and we certainly don't understand it completely. We don't have to understand everything completely in order to come up with effective working hypothesis about them. > but we can't hear anything above 20 kHz. Actually, often many of us can hear something above 20 KHz if it is an isolated tone and also loud enough. The question at hand is about something very, very different - whether or not we can hear the elimination of information above 20 KHz from music. > even with > percussive sounds with sharp attacks. and if we cannot > hear anything above 20 kHz, then > 40.0001 kHz sampling rate can store all of the > information we need. for practical reconstruction > purposes, 44.1 and 48 kHz are sufficient. That turns out to be a hypothesis whose contradiction is difficult or impossible to prove. > now, in *processing* sounds with some nasty > non-linearities in the process, it very well may be > necessary to upsample to 192 kHz or higher to do that > non-linear processing, and when it is done, LPF to 20 kHz > and downsample back to 48 kHz. but, except for > experimental purposes, 192 kHz storage or transimssion is > not necessary. Agreed.
From: Arny Krueger on 4 May 2008 16:29 "rickman" <gnuarm(a)gmail.com> wrote in message news:96f46e02-0219-4e88-81de-9a3e15579e8e(a)a70g2000hsh.googlegroups.com > I never said that *anyone* can hear sounds above 20 kHz. But I'm willing to agree that some people can do this, even in a blind test. However, its not the question at hand. > I said that there are people who can hear the difference > between systems that use higher sample rates than 48 kHz. I'm willing to agree that this can happen, but the explanations always seem to have this annoying tendency to not be the same as the question at hand. Just because a system uses a sample rate >= 48 kHz does not mean that it is free of audible flaws. The question at hand is whether or not introducing an ideal low pass filter is audible when listening to music, and when that low pass filter has negligable effects below 20 kHz. For example, a - 6 dB/octave filter with a - 3dB point at 20 kHz also causes 1 dB loss at 10 KHz, which some listeners can detect. > The real issue is about the electronics and how they are > measured *compared* to human hearing. Distortion is the > real issue. Filters introduce distortion and some people > can hear the distortion produced by converters running at 48 kHz. There are basically only two types of distortion - linear distortion and nonlinear distortion. Since you have not restricted your comment to nonlinear distortion, there is a possiblity that you are talking about nonlinear distortion, which is again not the question at hand. > I suppose it is possible that my friend was mislead (and > therefor myself). Actually, it is for all known practical purposes, a certainty. > I have no first hand knowledge of the > sound tests. Many of us do. Please see www.pcabx.com . > But he is not a person to believe in rubbish. Well, up until this question... ;-) > In fact, he is a person who is very intellectually critical. Apparently, not critical and well-informed enough. > This is one of the few times he > could not dismiss claims that run counter to the science. That's his problem.
From: robert bristow-johnson on 4 May 2008 17:03 On May 4, 4:22 pm, "Arny Krueger" <ar...(a)hotpop.com> wrote: > "robert bristow-johnson" <r...(a)audioimagination.com> wrote > in messagenews:4f0de839-fa35-4c51-93c8-751c12118f79(a)x35g2000hsb.googlegroups.com > ... > > now this isn't what Bob Katz did, but it is what i wished > > he did: called the original sound "A", and the LPFed > > sound "B". so sometimes the subject hears AA, sometimes > > AB, sometimes BA, and sometimes BB (all four permutations > > exist in equal quantity) and for each pair, the subject > > simply has to say if they think the sounds are the same > > or different. every time a subject says that AA or BB > > sound different, we count that as a false positive and > > subtract that count from the number of times the subject > > says that AB or BA sound different. likewise for false > > negatives. > > That is called same/difference testing. thanks, i'll use that term in the future. > It is an older form of testing, > valid enough as far as it goes. The most important question in a test is not > the order of presentation of variables (within reason) but rather whether > the listener can determine the identity of the so-called unknowns by any > means other than listening. that's what blind testing is about. > > > now Bob didn't do that, but he did something like it (i > > think it was ABX) and there was no statistically measured > > difference. > > At this point 100's if not 1,000s or perhaps even 10,000s of people have > done this kind of test, and they obtained the same results as Katz. Please > see www.pcabx.com. i have never completely accepted that ABX testing is better than "same/ difference testing" (what i used to call "AB testing") for the question we are trying to answer here. ABX is for answering the question: which is better? or which is closer? does synthesizer A sound more like a real piano or does synthesizer B sound more like the real piano? even though i am politically very liberal, as an engineer, i am *very* conservative. say the issue is Monster Cable. the claim is that Monster Cable sounds better than 10 gauge lamp cable. now, being conservative, i am willing to grant Monster Cable the benefit of doubt in that if people can actually hear a *difference*, i'm willing to let their claim that they sound better go without challenge. but if people *cannot* hear a difference, then i am not willing to accept, even tentatively, that Monster Cable sounds better. for it to sound better, it has to at the very least, sound different. since "better" and "worse" is a subjective thing (like clean vinyl vs. CD), i am too conservative to take a position regarding which is "better" if they're different. but i am not too conservative to say "bunk" about something being experienced as "better" if it is not experienced as different. "same/difference testing" does not cloud the issue with an ancillary question of "which one of these two sound closer to a third?" it only asks the question "do these sound discernably different or do they not?" and with same/difference testing, it is so easy to unbias the measure so we can even invite a Monster Cable salesperson to the test and let them know that if they guess with a bias that these sounds *do* sound different (which would support their case, but not fairly), we will know it (because we subtract the false positives from the true positives). i think for this kind of question, the Monster Cable sounds better or ultra-high sampling rates sound better or whatever, same/difference testing speaks to that question directly whereas ABX is not as direct. r b-j
From: Greg Berchin on 4 May 2008 18:03
On Sun, 4 May 2008 14:03:12 -0700 (PDT), robert bristow-johnson <rbj(a)audioimagination.com> wrote: >ABX is for answering the question: which is better? or which is >closer? does synthesizer A sound more like a real piano or does >synthesizer B sound more like the real piano? I could be wrong, but I understood that a proper ABX test consists of A, B, and either X==A or X==B. It is the job of the subject to determine whether X is identical to A or X is identical to B. This is very different from determining whether X sounds "more like" A or "more like" B. I also believe that a more effective test format would be ABXI, where "I" means "indeterminate", for those cases in which the subject willingly admits that they cannot make the determination. Greg |