Seems to me both sides just like to argue a lot. And objective test like a null test to settle the matter is not done.
Amen.
Point by point, in the Gearlutz summary, Lavry addresses each and every "issue" raised here over the last couple of weeks. I know a few around that could use a good reading, as much as I am sure I hear the grumbles of "no proof", "our hearing", "cymbals", etc... And after all, who the **** is Lavry to come against our beliefs?
No joke, paranormal abilities do not interest me. The other side that Lavry represents claimed no advantage of higher SR also showed no measurement done on real devices available in the market either. As I said, I strongly suspect both sides prefers the endless repeating debates. Perhaps the participants feel the thread is more alive that way.
I guess I could. The flatest high quality mic I have is a hyper-cardioid condenser. Probably won't sound great on a cymbal since the sound from one comes from all over the surface not just one spot. Well... unless maybe I experiment with distance and see what I can do with that. It might cause some HF loss...
Also, the cleanest preamps I have are not beautiful sounding. Suspect the files might be good for data, but not so much for triggering to use for making music.
What sample rate are you interested in?
Set aside your statement about measurements is flat wrong, here we go, exactly as predicted. I suspect it would make sense for you to also require experimental proof about the light speed limit. 100 years of theoretical analysis and application development based on it doesn't count, right?
Anyway, is there anything revealing you would expect from such measurements? And, after all, who owns the burden of proof? Why would anybody bother to provide measurements of what confirmed theory clearly shows it's impossible?
There's a bigger fish to fry out there, better ways to spend your money and time. Unless there's an audio business involved, of course.
If you follow the cymbal link from a few days ago, he has made a bunch of cymbal downloads available on one of the other pages there:
Cymbal Reviews with Spectral Analysis | Drummerworld Forum
here is the other link:
Welcome Secrets of Home Theater and High Fidelity
Last edited:
Thank you George.
Lavry, while smart and technically savvy, is very sloppy as a technical writer.
The gear sluts post is of no use. All he does is make blanket assertions.
The over sampling post, on the other hand, is far far superior in terms of content, but he still misses significant points.
(Saving, be right back).
On page 1 he shows NRZ. Note that makes no mention that a significant amount of energy is added to the signal as a delayed entity.
On page 2, he shows the magnified frequency plot of the NRZ data, and specically points out the hf roll off as being unacceptable for hi fi...and then, what????
On page 3, he shows NRZ 2x over sampling spectra, but neglects to explain why the images around 88 kHz are sloped while the primary below 22 are flat.. I want to know.
On page 5, he details interpolation. While he is a bit messy in explanation here with linear, the 2x zero point NRZ method is clear, the images go out to 88khz centered again, and as he explains, the filtering needs drop significantly.
On page 3, he does state that phase linearity is "drastically improved" as only a four pole filter is now needed. That statement should have been repeated directly after showing the 2x zero point NRZ results which mimic exactly the twice sampling rate data.
Throughout all his writings, he makes no mention of the linearity aspects of using NRZ to increase the signal energy. This is an assumption that I have not seen proven true.
The shorter the sample value is held, the more relaxed the filtering can be, albeit at reduced signal energy. What I would have immediately asked him is...if the NRZ is held for 1/3 of the data width, do the images immediately center around 132 KHz, 1/4 go to 176 kHz ?
Jn
Last edited:
Syn08, probably my english is rusty but I can not grasp what you are saying. What did I say about measurement that is incorrect?
I mentioned measurement in relation to the claim of "no advantage of higher SR" on real devices available in the market, Lavry only show graphs made assuming an ideal performing device. In my current setup I prefer to upsample 16/44k material to 24/96k on playback in foobar and suspect the difference to imperfection of my soundcard. I also hear an improvement of reproduction of the same material over the years on subsequent soundcards that I use.
I would expect less endless noisy debates to go on and more meaningful discussion here.
My rusty english interpret this statement as "theory says it is impossible to have an imperfect working DAC", which is obviously not what you are saying.
My understanding is that Lavry separates the issue of establishing an agreed upon (*) audible bandwidth from the issue of absolutely adequate (**) sampling rate for that bandwidth.
(*) my question is, except of cymbals and triangle bells (with their non harmonic overtones), what other musical instruments have the chance to project their output beyond 20kHz?
(**) Kotelnikov (1933)
http://ict.open.ac.uk/classics/1.pdf
(**) Shannon (1948)
https://web.archive.org/web/20100208112344/http://www.stanford.edu/class/ee104/shannonpaper.pdf
George
(*) my question is, except of cymbals and triangle bells (with their non harmonic overtones), what other musical instruments have the chance to project their output beyond 20kHz?
(**) Kotelnikov (1933)
http://ict.open.ac.uk/classics/1.pdf
(**) Shannon (1948)
https://web.archive.org/web/20100208112344/http://www.stanford.edu/class/ee104/shannonpaper.pdf
George
You claim that no measurements were done on real devices available in the market. In general, this is so wrong it even doesn't require quotes.
I have no idea what measurements could be shown to prove "no advantage of higher SR". Ever heard about the impossibility to prove a negative, and about the burden of proof rule?
And here we go again, "upsampling the CD format changes the sound", do you really believe upsampling brings more "information" to the original material? Is there a Maxwell daemon in your system that you would like to talk about?
And yes, your English could be rusty, that's not what I said. An imperfect DAC does not allow one to ignore the theory.
Real answer is: all of them. This comes down to what really maters for the SQ and where to draw the line, since a compromise is unavoidable. Next time we'll hear somebody claiming that sampling @1GHz is required since the HF content of musing creates intermodulation products that could be audible, while others will say 2GHz is better.
Then, how far higher than 20kHz in your opinion based on existing spectra from instrument’s output (not from vinyl playback recordings). Serious question.
George
As well you should.
A good publication should either stand on it's own, or reference suitable information.
I asked questions of his writings that should have been within the writing, but we're not.
When I referee any articles for publication, these are the type of comments I will have returned to the author for revision.
For high level physics articles, I sometimes have to rely solely on checking for internal consistency, as the topic may at points fall outside my wheelhouse. When I am asked to referee such articles, I make it clear up front if the topic is outside my domain, but at times there are really very few if any qualified experts in the world capable of critiquing the article, the journal editors must have referees.
For superconducting magnet manufacturing articles, THAT I live for. Sometimes it is a veritable feeding frenzy. High level motion control articles, sometimes I am reduced to sobbing. There is such a long ways to go in that discipline.
Jn
Ps. I note with more than casual humor, you did not even try to address the points I made.
Last edited:
Benchmark DAC-3 upsamples all PCM to 211kHz since they say it provides a more favorable oversampling ratio. Their dac measures as well or better than Lavry's, AFAIK.
Last edited:
Welcome Secrets of Home Theater and High Fidelity
Question: How many of you have time aligned speakers ?
(On the KEF 50 wireless, as there is no way to align the two ways by adjusting their relative distances to the listener, because the coaxial construction of the speakers, they use an optional digital delay.)
There is no use, on my point of view, to evaluate by listening any transient difference without this prerequisite.
- Status
- Not open for further replies.