There's no such thing as digital

[mauidan]

http://www.audiostream.com/content/draft

 

[Phelonious Ponk]

Fair enough; it's just a different analog. A much more precise analog with a system of recording and reading that is much less prone to error than grooves in plastic or magnetic particles on mylar. I'm good with that.

Tim

 

[jkeny]

The point is that digital is a concept (an abstract) - an agreed protocol - above a certain agreed voltage threshold it is interpreted as a 1 & below another threshold interpreted as a 0. Real world implementation of this protocol when transmitting digital is analogue i.e a waveform with finite rise & fall times - not a sudden, instant change from o to 1. Therefore when timing is important (as it is in digital audio) the point on the waveform where this voltage threshold is encountered is not precise - it sits somewhere on the upward slope of the wave (or downward slope of the wave). Any interference which disturbs the slope of the waveform shifts this threshold point & therefore the bit is changing at the wrong time. A bit at the wrong time is equivalent to a wrong bit - it is interpreted into the correct analogue amplitude but at the wrong point on the analogue waveform in the D/A conversion.

 

[Groucho]

This is the primary reason that digital audio has taken so many decades to come close to the sound of analog. When digital audio was introduced, none of the top analog designers of the day knew anything about it. So it was all designed by digital engineers. Digital engineers have gone through years of training where these problems were never mentioned.

This is precisely the point I was discussing in the thread on early digital recordings. It is taken as an undisputed fact among modern audiophiles that early digital was rubbish. Yet it seems that the very first digital recordings from the 1970s(!) are still regarded by classical music lovers as absolutely top notch, and we can judge for ourselves - the ones I've heard sound pretty good to me.

This whole thing sounds like some made-up 'wise words' that he will never be called upon to justify.

 

[jkeny]

This is precisely the point I was discussing in the thread on early digital recordings. It is taken as an undisputed fact among modern audiophiles that early digital was rubbish. Yet it seems that the very first digital recordings from the 1970s(!) are still regarded by classical music lovers as absolutely top notch, and we can judge for ourselves - the ones I've heard sound pretty good to me.

This whole thing sounds like some made-up 'wise words' that he will never be called upon to justify.

Perhaps it shows is that the main problems with digital were/are with the playback side rather than the recording side?

 

[Don Hills]

The point is that digital is a concept (an abstract) - an agreed protocol - above a certain agreed voltage threshold it is interpreted as a 1 & below another threshold interpreted as a 0. Real world implementation of this protocol when transmitting digital is analogue i.e a waveform with finite rise & fall times - not a sudden, instant change from o to 1. Therefore when timing is important (as it is in digital audio) the point on the waveform where this voltage threshold is encountered is not precise - it sits somewhere on the upward slope of the wave (or downward slope of the wave). Any interference which disturbs the slope of the waveform shifts this threshold point & therefore the bit is changing at the wrong time. A bit at the wrong time is equivalent to a wrong bit - it is interpreted into the correct analogue amplitude but at the wrong point on the analogue waveform in the D/A conversion.

Are you referring to jitter? Where, in a DAC, a sample is decoded into the right analogue value, but slightly earlier or later than it should have been?

 

[Groucho]

Perhaps it shows is that the main problems with digital were/are with the playback side rather than the recording side?

Is there a big difference in principle? According to the accounts linked to in earlier posts in the Early Digital thread, the playback was good enough for the classical labels to audition the recordings and adopt digital as their new medium (rather than direct-to-disc or tape), and it does seem that people could even tell it was good after transfer to LP (i.e. played back to the cutting machine).

I find it all rather astounding, and I'm thrilled to find that the earliest recordings are on Spotify (Premium of course) so I can have a reasonably high quality listen for myself. It can't be bad if recordings at the birth of a medium almost 40 years ago are still indistinguishable from the latest technology to listeners today, and still highly regarded by the classical music fraternity. It tells me that digital recording really has got something going for it!

 

[j_j]

Well, some facts:

"Digital" is a sampled, quantized ANALOG of a continuous-time, continuous level signal.

"Analog" is a magnetically stored, continuous time, continuous level ANALOG of the original continuous-time, continuous-level signal.

With digital, you take all the hit up front, and copies are free from further degradation as long as hardware isn't broken.

With analog, every play and every copy must be worse. Physics requires it.

In practice, analog storage has quite a bit more noise, practically speaking narrower bandwidth, and lots of distortion. Some of that distortion sounds good.

The proof, though, is in the pudding. You can make a digital copy of an LP that sounds like the LP. You can not make an LP copy of a digital signal that sounds digital. That's a clear result as far as "accuracy".

Now, why don't they put the euphonic distortions into digital (rather than the hypercompressed crap they put out now days)? Good question.

ETA: I'd almost go as far as saying that an ADVANTAGE of analog is that you can't brickwall the level like you can with digital. Yeah, that's a "flaw" in analog recording, but it prevents the worst kinds of abuse, too.

 

[Phelonious Ponk]

Well, some facts:

"Digital" is a sampled, quantized ANALOG of a continuous-time, continuous level signal.

"Analog" is a magnetically stored, continuous time, continuous level ANALOG of the original continuous-time, continuous-level signal.

With digital, you take all the hit up front, and copies are free from further degradation as long as hardware isn't broken.

With analog, every play and every copy must be worse. Physics requires it.

In practice, analog storage has quite a bit more noise, practically speaking narrower bandwidth, and lots of distortion. Some of that distortion sounds good.

The proof, though, is in the pudding. You can make a digital copy of an LP that sounds like the LP. You can not make an LP copy of a digital signal that sounds digital. That's a clear result as far as "accuracy".

Now, why don't they put the euphonic distortions into digital (rather than the hypercompressed crap they put out now days)? Good question.

ETA: I'd almost go as far as saying that an ADVANTAGE of analog is that you can't brickwall the level like you can with digital. Yeah, that's a "flaw" in analog recording, but it prevents the worst kinds of abuse, too.

I'll take neither, thanks. I don't find the distortions of tape, and particularly vinyl, to be "euphonic. I even prefer moderately compressed digital to those distortions. YMMV.

Tim

 

[j_j]

I'll take neither, thanks. I don't find the distortions of tape, and particularly vinyl, to be "euphonic. I even prefer moderately compressed digital to those distortions. YMMV.

Tim

I won't argue with that. Your preference is your preference, and that's all there is to it. I tend to agree, personally, for the most part.

 

[mep]

Well, some facts:

The proof, though, is in the pudding. You can make a digital copy of an LP that sounds like the LP. You can not make an LP copy of a digital signal that sounds digital. That's a clear result as far as "accuracy".

I disagree. Listen to Donald Fagan Nightfly on LP and tell me that it doesn't sound digital. Ditto for Dire Straits Brother in Arms to name two LPs off of the top of my head. And maybe this is just semantics at play here with your choice of words. My point is that some LPs that were sourced from digital definitely retain the digital sound.

 

[Groucho]

My point is that some LPs that were sourced from digital definitely retain the digital sound.

But the fact that some classical albums, apparently, don't have a digital sound presumably shows that digital is OK in principle, even if some pop albums do sound 'digital'..? Do we know that when we say 1980s pop albums sound digital we're not just locking onto the particular style rather than the sound of the medium? Or that we're not really commenting on the state of early digital reverb algorithms and the appalling but ubiquitous Yamaha DX7 synthesiser?

 

[Don Hills]

mep,
I assume we're discussing LPs cut from a digital recording, as distinct from LPs cut from an analogue tape recording. The differences between digital and tape sources can be quite significant, so I would expect some differences to survive the LP cutting process.

I suspect some of the debate here may be due to the interpretation of your remark that it "sounds digital". To some people this is a bad thing; to others it is a good thing. The same disagreement occurs between the producers of the recording. Classical music producers tended to be early and enthusiastic digital adopters because they wanted a close match between what they heard in the studio / hall and the final product. For many other genres, tape was often used as an effect - the "tape sound" was an integral part of the final product. It's not surprising that when consumers of these genres heard music produced without the "tape sound", they judged it to be less accurate. I'm perhaps over-generalising, but I hope you see what I mean.

 

[microstrip]

(...) The proof, though, is in the pudding. You can make a digital copy of an LP that sounds like the LP. (...)

I would be very happy if some one could provide me with copies of a few LPs I own that could prove it. I suggest a few titles : The King James Version [Sheffield Lab Direct Disc ], or the DG Rigoletto conducted by Giulini and sang by Placido Domingo. The digital versions of most of the LPs I enjoy are all really disappointing compared with the LPs - I find difficult to believe that it is just incompetence of the people who did the transcriptions.

 

[Phelonious Ponk]

I would be very happy if some one could provide me with copies of a few LPs I own that could prove it. I suggest a few titles : The King James Version [Sheffield Lab Direct Disc ], or the DG Rigoletto conducted by Giulini and sang by Placido Domingo. The digital versions of most of the LPs I enjoy are all really disappointing compared with the LPs - I find difficult to believe that it is just incompetence of the people who did the transcriptions.

Proof is going to be hard to come by with you in Portugal, micro, but I'd bet the deed to my house that I could make a digital copy of your favorite album, ripped from your turntable, and blind, played back through your system, you wouldn't be able to tell it wasn't your LP. I doubt it would even require "high-end" ADC/DAC. Basic studio stuff would do just fine. In fact, I wouldn't trust the high-end stuff; I'd be afraid it would leave its sonic signature and I'd lose my house.

Tim

 

[j_j]

mep,
I assume we're discussing LPs cut from a digital recording, as distinct from LPs cut from an analogue tape recording.

Doesn't matter. Cut a Redbook recording from the LP, play it opposite the LP.