"hearing" may be the only sense that matters if you listen through headphones, but there is a large amount of experimental evidence that humans "sense" frequencies much higher than they hear, certainly to 30 kHz and possibly as high as 100 kHz. So the question shouldn't be what can people "hear" but rather what is the frequency range of what we want to listen to. Further, I think it's likely that the filtering in both A>D and D>A is what causes at least some people to have problems with much digitally recorded and played music; I doubt we really have a good handle on what is really needed in good digital filter design. Just a simple example? We don't currently have a DAC playback filter with both good impulse response and flat frequency response; you can have one but not both (well, maybe almost both if you have a dcs Vivaldi).
Is the dynamic range of CD sufficient?
- Thread starter Al M.
- Start date
How important is impulse response? The filters on the AD filtered it out. You can generate test signals that way, but the music recorded through an AD doesn't have any impulses that cause issues for a DAC.
I don't know of credible evidence people sense frequencies to 30 khz. There is some they sense things around 100-120 khz. Mostly when such sound enters through your eye sockets and stimulate nerves in your skull. Do we need response to 120 khz now? We don't have many mic's and effectively no speakers to reproduce it. Radar at 6 ghz can cause phantom sounds too, but I don't think we need 6 ghz digital recordings. The digital filters at the output of a DAC are mostly a goose chase of nothing in my opinion.
By "experimental evidence" I assume you mean evidence that is the result of experimentation. Got any links? I'd love to take a look at the evidence, and the methodologies and numbers of the experiments, because there is a huge body of evidence from decades worth of repeated, varied, verified and reviewed studies indicating that humans are don't sense, or respond to auditory stimulus outside of the audible spectrum at all. A comparison might be in order.
Tim
http://en.wikipedia.org/wiki/Hypersonic_effect
Basically the Oohashi stuff. Others have been unsuccessful duplicating the results he got for hearing to 30 khz. I used to have some links for the business about 100-120 khz. People reliably identified it as different in DBT's. I forget who figured out it was through eye sockets.
I don't understand this at all. all sound , music especially, is made up of impulses, and it's an important part both of musical enjoyment and the differences between live and recorded sound (in any medium).
Amongst the many things one can count on in audiophile blogs is that about once every 6 months someone will trot out the tired Oohashi *test*.
To an engineer an impulse is a pulse with zero width and infinite frequency content. Much different than what a musician or listener may mean by the term "impulse". Fourier says music is made up of a series of sine waves of varying amplitudes and frequencies, not impulses. Impulse to an engineer is not the same as "transient response", "attack", "decay", etc.
I tend to agree, but FWIWFM I use resolution to refer to bit depth, and sampling rate to refer to bit rate. Perfect (let's say "near-perfect") implementation implies a lot of care in both. Filter artifacts are usually associated with bit rate instead of bit depth, though there are of course plenty of frequency and time domain filter artifacts.
Semantics are a pain. "Noise floor" to me means just that, but I certainly agree that reverberation and reflections can smear or muddy the sound and degrade precise imaging. Most people prefer the more "live" sound they add, but I tend to prefer a deader room that exhibits the image in the recording with less room influence. Perhaps partly because I have found over the years that the room that sounds best for small groups does not sound good for large orchestras, and so forth. A matter of taste.
<This is all speculative opinion, though I believe it based upon my experience making, recording, mastering, and listening to music for the past few decades.>
Well, thinking on it, the music is probably the biggest limitation in dynamic range. Not a lot of music has dynamic range of even 80 dB, though I suppose most gets there if you consider the silence between songs.
The mix is usually designed to work over a wide range of systems so the music sounds good to the most people. That, IMO, typically results in dynamic range constrained to suit everything from cars to MP3's to average audiophile systems. That has also given rise to audiophile labels producing recordings with wider dynamic range, natch.
As for me, probably because I'm getting older, I have a harder time appreciating movies that go from louder than I can stand to so soft I can't make out the words...
DonH50 pretty much covered this.
What you are calling impulses in music aren't causing the ringing you see in magazine tests. And anything that would activate that ringing on the playback end has already been filtered out. So having a non-ringing output filter that also alters frequency response is getting you nowhere. For that matter I can show you an EQ curve you can apply to your Redbook CD's that will get rid of that ringing were it to exist. It will sound different as it alters the FR. Changed FR sounding different isn't exactly breaking news or shouldn't be.
Last edited:
I use a Wyred4sound class D amp. It is spec'd for 80 uvolts out quiet, but I have actually measured it as more like 50 uvolts. I also have a Tact digital amp which is around 35-40 microvolts. Now if I was using efficient speakers that might matter, but my Soundlabs are anything other than efficient.
Your point is one I think lost on many. To achieve 20 bit performance overall is perhaps possible though it is likely few at any price have such a setup in their home. Most good systems achieve maybe 18 bit performance. And only that electrically. Figuring in noise levels most don't even get that.
I don't know if my room is actually more "dead" sounding than prior to room treatment -- all my tube traps are set towards the diffusive side rather than the absorbent one, and the sound panels are diffusers, not absorbers. Interestingly, while in the hand clapping test all the room echo is gone and consequently conversations in the room sound drier, the music coming from the system sounds acoustically much more alive than before. The extra liveliness comes from letting all the ambience and air of the recorded venues come through, which was completely suppressed before by all the room reverberation and reflections; that had also resulted in a flat, congested soundstage without any depth. Both for small groups and large orchestras everything sounds more alive.
Last edited:
My point. We nave nearly 100 years of audiology in which the same result has been repeated over and over again -- feed them a signal above 20khz and humans do not respond at all. It may as well not be there.
Tim
I should have noted in my earlier post about dynamic range that, while many microphones specify a dynamic range of 140+ dB, the majority (all I saw in a quick look) include pads to reach the highest SPL's. Their actual SNR is ~80 dB, commensurate with the dynamic range of everything else. At least for the few mics I am familiar with and checked around the $1k range (Neumann, AKG, Rhode, Earthworks, etc.)
Except those experiments (often quoted as "evidence" for audibiliity of > 20 kHz signals) where the HF signals causes intermodulation distortion in amps and speakers that includes sub-20kHz components.
A few quick notes:
1. Deciding what channel to use to transmit/deliver is not the same as what one hears. The channel needs to be shown objectively to be transparent. To do that, it needs to have good bit of headroom above what a normal population survey would indicate.
2. We can't control what "music" is. Its spectrum can be whatever it wants to be. Per #1, we need to achieve transparency for all cases if that is our stated objective.
3. Dynamic range of mic, etc. does not set the content dynamic range since I can mix multiple elements, add effects, etc and with it, create louder or fainter notes.
4. There is almost no economic need to shrink the resolution. The people who want high resolution can afford the bandwidth and storage easily. In that sense, I see no need to justify > CD quality as being necessary. If the source is > CD, let's have it
.
1. Deciding what channel to use to transmit/deliver is not the same as what one hears. The channel needs to be shown objectively to be transparent. To do that, it needs to have good bit of headroom above what a normal population survey would indicate.
2. We can't control what "music" is. Its spectrum can be whatever it wants to be. Per #1, we need to achieve transparency for all cases if that is our stated objective.
3. Dynamic range of mic, etc. does not set the content dynamic range since I can mix multiple elements, add effects, etc and with it, create louder or fainter notes.
4. There is almost no economic need to shrink the resolution. The people who want high resolution can afford the bandwidth and storage easily. In that sense, I see no need to justify > CD quality as being necessary. If the source is > CD, let's have it
.
A few quick notes:
1. Deciding what channel to use to transmit/deliver is not the same as what one hears. The channel needs to be shown objectively to be transparent. To do that, it needs to have good bit of headroom above what a normal population survey would indicate.
2. We can't control what "music" is. Its spectrum can be whatever it wants to be. Per #1, we need to achieve transparency for all cases if that is our stated objective.
3. Dynamic range of mic, etc. does not set the content dynamic range since I can mix multiple elements, add effects, etc and with it, create louder or fainter notes.
4. There is almost no economic need to shrink the resolution. The people who want high resolution can afford the bandwidth and storage easily. In that sense, I see no need to justify > CD quality as being necessary. If the source is > CD, let's have it
!. Try this one again. Talk real slow.
2. We don't control it, no, but it's not hard to make a case that it's not music if it's not audible. Why does a spectrum outside of our range of hearing have to be transparent?
3. OK
4. It's not about economics. We have no good reason to believe there is an advantage in reproducing supersonic content. We do have good reason to believe there is a disadvantage in reproducing it; IMD.
Tim
Similar threads
