What is the cause of "digital glare"?

[andreasmaaan]

Thanks all, this is really interesting

I was also under the impression that even prior to CD there had been widespread use of digital effects and processors in studios. Not PC-based stuff, but rather hardware units. Was I mistaken about this?

 

[RayDunzl]

I was also under the impression that even prior to CD there had been widespread use of digital effects and processors in studios. Not PC-based stuff, but rather hardware units. Was I mistaken about this?

Our Rocking Teen Combo (we were all about 30) had a Lexicon Digital Delay box before we ever saw a CD. It sounded, well, digital...

Used it for adding a little reverb/echo to vocal and whatever was fed to it on the effects bus.

It had some interesting "effects" available beyond the basics. One was to modulate the timing of the delay/replay dependent on the attack of the incoming signal.

One last set at the Liquor Lodge in Tarpon Springs, a drum solo occurred, for some reason there were still a few people on the dance floor, still "dancing".

I dialed in some of that, think of a snare going beeeyoingwhannnng, and the dancers sort of melting into it...

Half of them ended up on the floor.

---

We made some recordings, still have them someplace, of course.

Imagine a rousing last tune of the gig rendition of Styxx - Born for Adventure, and at the end, no sound from the remaining audience is heard, just a seemingly stunned silence punctuated with the distant plaintive sound of a Pac-Man machine back past the pool tables, waiting for somebody to feed it, and then sounds of the band beginning to break down the gear.

 

[Pluto]

my understanding is that the Sony PCM adapters were often used for early A/D, and they were limited to 14-bit audio

I have here, a Sony PCM-501ES (switchable 16 or 14bit)) that still works as well as they day I bought it new. What a revolution that was. Within a few days of acquiring the 501 I had put two of my Revox A77s up for sale. The major – not insignificant – problem was that there was no means of precision editing this format so anything that required a serious edit had to be laid-off to another format. But as F1 format recordings had to be transferred to ¼" for delivery or disc-cutting in any case, this wasn't really a problem. The real value of the F1 format was its use at the original acquisition phase as the quality was equal to, or better than, analogue tape with Dolby A, with sufficient headroom to avoid the need to push your luck with levels and none of the tedious alignment necessary to ensure that a Revox was performing to the best of its ability. The very best quality super-deluxe VHS tapes costed three or four bucks apiece and they ran for THREE HOURS

The F1 format (largely the creation of Sony) was the first step in the revolution that has made top quality recording available to everyone. Of course, whether that is a good thing, or not, is another debate....

 

[Pluto]

my understanding is that the Sony PCM adapters were often used for early A/D

These adapters are not really of great general use as converters because most of the models had no digital output, as we understand it today. They had a video output which was the signal recorded on your VHS, βmax or (if you were really fancy) U-Matic machine but that was it. One later model (the 601, I think) did incorporate an SPDIF interface but, in general, those units that offered SPDIF did so by the grace of certain after-market manufacturers (notably Audio & Design Recording in Reading, England).

Another issue to be borne in mind is that some PCM adapters used a single converter, time-division-multiplexed between the two channels. This means that, without taking further corrective measures, an after-market SPDIF interface would have a timing error (of half the sample-rate) between the two channels. This was fully corrected when using the internal analogue interface but could be an issue with some of the early SPDIF interfaces. Back in the mid-eighties, correcting this error (which we would do today in the blink of an eye) was not so easy. For this reason, I suspect, the SPDIF interface for PCM adapters didn't really catch on like it otherwise might have.

 

[Frank Dernie]

and we cannot (yet) accurately measure the subjectively filtered signal at the inner ear or auditory cortex level.

But why would we need to?
In the case of hifi equipment the only link between the components is an electrical signal and if this does not change from one component to another then the movement of the loudspeaker drive units can't and won't change either. If the sound being radiated into the room is the same then whatever is the mechanism of our hearing is irrelevant, since exactly the same thing is happening in the room we are listening in.
If one of the pieces of equipment produces frequency response changes or distortion the movement of the loudspeaker diaphragms will indeed be different, we can certainly measure and maybe hear this if it is a big enough change, regardless of the way the ear works.
This is common misinformation put about in some hifi forums and is a "red herring"

 

[Pluto]

...the only link between the components is an electrical signal and if this does not change from one component to another then the movement of the loudspeaker drive units can't and won't change either

Precisely.

If I was asked, ’what is the single most important bit of kit in the entire recording chain?’ – I might say it's the little switch that selects, for listening, either the output of the console or the output of the recorder. If there is no audible difference between the two, what more would you want?

 

[Guermantes]

Thanks all, this is really interesting

I was also under the impression that even prior to CD there had been widespread use of digital effects and processors in studios. Not PC-based stuff, but rather hardware units. Was I mistaken about this?

Yes, Lexicon, AMS and Eventide manufactured digital processors in the '70s. This site has details on a few of them: https://www.vintagedigital.com.au/vintage-digital-effects/

Then, of course, there were the digital delay lines used for vinyl disc cutting from the late seventies, but by this stage Denon and other companies had started doing digital recordings. The writing was on the wall for pure AAA chains.

 

[Guermantes]

Our Rocking Teen Combo (we were all about 30) had a Lexicon Digital Delay box before we ever saw a CD. It sounded, well, digital...

I bought a Roland SDE-1000 digital delay while in high school because it made my guitar sound like The Edge

 

[Pluto]

I remember that ad

The box on the right was 19" wide and 5 or 6U high – all that just to get 5 seconds of delay.

I wonder how many gramophone records, vaunted today as the supreme peak of the audio art, lauded for their analogue purity and unsullied by digital pollution, have actually passed through one of these devices (or something similar) without anyone, at the time, noticing or caring? Records proudly wearing that cherished AAA badge, that have secretly flowed up and down the digital river and nobody cared. How sad.

But the ATR-100 was, quite possibly, the best tape machine ever made. They cost and arm and a leg these days, if you can find one.

 

[Halcyon]

But why would we need to?
In the case of hifi equipment the only link between the components is an electrical signal and if this does not change from one component to another then the movement of the loudspeaker drive units can't and won't change either

Well, having been at this also since the 80s and studies psychoacoustics, this is what I've learned. YMMV:

1. If manufacturers could design only by measurements, they would. Doing ABX double-blind sensory evaluations is a major PITA and very expensive. Ask Sean Olive.

2. Every single decade we have claimed we finally know how to measure a singnal so that we know what is pure, unadulterated, "like a wire without a loss", etc. And every single decade we get better measuring instruments and we find new sources of of non-linear distortion (we didn't previously consider) or find that "oh, human hearing can actually distinguish signal below noise, certain q IMD spikes, a certain order harmonic distortion or something else" better than we previously thought.

3. Properly done double-blind ABX tests still show that we don't know all. Either some listeners are psychic, we get very (un)lucky with fairly high trial counts OR they actually can distinguish (statistically, not always 100%) signals we thought were inaudible and/or we don't yet know how to measure a difference for.

With these experiences, I've grown more humble.

I used to think that under certain electrical (loaded) measured specs, say an amplifier A would be totally transparent from amplifier B. Then I run into these freaks who can distinguish them. Who am I to say they can't hear it, when they demonstrate it?

So it's back to the drawing board and having an open mind. Let the science progress by trials and experiments, not by declarations.

 

[solderdude]

1. If manufacturers could design only by measurements, they would. Doing ABX double-blind sensory evaluations is a major PITA and very expensive. Ask Sean Olive.

Yes, when transducers and acoustics are involved this is totally valid. What's it have to do with amplifiers, cables and DACs and stuff ?

2. Every single decade we have claimed we finally know how to measure a singnal so that we know what is pure, unadulterated, "like a wire without a loss", etc. And every single decade we get better measuring instruments and we find new sources of of non-linear distortion (we didn't previously consider) or find that "oh, human hearing can actually distinguish signal below noise, certain q IMD spikes, a certain order harmonic distortion or something else" better than we previously thought.

Yes too. It is already widely known humans can hear below noise levels for decades. Ask any person who decodes morse code. The better guys can still hear morse code where others just detect 'noise'.
Indeed there are studies that show humans can detect single or dual tones under the noise floor. With music I haven't seen any research showing this. Reason ... dynamic range of music is not the same as a tone. We can't hear music below the noise level or in any case won't enjoy it.
It's why there is noise shaping as well. Listen to Ethan's test files with and without dither.
The problem with measuring the hearing is that at some point we have to use transducers and subjectivity and 'luck' comes into play.
Measure 'reaction speed' for instance and you get a bell curve alike plot. Can some folks really react within a few ms or was it plain luck ?
In some cases the test method is flawed or results are interpreted incorrectly.

3. Properly done double-blind ABX tests still show that we don't know all. Either some listeners are psychic, we get very (un)lucky with fairly high trial counts OR they actually can distinguish (statistically, not always 100%) signals we thought were inaudible and/or we don't yet know how to measure a difference for.

Do you have any evidence of this ? What tests were these ? Link to research ?
Anecdotals don't count.

With these experiences, I've grown more humble.

I have researched (not studied) this as well for many decades and found the exact opposite.
Have met quite a few people claiming things but weren't able to do what they thought they could.
Researched my own hearing (a lot when I was younger) and every time found the things I thought I could do I actually couldn't.

So it's back to the drawing board and having an open mind. Let the science progress by trials and experiments, not by declarations.

Progress can only come from re-examining things and improving performance.
Isn't that what the progress in science is all about ?
Isn't that what's done daily by most researchers ?
How can there be progress if science only used declarations ?
Is there any research involving 'glare' in audio ?

 

[Blumlein 88]

Well, having been at this also since the 80s and studies psychoacoustics, this is what I've learned. YMMV:

1. If manufacturers could design only by measurements, they would. Doing ABX double-blind sensory evaluations is a major PITA and very expensive. Ask Sean Olive.

2. Every single decade we have claimed we finally know how to measure a singnal so that we know what is pure, unadulterated, "like a wire without a loss", etc. And every single decade we get better measuring instruments and we find new sources of of non-linear distortion (we didn't previously consider) or find that "oh, human hearing can actually distinguish signal below noise, certain q IMD spikes, a certain order harmonic distortion or something else" better than we previously thought.

3. Properly done double-blind ABX tests still show that we don't know all. Either some listeners are psychic, we get very (un)lucky with fairly high trial counts OR they actually can distinguish (statistically, not always 100%) signals we thought were inaudible and/or we don't yet know how to measure a difference for.

With these experiences, I've grown more humble.

I used to think that under certain electrical (loaded) measured specs, say an amplifier A would be totally transparent from amplifier B. Then I run into these freaks who can distinguish them. Who am I to say they can't hear it, when they demonstrate it?

So it's back to the drawing board and having an open mind. Let the science progress by trials and experiments, not by declarations.

I may reply in more detail when I've more of a chance. In short everything you've written here is revisionist history. It doesn't accurately describe what happened in the last 30 or 40 years.

 

[Blumlein 88]

Well, having been at this also since the 80s and studies psychoacoustics, this is what I've learned. YMMV:

1. If manufacturers could design only by measurements, they would. Doing ABX double-blind sensory evaluations is a major PITA and very expensive. Ask Sean Olive.

I think Sean Olive would say he does design by measurements. Double checks himself double blind. Toole said they long ago could have a better than 99% correlation with listener results for small speakers that don't go below 100 hz. Just using the spin-0-rama results.

2. Every single decade we have claimed we finally know how to measure a singnal so that we know what is pure, unadulterated, "like a wire without a loss", etc. And every single decade we get better measuring instruments and we find new sources of of non-linear distortion (we didn't previously consider) or find that "oh, human hearing can actually distinguish signal below noise, certain q IMD spikes, a certain order harmonic distortion or something else" better than we previously thought.

Knowledge like this has been around for a long time. Sure more details of more situations is figured out. That is how science works as it advances. For wire or gain without much distortion the basic levels seem well established and I've not seen such over-turned. Are you confusing actual science or practical research with marketing and buzzword based design? THD/IMD of less than .1% (-60 db) seems to be the point of inaudibility. Some people will have abilities either side by a bit perhaps. Noise is obviously audible at lower levels in the right conditions. And with music those are probably more than enough. Again nothing new. Knowledge of hearing into noise floors is old old fact.

Until the advent of digital, sources themselves weren't flat enough or consistent enough to be audibly innocuous. Some preamps and a few amps were. We knew the limits though marketing liked to ignore those for their own purposes. Speakers still aren't.

3. Properly done double-blind ABX tests still show that we don't know all. Either some listeners are psychic, we get very (un)lucky with fairly high trial counts OR they actually can distinguish (statistically, not always 100%) signals we thought were inaudible and/or we don't yet know how to measure a difference for.

Unless you have some information to back this up it isn't credible. Do enough tests and sure enough some just thru chance get lucky, and appear to hear something when they really don't. Statistics isn't surprised by that.

With these experiences, I've grown more humble.

I used to think that under certain electrical (loaded) measured specs, say an amplifier A would be totally transparent from amplifier B. Then I run into these freaks who can distinguish them. Who am I to say they can't hear it, when they demonstrate it?

So it's back to the drawing board and having an open mind. Let the science progress by trials and experiments, not by declarations.

If people show they can distinguish something then great. With real loads very few amps are faultless even now. It isn't something new or unexpected. Amps that are identical spec'd into a resistor may not be into a speaker.

 

[Frank Dernie]

If people show they can distinguish something then great. With real loads very few amps are faultless even now. It isn't something new or unexpected. Amps that are identical spec'd into a resistor may not be into a speaker.

Very much so.
When it comes to power amplifiers I am still somewhat sceptical about the tests. It is obvious that low distortion into a simple resistor is probably only slightly relevant. The effect of reactive loads - ie speakers - on amp performance both from the pov of distortion, clipping and frequency response changes is, for me, the most important criterion.
There are few such tests published but the Stereophile tests showing audible levels of distortion from many of the popular models and output impedance magnitudes causing sufficient changes in speaker frequency response definitely indicate why some power amplifiers may well sound different to well engineered ones. Unfortunately they only have one simulated speaker load, much better than a resistor but far from the whole story.

 

[tmtomh]

Digital “glare”

Way back, let's say the eighties, when an ‘outsider’ listened to studio sound, a very common remark was how ‘harsh’ it seemed. We put this down to the fact that John Doe was not used to listening to a reasonably clean top-end that had been neither mangled by the vinyl production process nor heavily attenuated by the performance of cassette tape >6 or 7kHz.

With the arrival of the CD, I suspect that huge swathes of the listening public experienced, for the very first time, the true sound of a reasonably flat response up to the typical limits of human hearing. While we are (thankfully) largely free of analogue cassette tape, I wonder how many so-called enthusiasts set their personal sonic reference point to a level defined by the sound of vinyl...

I think this gets to the heart of it. Digital "glare," while a very subjective term, clearly refers to the frequency range extending roughly from the upper midrange to the mid-treble. At a rough guess I'd estimate 3-8khz.

IMHO there were several mutually reinforcing reasons that this became a thing upon the introduction of CD players to the marketplace:

1. Vinyl playback tends to produce frequency-variable channel crosstalk that disproportionately impacts the mids and treble (because humans are more sensitive to directional aka stereo L-R cues in the higher frequencies);
2. Vinyl playback tends to produce resonances that often give it a warmer, softer sound, which in somewhat more objective terms means that it adds 2nd-order bass harmonics that thicken up the mid-bass - and bumping up mid-bass is a very effective way of taming the perception of treble in a way that some (many?) listeners find pleasant - arguably even more effective (more "pleasant" sounding) than turning down the treble directly;
3. As noted by Pluto, vinyl and cassette's relatively low SNR/SINAD and imperfect frequency response reduced the clarity and therefore perceived "bite" of treble frequencies in particular;
4. As noted by a few folks already, some early CD players had issues of their own - DAC chips, filters, and/or analogue sections - that contributed to perceived sonic problems.

All of these issues aside from #4 were about the nonlinearity and euphonic distortion of analogue media - in other words, they were about long-entrenched listening habits rather than actual high fidelity to the musical source material. These listening habits still rule the roost when it comes to the audiophile press - and when it comes to the listening preferences of a lot of people who are into hi-fi and music. People over 50 came of age with vinyl and cassette, and so that's the sonic profile that they heard in their formative years. And a good number of people under 40 are into vinyl for reasons that have nothing to do with accurate sound reproduction, but rather with the idea of it, the tactile experience, and the notion that vinyl gets one closer to the music in a "feel" way rather than a clinical fidelity way.

I first got into hi-fi around 1981 (I was only 12 but my father was really into it, so I went around to hi-fi shops with him on the weekends, and I read all the magazines he subscribed to). I remember very clearly in that 1981-84 period, when digital recording was emerging as a powerful new trend but CDs had not yet become a widespread consumer product - at that time there was pretty much zero concern about, or criticism of, digital. Vinyl cut from digital recordings was considered a best practice for audiophile product, and top-shelf for audiophile sound. No one started complaining until the reviews of the Sony CDP-101 appeared in Stereophile and the other audiophile magazines.

In fact, I might be overstating things, but I would guess that if one is looking for the moment when the cleavage between measurements and subjective impressions really took root in the audiophile press, it's likely to be around 1983, when the audiophile press started hating on most of the early CD players and could not square what they were hearing with the usually superlative measurements of the units.

 

[Frank Dernie]

Maybe I am being unfair but I think a lot of the Journalist’s mates in the business owned smallish specialist companies who hadn’t a hope of producing a CD player and they started inventing things wrong with CD to convince their customers not to go there, and to keep buying the record players.
Here in the UK the Linn proprietor slagged off CD and digital in general at every opportunity but was unable to tell the difference the early Sony PCM-F1 system in line and a straightforward interconnect.
At least he had the good grace to admit it.
It also seems a lot of people prefer the high frequencies to be rolled off, as they are with most record players and radios.

 

[Willem]

Indeed, the audiophile press seems to be in collusion with the small uncompetitive manufacturers, be it of vinyl decks or valve amplifers. Each time I attend live performances of (mostly larger scale) classical music, I am struck by the sometimes harsh glare of how music really sounds.

 

[Thomas savage]

Indeed, the audiophile press seems to be in collusion with the small uncompetitive manufacturers, be it of vinyl decks or valve amplifers. Each time I attend live performances of (mostly larger scale) classical music, I am struck by the sometimes harsh glare of how music really sounds.

Oh they definitely are in cahoots, HI End audio is a very small world and the reviewers ( can't call them journalists really) know they have to work together with the little manufacturers to keep it all going, Especially in certain markets like the UK and U.S. .

It's a bit of a joke really and it's why most reviews are useless with little or no reporting on actual issues of ownership like reliability or quirks certain designs have. Iv been on the inside when the 'talk' happens between the importer or distributor and the reviewer, anything negative is simply not allowed. Some times the reviewer will sneak something in but you will need to be skilled at reading between the lines to spot it.

Bit off topic , my apologies.

 

[Pluto]

I think a lot of the Journalist’s mates in the business owned smallish specialist companies who hadn’t a hope of producing a CD player and they started inventing things wrong with CD to convince their customers not to go there

Phew – a hugely cynical supposition but, quite probably, correct in principle. Talking of cynical …

It's 1982 and we're at the monthly sales meeting of a well-known analogue audio manufacturer.​

Chairman: Right chaps – we seem to be approaching a time when all these new-fangled CD player things will sound just like the master tape and they only cost a few hundred quid. Damn it all, the Japs will be selling them in the supermarkets next and our customers will go there, lidl by lidl, instead of their audio dealer. What do we do?

Marketing exec: Well Sir, it's rather simple really. We have our journalist friends work on the cult of “subjectivism”.

Chairman: What the blazes is that?

Marketing exec: It's when the journos. write about how much they simply luuuv our products to bits in spite of the fact that they have a frequency response like a donkey's hind leg. In fact, it doesn't really matter how cheaply the stuff is made because as long as our friends tell their readers just how great this new model sounds and the warm rosy glow it gives them, nobody will give a monkey's about anything else.

Chairman: But what about journalistic integrity?

Marketing exec: Well Sir, we've already started to address that little err.... conundrum. We've had lunch with the publisher of ”HiFi Ruin”...

(interrupting) Accountant: Yes Sir, we spent £57,000 plus expenses with them last year...

Marketing Exec: ... and we discussed the idea that their coverage of the product ought to be a little more sympathetic than it has been lately... we'll send the journalists a case of sherry at Christmas and let them keep the occasional product... think of it as the oil that lubricates the cult of subjectivism.

Chairman: That works for me! Trebles all round!

 

[Killingbeans]

Each time I attend live performances of (mostly larger scale) classical music, I am struck by the sometimes harsh glare of how music really sounds.

Gets me thinking about my old grandma who refused to turn on her hearing aid, because the real sounds of the world were just: "horrible noise"

I have the sneaking suspicion that most of the people who believe in the magic of ultrasonics in music and go for 24bit/768kHz files/DACs, RF bandwidth amps and/or 'supertweeters', actually also find any frequencies above 7 or 8KHz to be agonizing (if at all audible).