P
pianodano
Guest
Wrong! The thread title (and I suspect it's intent) asks why one and not the other.
I had been thinking a lot about that myself. But at the time I replied, I didn't feel like going back to page one and reading the original posters 1st post.
mshilarious
Banned
OK, this is kinda like eating your cake and having it. You can't accuse me of relying on measurement without also acknowledging that these supposed flaws are also based on measurement of the same data.
The visual representation of a 10kHz sine wave at 44.1kHz looks like a square wave, that was argument #1 (since set aside by some). Why is the visual representation of the time domain accurate, but the visual representation of the frequency domain inaccurate? It's the same data! Does a 10kHz sine wave sample sound like a square wave? Does it contain any 20kHz harmonic when represented in the frequency domain, or played back through a D/A? No. Does it contain any ultrasonic harmonics when played back through a D/A? Yes, but as I have measured, at less than -110dB relative to the fundamental. Can you hear a -110dB harmonic with a full scale fundamental? You can answer that for yourself using much lower frequencies, to make the test easier. But that still isn't a fair representation of just how low such distortions sound, because none of them are audible spectrum.
If we decide we truly need ultrasound, say to 40kHz, you do need to set your converters to 88.2 or 96kHz. You also need to give some hard thought to your microphones and your tweeters. Pretty much, ribbon tweeters will be mandatory--metal domes have ultrasonic resonant frequencies; fabric domes have flatter ultrasonic response, but with poor transient accuracy. I've been thinking about adding ribbon tweeters myself, but I'd have to change my crossovers and add a midrange too, so I've put it off . . .
Your comments about Heisenberg apply equally to digital and analog recording, by the way. Analog recording is a continuous measurement so far as human beings are concerned, but at the subatomic level, it has to involve individual electrons. So if playing back a tape is a measurement in the sense of quantum mechanics, it is equally unreliable.
The good news for the h8rs out there is I have completed my "real world" test of quantization distortion using my organ pipes, and I will post those files this evening, at which point I'll be done unless there is another aspect of digital recording anyone would like me to test . . .
mshilarious
Banned
apl
Stand Up Comity
That's a crappy definition, especially for a Health section.
Objective data is information gathered without input from the gatherer.
For example, if you ask me how hot it is outside, if I say, "Not too bad," that's subjective. If I say, "75° F," that's objective because that number doesn't depend on me.
apl
Stand Up Comity
So you want a computer, software, and hardware mated together that you know will work? I hate to say this, but there's always Mac and ProTools.
apl
Stand Up Comity
LOL, I went back to the original post to remind myself of what we're discussing. There was no original question. It's an article that contains gross misinformation about how digital works. Specifically, it uses the connect-the-dots fallacy. As I said before, you don't connect the dots with straight lines, you connect them by putting in a sinc function of the amplitude of the dot and adding it to all the other dots' sinc funtions.
. Oh, and just to be a nit-picky old bastard, I'll point out that there was a question in the first post.
ausrock
Well-known member
P
pianodano
Guest
Just to be clear, I personally think there is very little wrong with digital sound today. But just to throw a monkey wrench into the works, I wish they could include algorithyms that would remove some of the odd harmonics introduced by modern circuitry and add even ordered harmonics. I am old enough to remember using a Silvertone twin twelve as a young teenager and hearing for the first time the first Silvertone SOLID STATE 6-10. I am confident that the solid state versus tube debate was NEVER settled at all and, I have long suspected that perhaps a part of what people think they don't like about the sound of digital is not caused so much by the process as the, for the most part, cheap circuitry and componets used. Hence the resurgence of boutique tube based expensive gear to help remove the edge. I do know that here, where my (partly digital) system goes into a analog console with outstanding eq, it sounds fine.
apl
Stand Up Comity
There are sincs between the dots. I was gonna do up a bit of pics in Excel, but Lavry's done a much better job in the first nine pages or so.
Dr ZEE
Anti-Pro Circles Insider
Uh, cool. I get it.
"Not too bad" is pretty objective to me. I gathered by asking you. No input from me. So it's pretty damn objective to me as a gatherer, well, according to your tight definition, it is.
Now, may I try to make up a definition for something. Just let me try. I'll do my best.
Hmmmmm, hammmm, hmmmmmmmmmm, here I go:
Definition of Engineering!
Engineering - putting things together so it works.
Did I do good. ?
hmmmmmmmm
I guess not.
Dr ZEE
Anti-Pro Circles Insider
Really? I take your word on that one. After all you are an engineer, you can figure out what and how to put something cool in there
and After all "it" must work, and so it does. there is still a bit of a problem, though. there are some freaks out there who still unhappy with what they hear.
/respects
W
wado1942
New member
OK, you hit on a pet peve of mine. Tape does not compress, never did and never will. Compression implies a time factor which tape does not. Tape DISTORTS when driven too hard. But people often confuse it as compression because the bias oscillator interferes with high frequency content, hushing HF componants as you approach the saturation point. Therefore the added harmonics are supressed.
At any rate, digital is not offensive because it fails to add problems. That's completely backwards thinking particularly since albums recorded on analogue tape and dumped to digital still sound digital to me. I remember the singer of my band commenting after I noted my experimental listening tests with analogue vs digital recordings of the same piece. He said "of course they'd like analogue better because its flaws cover up the flaws of the recording". This is like saying my car which is splattered with mud looks better because the mud covers the flaws in the paint job. Completely backwards and false. Sorry.
So anything with a 20Hz-20KHz frequency response would sound perfect? Dude, there's more to sound quality than frequency response. There's more to it than distortion and noise. If frequency content is all that mattered, we would have a whole new world from what we now know. Case in point, or years people have been noting that vacuum tube amplifier circuits sound better than solid state. There's been all sorts of studies on why and none of them were conclusive. Frequency response, distortion, noise were all very similar. A lot of people said that "tubes distort assymetrically and this is more pleasing than the symmetrical distortion of solid-state". Well, that may be somewhat true but it's not a logical conclusion. Most pro gear will have distortion characteristics below .1% at nominal levels. This is belows what would even be noticable and the only way to really get some noticable distortion (more like 1%) is to slam the equipment with high levels. Something that's not recommended or practiced with any regularity except by people who have no clue what they're doing. About 5 years ago, they came up with a new form of test equipment that can measure dynamic distortion. They discovered that while opamps show very little harminic distortion, they had absolutely massive dynamic distortion caused by the negative feedback loops required to maintain low harmonic distortion. Our ears are more sensitive to dynamic distortion, but it's been impossible to test for it till recently. What I'm saying is, our ears are telling us something is wrong with digital recording, we just haven't been able to test why. 20 years from now, that may change and a lot of people will feel very foolish.
I'll say this, as a mastering engineer, people are more offended by added stuff than missing stuff. A 3dB cut at 100Hz is not very noticable but a 3dB boost is very noticable. A downward expander is less noticalbe than an upward expander. The live feed from my board sounds better than what I get back from my 1/4" deck or my computer running 88.2KHz. Now if not adding distortion made sound worse, than the feed off of my 1/4" deck would sound better than the live feed from the board. This just isn't the case. This suggests to me that digital recording adds something to the signal that's offensive to human ears that science has not yet been able to pinpoint. We're getting close to a solution, but nothing so great yet.
I'll add one more thing, a 128Mbps MP3 has frequency response, distortion and noise levels very similar to CD, yet nobody in the know will argue that they sound the same.
Last edited:
evm1024
New member
I think you really need to step back and look at it from Dr Zee's point of view.
So we sample. We get (in theory) the value of the analog at a specific moment in time. Some time goes by and we sample again.We get another value.
In reconstruction of the original signal we apply a function to the values between the samples.
The original function was to settle to the samples value and then at the next clock tick move to the next samples value letting RC do the "smoothing". Classical stair step. Well that kinda sucked.....
Refine the time constants a few times, change the function to a more sophisticated one and get closer to reality. But still a function that at some level is an approximation of what SHOULD be.
There lies Dr Zee's point. All the math used to recreate the space between the samples is a construct that has a high probability of being the same as the actual signal. But let's not forget that it is not the actual signal.
apl
Stand Up Comity
Uh, cool. I get it.
"Not too bad" is pretty objective to me. I gathered by asking you. No input from me. So it's pretty damn objective to me as a gatherer, well, according to your tight definition, it is.
Good point. Your objective dataset includes my subjective dataset, ie, you can objectively quote me. But you don't know how hot it is.
Dr Z said:
That is not disputed. A lot of people like the way analogue sounds compared to digital.
apl
Stand Up Comity
I'd like that. I can't quite figure out what his POV is.
evm said:
That's not how D/As work.
evm said:
You don't get the original signal back from analog, either.