Just curious as to why still analog??

[regebro]

Good explanations. In analog the signal i stored in a way that is an analougue to the acoustic wave. In digital it's stored as numbers. What people do is look at the numbers as if it was an analogue and say "hey, that looks nothing like the original wave, it's jagged and crooked and the information between the sample points are missing!" But that is a false conclusion, because you are looking at the numbers representing the wave, not at the the wave as it is output after a D/A converter, which is the relevant part.

It is a bit like looking at all the pieces of a kit-car and say "that looks nothing like a car!"

 

[cjacek]

Could it be then that the real difference between the sound of Analog vs Digital in NOT about HOW it is stored or recorded but rather on WHAT it is being recorded on ? Put simply: Analog tape sounds one way because the magnetic surface makes it sound that way. Digital sounds another way because the polycarbonate substrate, a reflective metalized layer, and a protective lacquer coating (or whatever material is used for Hard Discs, Flash Cards, DVDs etc ....) make it sound that way.

~Daniel

 

[regebro]

Could it be then that the real difference between the sound of Analog vs Digital in NOT about HOW it is stored or recorded but rather on WHAT it is being recorded on ? Put simply: Analog tape sounds one way because the magnetic surface makes it sound that way. Digital sounds another way because the polycarbonate substrate, a reflective metalized layer, and a protective lacquer coating make it sound that way.

~Daniel

Then digital tape would sound like analog tape, so no.

But you are partly correct. Analog tape sounds like analog tape largely because of the medium. The non-linearities of tape is a significant part of it's sound, including the famous tape compression.

Magnetic tape is not an easy format to record on. Early recordings sounded like crap, and was not usable for anything but voice recordings. Not until two guys at AEG, by mistake, in 1941 notice that overlaying a high-frequencey signal improves the sound significantly does it gets usable for music, and even then it wasn't very good. It still needs a whole bunch of filtering mechanisms to make the frequency response anything even remotely like flat.

Here are two enlightening pages on that topic:

http://www.endino.com/graphs/
http://www.digitalprosound.com/Htm/TechStuff/2000/Aug/AnalogTape2.htm

Recording to tape, even high-end tape, will impose a certain sound on whatever you record.

"To me, analog is unpredictable; it does that funny thing to the bottom end. You work really hard on the bottom to get it exactly right, and then you play it back on your analog tape, and it's like, 'Oh, what happened there?' The storage medium is making decisions about what the bottom end should sound like." - Bob Clearmountain

High end digital does not do this. Low end will probably fuck things up, just as low end analog will, but high-end digital is WAY more accurate than analog tape is. And this, to me, is the power of analog: It will make things sound *good* and *coherent* because it imposes it's sound onto whatever you record. Digital just plays it back as you record it.

 

[cjacek]

And this, to me, is the power of analog: It will make things sound *good* and *coherent* because it imposes it's sound onto whatever you record. Digital just plays it back as you record it.

I think this is a good statement. The way I understand it is that Analog creates this "recipe" fom all the ingredients (of sound in our example) that just makes "sense". Digital on the other hand can very accurately "display" each of the "ingredients" but it's not very "coherent" or all of these "ingredients" don't necessarily "marry" well together to form a pleasing sound, like Analog can.

~Daniel

 

[regebro]

That's my understanding too.

What you can say is that because if the inherent huge limitations in analog, especially in tape, but also in creating mixers (remember the stupid summing hype a year or so ago) most constructors has focus on making every piece of equipment sound good, because you'd end up in a place where incresaing the accuracy in one way (say, frequency response) would decrease it in another (say, phase-coherency). So what do you do? You take the best-sounding compromise! With digital, people have been focusing on making it sound accurate. This is for two different reasons:
1. It's easier than making it sound good, because "accurate" can be measured and computed, "good" can't.
2. Up until the 70's "accurate" and "good" were one and the same, and it is still often confused. That confusion seems to me to be the base of many peoples wish to find something "wrong" or "missing" or "destructive" in digital technology.

There is absolutly no doubt whatsoever that of all my mic preamps, my Art Tube MP is the least "accurate". It has a distinct high-frequency rolloff. Still, it's the one all vocalists I have recorded (which, including me, is the amazing numebr of four!) prefer! It just makes them sound better. What is it tha makes them sound better? Well it is the high-frequency rolloff! It is the incaccuracy that makes the preamp sound good on vocals!

What we percieve as good is not the same thing as accurate.

 

[evm1024]

This calls for visual aides.

Se attached file. The two waves, red and blue, has a phase relationship that is WELL below the sample-rate. But as you see, because of their difference in phase, they get sampled at slightly different points in the wave, and the samples values therefore are different. When recreated, therefore, the phase relationship will be preserved. So, yet again, how small phase differences that can be preserved is not a function of samplerate, but a function of samplerate and bitdepth. (and of course, of the A/D converter, but that's implementation dependant).

This is the theory HOWEVER the real world intrudes and intorduces jitter in the clocks, quantization error, integeration time constants that very by frequency in the sample window, drift in the hold capacitor, outright missing codes in the conversion, mis conversions and so on. This means that the phase relationship is not preserved in the order of one bit. The information is lost. You show us a sine wave that you sample presuming a capture system that has no faults and do not show us the D/A side where the sine wave needs to be regenerrated and that has it own set of errors that remove alldata in the range that you are speaking of.

Dude just a 0.001 % jitter in the conversion clock (go take a look at the clock specs and conversion times) completely removes all that you said in terms of precsion. The phase information that you speak of is lost in the noise.

Plus music is not a constant monotonic sine wave but rather a very complex waveform.

Tektronix once offered a Transient Wave Analyser that had 500 pS precsion (as compaired to your proposed 346 pS limit). This machine cost $120000 and required about a month of hand tuning to get that kind of precision. It was speced for that precision only at 72 degrees plus or minus 2 degrees. Oh and I might add that it was a 7 bit flash converter.

But I guess this thread is all about the real world and ignoring it. Some like "Digital" sound, some like "Analog" sound and others cannot hear the difference.
Regards

 

[My Name]

That's my understanding too.

What you can say is that because if the inherent huge limitations in analog, especially in tape, but also in creating mixers (remember the stupid summing hype a year or so ago) most constructors has focus on making every piece of equipment sound good, because you'd end up in a place where incresaing the accuracy in one way (say, frequency response) would decrease it in another (say, phase-coherency). So what do you do? You take the best-sounding compromise! With digital, people have been focusing on making it sound accurate. This is for two different reasons:
1. It's easier than making it sound good, because "accurate" can be measured and computed, "good" can't.
2. Up until the 70's "accurate" and "good" were one and the same, and it is still often confused. That confusion seems to me to be the base of many peoples wish to find something "wrong" or "missing" or "destructive" in digital technology.

There is absolutly no doubt whatsoever that of all my mic preamps, my Art Tube MP is the least "accurate". It has a distinct high-frequency rolloff. Still, it's the one all vocalists I have recorded (which, including me, is the amazing numebr of four!) prefer! It just makes them sound better. What is it tha makes them sound better? Well it is the high-frequency rolloff! It is the incaccuracy that makes the preamp sound good on vocals!

What we percieve as good is not the same thing as accurate.

Spot on!!!!!

This is a great post, and should be used to explain to anyone who wonders why analog is still popular and does sound "better".

 

[regebro]

This is the theory HOWEVER the real world intrudes and intorduces jitter in the clocks, quantization error, integeration time constants that very by frequency in the sample window, drift in the hold capacitor, outright missing codes in the conversion, mis conversions and so on. This means that the phase relationship is not preserved in the order of one bit.

Of course not. Just as the phase relationship of two signals run through an analog mixer is never perfect.

The original claim was that it was impossible for digital to record phase relationships signal delays which were lower than the space between two samples. I explained above why this statement was incorrect.

You show us a sine wave

No I don't. I specifically avoided using a simple sinewave.

that you sample presuming a capture system that has no faults

Yes. This is theory. I'm explaining why the theoretical claim was incorrect. I also specifically noted that practical performance is implementation dependent. Just as with analog.

and do not show us the D/A side where the sine wave needs to be regenerrated and that has it own set of errors that remove alldata in the range that you are speaking of.

No, it hasn't.

See, you wanted an intelligent conversation. But your reaction to that intelligent conversation, and your reaction to facts that contradict your preconceptions, is making these kinds of silly excuses for not accepting the facts. Any time now, you will revert to a pissing contest. I'd like to state here an now, that this is NOT my fault.

Dude just a 0.001 % jitter in the conversion clock (go take a look at the clock specs and conversion times) completely removes all that you said in terms of precsion.

No, it doesn't.
1. The jitter has to be in the same order of magnitude as the phase relationship to cover it up.
2. Your argument here is basically "if you make crappy ADs and DAs then digital will sound like crap". Well, DUH!

Tektronix once offered a Transient Wave Analyser that had 500 pS precsion (as compaired to your proposed 346 pS limit).

You are confusing the theoretical limit with the limits of practical implementations, first of all. Second of all, you went from tlking about the phase relationships between two different waves, to transient wave analyzers, which are a completely different thing. I don't see how it applies.

But I guess this thread is all about the real world and ignoring it.

Ah, and here came the pissing contest. Surprse, surprise.

Next time you are about to say that you want an intelligent conversation, I would suggest that you shut the fuck up until you actually mean it.

That said, I note that you actually have absolutely no actual criticism towards the explanation I did, and I therefore regard the matter as settled. There is nothing inherent in either 44.1 or 96 kHz sampling that makes it impossible to capture the 10us delays that we were supposed to be able to discern.

Note also that we use theis sensitivity to delays between signals to locate stuff. If 44.1 Khz digital was not able to capture this, then the pseudo.achostical systems used to do fake "3d positioning" in headphones would not work. They do work, hence digital can capture it.

Case closed.

 

[evm1024]

Funny How we went to in theory...

In theory tape does not hiss and digital does not jitter. Positioning of a sound source in a sterio system like your ears is found through phase relationships between the ears, through intensity differences and through the direction that the sound enters the ear plus whatever else we do not know about. The brain uses many clues. In reality there are very few systems that perserve phase relationships at the one bit level. It should be noted that light travels 300000 m/S in a vaccuum and an electrical signel travels at a speed slower than this through a wire (though very clos to this speed) thus an electrical signel will travel just 1 cm (0.30") in 333 pS. So if the cables from your microphones are not the same length to less than one cm your 333 pS resolution is lost. We do not even need to take into account jitter....and all the other things that are inherient in digital or analog systems.

We are having an intellegent conversation. In truth. You present an argument that I think is so off the mark that I take exception to it and write about it. You read my comments and say no,no,no. I read your rebuttial and present my own arguments and thus it goes. Sure there are some barbs tossed and some kidding but that is all the part of the give and take in an intellegent comversation.

I think that you are in error. Nyquist says you are in error and no one has been able to disprove Nyquist. I show you some of the reasons why nyquist trumps and you disagree. We both retrench and check our facts.

From my reading of your post there was not an indication that you were speaking theroetically. I read it as a fact of existing digital systems. Perhaps I was wrong in this understanding or perhaps you are changing the spin on your assertion. It does not matter too much one way ot the other.

If we are going to speak in theory then tape would not hiss and phase would be preserved and the dynamic range of tape would be greater than the dynamic range or the ear....Digital would not have quantization error, phase would be preserved and the dynamic range would be greater than the ear. Oh and both systems would have flat bandwidth from DC to 100 kHz and be undetectable one to the other,

In the real world you can hear the differences (note I do not say that one is better than the other).

I remember years ago when Sterio magizine did a comparision of CD players. They found in blind AB testing with a range of music and reviewers that the differences between players was less noticable than the differences in individual CD of the same music. Read that again. Perhaps that has changed.

I was once told that if you could make the other person in an argument resort
to swearing then they admit defeat. Have you conceded?

regards!

 

[acorec]

In theory tape does not hiss and digital does not jitter. Positioning of a sound source in a sterio system like your ears is found through phase relationships between the ears, through intensity differences and through the direction that the sound enters the ear plus whatever else we do not know about. The brain uses many clues. In reality there are very few systems that perserve phase relationships at the one bit level. It should be noted that light travels 300000 m/S in a vaccuum and an electrical signel travels at a speed slower than this through a wire (though very clos to this speed) thus an electrical signel will travel just 1 cm (0.30") in 333 pS. So if the cables from your microphones are not the same length to less than one cm your 333 pS resolution is lost. We do not even need to take into account jitter....and all the other things that are inherient in digital or analog systems.

We are having an intellegent conversation. In truth. You present an argument that I think is so off the mark that I take exception to it and write about it. You read my comments and say no,no,no. I read your rebuttial and present my own arguments and thus it goes. Sure there are some barbs tossed and some kidding but that is all the part of the give and take in an intellegent comversation.

I think that you are in error. Nyquist says you are in error and no one has been able to disprove Nyquist. I show you some of the reasons why nyquist trumps and you disagree. We both retrench and check our facts.

From my reading of your post there was not an indication that you were speaking theroetically. I read it as a fact of existing digital systems. Perhaps I was wrong in this understanding or perhaps you are changing the spin on your assertion. It does not matter too much one way ot the other.

If we are going to speak in theory then tape would not hiss and phase would be preserved and the dynamic range of tape would be greater than the dynamic range or the ear....Digital would not have quantization error, phase would be preserved and the dynamic range would be greater than the ear. Oh and both systems would have flat bandwidth from DC to 100 kHz and be undetectable one to the other,

In the real world you can hear the differences (note I do not say that one is better than the other).

I remember years ago when Sterio magizine did a comparision of CD players. They found in blind AB testing with a range of music and reviewers that the differences between players was less noticable than the differences in individual CD of the same music. Read that again. Perhaps that has changed.

I was once told that if you could make the other person in an argument resort
to swearing then they admit defeat. Have you conceded?

regards!

Damn, where did you come from? A cohesive, logical argument with facts and examples to back it up?

Get out of town. You must be new 'round here.

 

[acorec]

This is the theory HOWEVER the real world intrudes and intorduces jitter in the clocks, quantization error, integeration time constants that very by frequency in the sample window, drift in the hold capacitor, outright missing codes in the conversion, mis conversions and so on. This means that the phase relationship is not preserved in the order of one bit. The information is lost. You show us a sine wave that you sample presuming a capture system that has no faults and do not show us the D/A side where the sine wave needs to be regenerrated and that has it own set of errors that remove alldata in the range that you are speaking of.

Dude just a 0.001 % jitter in the conversion clock (go take a look at the clock specs and conversion times) completely removes all that you said in terms of precsion. The phase information that you speak of is lost in the noise.

Plus music is not a constant monotonic sine wave but rather a very complex waveform.

Tektronix once offered a Transient Wave Analyser that had 500 pS precsion (as compaired to your proposed 346 pS limit). This machine cost $120000 and required about a month of hand tuning to get that kind of precision. It was speced for that precision only at 72 degrees plus or minus 2 degrees. Oh and I might add that it was a 7 bit flash converter.

But I guess this thread is all about the real world and ignoring it. Some like "Digital" sound, some like "Analog" sound and others cannot hear the difference.
Regards

In defense of BOTH of you, these same arguments are going on over at Pro Sound Web all the time with people like Dan Lavery and phDs in digital electronics. Nobody ever seems to win the argument, but discussing these things leads to better design and more/faster research to move digital recording along.

BTW: Flash converters were one main reason that digital recording sucked. It has gone the way of the dinasoar in favor of 1 bit converters.

 

[regebro]

In theory tape does not hiss and digital does not jitter.

Correct. And therefore, we need to separate theroy and practice when this is discussed. If you say "In theory, this is impossible, because X" and I show X to be false, you can not come with the answer "well practically..." That's a bogus answer in that situation.

In reality there are very few systems that perserve phase relationships at the one bit level.

See. You persist. I have never claimed anything else. I now looked up the original statement by SteveMac:
"It's been said that the human ear is capable of percieving spatial differences in it's surroundings at a rate of something like 15 microseconds. They say sample rates of 96 still fall short of this." I showed that this statement was false.

Case closed.

We are having an intellegent conversation. In truth. You present an argument that I think is so off the mark that I take exception to it and write about it.

For that to be an intelligent conversation you need to present valid counter arguments. You have not done that.

I expect you to now agree that the original statement is false. Or at least drop this particular subject and bring up some other subject. You also indicated that I had confused bit depth and sample rate. The confusion however was yours and yours only. An excuse would not be out of place.

I think that you are in error. Nyquist says you are in error

No, he most certainly does not. Nyquist agrees with every little micrometer of what I'm saying and I agree with him.

 

[evm1024]

ractically..." Snip!

See. You persist. I have never claimed anything else. I now looked up the original statement by SteveMac:
"It's been said that the human ear is capable of percieving spatial differences in it's surroundings at a rate of something like 15 microseconds. They say sample rates of 96 still fall short of this." I showed that this statement was false.

Case closed.

Snip!

Let's see 96000 conversions per second = 10.42 uS per conversion period. This is nowhere near 300pS that you spoke of... 15 uS is not greater or equal than 2 times 10.42 uS. The the signel is not reproduced and thus the phase relationship is not preserved.

Let me ask a question of you. Did you say or did you not say that a 44.1 kHz/16 bit system was able to reproduce 346 pS?

regards

 

[Dr ZEE]

hopefully no one missed the significance of this historical snapshot from 1997 and his reference to the rise of tube technology, and his statement that there is a lot more bad digital than good.

I guess you're righ, I kinda' missed a 'time mark factor' there.
I've come across this article, which touches the effect of 'time line' (old vs. new attitude) factor in recording equipment evaluating process. Also many other interesting points there. Check it out, don't ignore . Here: Rules of the Game by James Boyk

/later

 

[evm1024]

Do read that article

Dr Zee, Very nice reference, a must read.

It makes me remember a time when I was over at an organists house to listen to a 33 rpm recording of some Bach. I was apprentising as an organ tuner at the time. Anyway, the record was so scratched up it was almost unlistenable. Hiss, scratch pop, click. It was worse than my grandmothers 78's.

But the organist did not hear any of the recording or playback defects. He only hear the phrasing of the playing and the stop selection. The state of tune of the organ, the voicing and the room acoustics.

Hearing is inthe brain.

Regards.

 

[regebro]

Let's see 96000 conversions per second = 10.42 uS per conversion period. This is nowhere near 300pS that you spoke of...

Ok, so you obviously did not understand the graph. If you want you are welcome to expand on what in the graph you don't understand, so I can make it clearer, since I assume it is a lack of explanation from my side that is lacking.

Let me ask a question of you. Did you say or did you not say that a 44.1 kHz/16 bit system was able to reproduce 346 pS?

(My emphasis on system).

1. I did not say that. The statement is absurd.
2. Stop the pissing contest already.

 

[acorec]

Then digital tape would sound like analog tape, so no.

But you are partly correct. Analog tape sounds like analog tape largely because of the medium. The non-linearities of tape is a significant part of it's sound, including the famous tape compression.

Magnetic tape is not an easy format to record on. Early recordings sounded like crap, and was not usable for anything but voice recordings. Not until two guys at AEG, by mistake, in 1941 notice that overlaying a high-frequencey signal improves the sound significantly does it gets usable for music, and even then it wasn't very good. It still needs a whole bunch of filtering mechanisms to make the frequency response anything even remotely like flat.

Here are two enlightening pages on that topic:

http://www.endino.com/graphs/
http://www.digitalprosound.com/Htm/TechStuff/2000/Aug/AnalogTape2.htm

Recording to tape, even high-end tape, will impose a certain sound on whatever you record.

"To me, analog is unpredictable; it does that funny thing to the bottom end. You work really hard on the bottom to get it exactly right, and then you play it back on your analog tape, and it's like, 'Oh, what happened there?' The storage medium is making decisions about what the bottom end should sound like." - Bob Clearmountain

High end digital does not do this. Low end will probably fuck things up, just as low end analog will, but high-end digital is WAY more accurate than analog tape is. And this, to me, is the power of analog: It will make things sound *good* and *coherent* because it imposes it's sound onto whatever you record. Digital just plays it back as you record it.

Yes! All true. The problem is that digital (if done right) will record what is being presented to the mic (or other source) and analog changes the source in a pleasing way.

 

[monty]

This is insane...you people need meds!!!

 

[Dr ZEE]

Hearing is inthe brain.

Hey, you are pushing it. Stop NOW! heh heh

Yeah, yeah yeah, ...old scatchy records. - Me and Bach in one room. Candles, fireplace crackles, glass of aged Cabernet... what else? ... sentements, man, pure 'baseless' sentiments

If you record piano nocturne on wax cylinder it is going to sound like wax cylinder, but in the universal swing of things it will bring the performer as close to the listener as it gets

If you record piano nocturne on cd, it is going to sound like cd (Yes, baby !!!!), but the link between the performer and the listener will be cut off by the scissors of technological arrogance in the 'comparator'

If physical touch does matter, then use wax, when sending audio-letter to your loved one.
**********

/respects

p.s. 1: technology, developing its product, based on presumption of inferiority and inability of man IS arrogant.

p.s. 2: There are no mysteries in any technology - it is all documented and accessible by an authorized user.
Mysteries are property of The Nature, accessible by a curious mind.

p.s. 3: I'm AM an ignorant lunatic, so I am free to say so

 

[arjoll]

Moved on from insane.....

This is insane...you people need meds!!!

No, I disagree . It has moved on from potential insanity (and, all respect to Dr Zee, posts which just didn't make sense to me!) to be a relatively interesting and lively discussion!

I think its telling that the general consensus here is that analogue does, in fact, sound better in a lot of cases. I think that in the last 'digital vs analogue' thread Regebro was completely misunderstood (look at his comments in this thread about the tube preamp, and the mention of the 'pleasing' sound of analogue tape), here everything is making a lot more sense. Its about using the best tools for the job.

Good digital may be, strictly speaking, more accurate. Good analogue can sound better because of innaccuracies of the medium.

BTW I'm writing this from the on-air studio of the local Community Radio station doing my regular Thursday night show. Here you get to hear the worst of both mediums - a playout system based on CDs ripped and encoded at 256kbps MPEG2, pushed through a worn out A&H desk, processed badly, pumped through a cheap stereo encoder and broadcast on FM, with its absolutely crap stereo separation and s/n ratio. Yet somehow I still enjoy the music.....