About noise and de-noizing... from Eroc

Sadly, I know only too well what goes on

Why would I want to withdraw my argument that processing samples in the digital domain is a recipe for lifeless sounding CDs?

 

Simply copying an analog tape to another analog tape will compromise the integrity of the original waveform.

 

There is still lots of misinformation about digital out there...

 

Just as analog processing will do! In either case you're changing the waveform.

Again, of course. You *don't want* the original waveform - that's why you're doing processing.

Answer me this - can you *prove* that it is *impossible* to get identical results using analog and digital EQ? I mean, if a digital effect can transform the signal in exactly the same way as an analog effect, what difference does it make?

How about this: you record the same tape to digital two ways - one with EQ and one without. You then compare the two files digitally and determine exactly what was changed between the two. Then you write a piece of software that will reproduce that difference exactly - digital EQ. How is the digital EQ any less "accurate" than the analog EQ, if the results are exactly the same?

 

Malc,

I see from your profile that you have a Plextor CD burner. If you don't have one already, download a demo of some quality audio editing software and do your own experiments.

 

I think we are really splitting hairs here.

Sure , TECHNICALLY, dumping an analog waveform onto a Pro Tools station and sampling it does destroy the analog wave form.

BUT, in the real world, doing this the right way, using high quality converters and a high sample rate, and paying attention to certain things renders a digital file that can sound so close to the original waveform as to be identical.

Sure it ISNT really identical, but if you cant tell in the listening stage, who cares? We LISTEN to music, we dont watch the waveform on an O Scope!!!!

 

The problem I have is when the word destroy is used. The sound is not destroyed if, as mikey says, you use quality converters going in to the software.

Pro Tools isn't the only quality software out there, either!

 

You seem to think that, just because it is possible to modify a waveform in the digital domain that this means that it will sound OK.

You're missing the point - changing the waveform in analogue is part and parcel of the analogue process. Every piece of analogue equipment has it's own "sound". EQing in analogue is just an extention of the fact that you can change the character of a waveform by using various electrical components, adjusting their parameters - this is not something that is alien to the original analogue signal.

A digital recording relies on sampling the analogue waveform to encode the waveform in a digital form. The whole theory of digital recording relies on being able to recreate the encoded analogue waveform from those samples. Changing the samples has nothing to do with the original intention of the process. The fact that you can play with the samples to create effects such as EQ, Reverb, NR etc, etc... does not mean that you will end up with a musical sounding result.

Who said that *no* processing was used (although in the case od RCA Bowie, DCC golds and the Japanese "Abbey Road" CD I suspect this is probably true)? All I said was:

"You don't need to be Einstein to realise that the most natural, analogue sounding CDs (I'm talking about re-issues of analogue recordings here) are the ones with the least digital processing (ideally none!). "

And I stick by it!!

 

The discussion that has developed from this Eroc thing is a serious one; but does anyone share my suspicions that we are being kidded? (re-read the initial posting)...

 

Friends,

It has been awhile since we had a good, engaging Analog v. Digital discussion - seems like "back in the day!" Anyway, have fun guys!

Bob

 

Nor does it mean that you won't end up with a "musical sounding result", either.

You seem to be far too concerned about the means rather than the end. Again: if processing digitally results in exactly the same waveform as processing in analog, who cares about the "original intention of the process"?

And where exactly is it written that the "theory of digital recording" only involves sampling and not processing?

 

OK, semantics always gets in the way doesn't it?!

When I say "destroys the integrity" I mean that, by altering samples, you have created a digitally encoded signal that no longer represents anything that existed in the analogue realm.

Suppose you EQ in the digital domain. The result will sound very similar to an analogue EQed signal. It is a subtle effect, but the strangely unnatural sound that I detect on digitally processed signals is a result of the "destroyed integrity" I refer to.

In an extreme example, take the new Free re-masters that Peter Mew mastered. They sound strangely cold and uninvolving to me. In the car or as background music you may not notice that something is wrong but sat in front of the speakers it just doesn't sound right. Mew has "destroyed the integrity" of the original analogue recordings.

So, when I say "destroyed" don't expect to hear distortion or noise - it is a subtle effect but one that sucks the life right out of the music leaving the listening experience a cold univolving one.

 

Again, there is still lots of misinformation about digital out there...

 

So, the problem you have is with the engineer, and what that engineer did, NOT digital!

So, can you conclusively prove this?

 

But it's not necessarily something that *couldn't* exist strictly in the analog realm.

But you're trying to say that since some digitally mastered recordings you've heard sound bad, that *all* digital is necessarily. That's kind of like saying "since I hit my thumb with a hammer, all it's good for is hurting people and it can do no good."

 

I get a kick out of the recurring Peter Mew bashing on this site.
I have his Bowie discs, and they sound no more or less processed than the "Hendrix Family Authorised Edition" of Jimi's cd's, which (as chance would have it) have been universally praised on this site.

This guy is not the devil! I heard 'Wingspan' played on a pair of Wilson WATT Puppies last year and it was SOMETHING ELSE! The piano was very realistic and dimensional.

AND I am used to the Wilsons, it was not a 'shock of the new' reaction.

 

Grant & Luke,

somehow you have gained the impression that I am anti-digital.

In fact, as a recording system, I think digital has a lot to offer - especially DSD technology.

What I don't like is the use of digital as a processing tool.

The Nyquist theorem has proved very successful. Nyquist showed (way back in the 1920s) that digital encoding of an analogue waveform and then its subsequent decoding could produce an accurate recreation of the original signal.

So, we sample an analogue waveform for digital encoding purposes with the aim to either store or transmit the waveform using the advantages afforded by digital electronics (low noise etc...). And, using Nyquist's theorem, we know that we are able to reproduce an accurate (the actual accuracy depend upon factors such as sample rate and wordlength) facsimilie of the original waveform.

So far, so good.

Now, while in the digital domain we choose to change the values of the samples to perform some desired modification to the waveform (maybe EQ, reverb, NR etc...).

But wait! We are no longer able to take advantage of the Nyquist theorem!! We can no longer accurately reproduce the signal that we encoded into digital.

OK, but we wanted to change the waveform anyway so what's the problem?

The problem is that the analogue waveform that we get back from the modified samples represents an "unnatural" waveform. By "unnatural" I mean that it could not occur this way naturally. Why is this important?

Well, take analogue and digital reverb as an example:

In analogue reverb you generate mechanical vibrations in a system (be it a spring, a plate, a chamber etc...) and pick up the echoes produced in the particular sytem of your choice. The factors governing the response of your chosen system are almost infinite. There will be resonances at different frequencies, unpredictable distortions etc etc.... The overall effect is a reverb that is pleasing to the ear due to its complexity. Afterall, we spend our lives hearing echoes of infinite variety all our waking hours, why should it be surprising that analogue reverb sounds pleasant to the ear.

In digital reverb, you use a digital processing algorithm that attempts to emulate analogue reverb. However, due to the complexity of real world analogue reverb, it is simply not possible to reproduce, digitally, the almost infinite idiosynchracies and quirks that makes analogue reverb what it is. In fact, most digital reverbs use very simple algorithms that only create a homogeneous reverb (ie without any resonances).

So, what are we left with?

If we apply digital reverb to a signal we get back an "unnatural" reverb that while sounding similar to analogue reverb, lacks all the subtle complexity that makes analogue reverb so wonderful.

Any use of digital processing leads to an encoded signal that can no longer represent a "natural" analogue waveform. For some this of no importance whatsoever. For me, it inevitably leads to a poor sounding result.

How many modern CDs sound lifeless and uninvolving, but you can't quite put your finger on why? I know what I think is the culprit......

Of course, the more digital processing applied, the worse the outcome (just listen to some Mew discs to see how the life can be sucked from well recorded analogue masters).


So you see, it is not that I am against digital recording - Nyquist's theorem is dear to my heart - but I do not have any affection for digitally processed audio.

 

Thom,

have you heard the Free remasters Mew did recently? Compared with the original CDs they sound sterile to me. I guess it all boils down to personal preference at the end of the day.

 

Marc,

While I don't know where I stand on this debate - I'd like to say your words aren't all landing on deaf ears (no pun intended ). Logically I get what your saying, and I think you have a valid point/concern.

Luke & Grant,

I agree with Marc in your responses seem to indicate that you're both missing his points. Grant - just saying "there's lots of mis-information about digital" isn't a reasonable rebuttal to what Marc has written. If you want to take him to task, I think it's only fair to respond to his points in the same logical manner as he's presented them.

HZ

 

Hey Malc,

Very well thought out posts. You have a considerable knowledge of this subject and I've enjoyed reading your posts. They'll get a man thinking.

Like HZ, I don't have allegiance to either side of the debate as I feel both have valid points. I do have an affectation for the analog world though.

 

Malc,
Thanks for taking the time to post this. I have learned a lot.
John

 

What can I say? We all have our personal likes and dislikes! But it does sound to me as if you are basing your opinion on math.

 

Hehe

Hey, but don't knock math - it's what the engineers used to create all this cool audio technology (analog AND digital) in the first place!

HZ

 

Oh, believe me! I'm not knocking it! I just wish I could understand it! My favorite subjects in school were music, english and history.