What is the cause of "digital glare"?

[confucius_zero]

In audio, you'll have some critics saying that:
- "this class D amp has digital glare"
- "this ESS DAC chip has Sabre glare"

I wonder what is glare, what causes it and would this glare be related to harmonic distortion that causes the treble to lack control or else?

Thank you!

 

[amirm]

There is no such thing. It is a made up thing by people who don't know how to perform controlled listening tests. If they did that, they would not hear any such thing!

 

[confucius_zero]

There is no such thing.

Alright thank you

 

[Sergei]

In audio, you'll have some critics saying that:
- "this class D amp has digital glare"
- "this ESS DAC chip has Sabre glare"

I wonder what is glare, what causes it and would this glare be related to harmonic distortion that causes the treble to lack control or else?

Thank you!

Don't know what they mean by "Sabre glare". In my experience, any contemporary DAC with a properly designed filter and output stage doesn't introduce anything I would call "glare". Perhaps they refer to some cheap Sabre-based DACs, with a rudimentary and/or overloaded output stage? Otherwise, quality DACs, Sabre-based or not, shall be subjectively virtually indistinguishable under controller listening tests, just like Amir says.

Class D glare is easier to encounter, especially in older designs, and in budget consumer electronics. The distortions of higher frequencies at highest and/or lowest output levels are usually caused by insufficiently high switching frequency, in combination with insufficiently sophisticated and/or non-linear low-pass filter, and excessive switching interference. See https://en.wikipedia.org/wiki/Class-D_amplifier#Design_challenges for discussion of some of these issues.

The design challenge of a transparent AB amp was mostly met in the 1980s, all the relevant patents expired long ago, best schematics and components are widely known, and today one can buy an incredibly good AB for not much money. The D class is not there yet. The designs are actively revised and patented. You can buy a really nice D, yet it will still cost a pretty penny (e.g. https://epc-co.com/epc/GaNTalk/Post...Sound-Quality-and-Efficiency-to-Class-D-Audio). Majority of D amps sold today can't touch the transparency of a classic AB at the higher audio frequencies.

I like using class D amps for subwoofers - that's the best match, as the major D issues diminish rapidly at lower audio frequencies, while the power efficiency advantage shines. None of my current home gear uses class D for tweeters, where the D issues are most noticeable. Class D driven tweeters are not often encountered in pro studio gear either. High-quality class D amps are OK for mid-range in my opinion - driving woofers in a center speaker of a home theater for instance.

 

[stereo coffee]

There is no such thing. It is a made up thing by people who don't know how to perform controlled listening tests. If they did that, they would not hear any such thing!

Would it be fair to say, performing controlled listening tests each time might be too formal- when simply listening to music
is what is being done. For instance humouring this a bit does one need a particular type of clothing for controlled listening tests
or will a bathing costume do the job ?

 

[confucius_zero]

The design challenge of a transparent AB amp was mostly met in the 1980s, all the relevant patents expired long ago, best schematics and components are widely known, and today one can buy an incredibly good AB for not much money.

I can attest to that. Bought the Yamaha RS202 and was blown away by the sound!

 

[amirm]

Would it be fair to say, performing controlled listening tests each time might be too formal- when simply listening to music
is what is being done.

You just need to do it once or twice so you get calibrated on how reliable your perception is. Take a case of "night and day difference" and test that blind. If you fail to hear such difference, then you will think twice the next time you trust such judgements.

For instance humouring this a bit does one need a particular type of clothing for controlled listening tests
or will a bathing costume do the job ?

No, nothing that funny. Just have a loved one do the AB switching a few times and see how consistent you are in your detectiona ability. If levels are different, you need to match those and that requires a bit of equipment. For other things like cables and such, you don't need to worry about that.

 

[Hugo9000]

If you do blinded testing, you don't have to worry about any glare hahaha I couldn't resist, sorry!

(also removes problems like veils and any other visual descriptors of what is supposedly a sonic characteristic hahaha!)


also, less likely to "notice" night and day differences without the light cues! hahahahaha!

 

[confucius_zero]

I saw what you did there @Hugo9000

 

[solderdude]

I have seen/read many 'subjective' reports on devices using Sabre DACs yet some were reported to not have any 'glare' nor the typical 'Sabre sound' but the majority (even some expensive ones) were said to have the infamous 'glare' or 'Sabre sound'.
This confuses me ... does this mean the 'glare' is not coupled/inherent to Sabre DAC chips ?

Does 'Sabre sound' truly exist ?
Is it the opposite of 'R2R sound' ?
Do filter settings (if possible) affect the glare ?
Can Sabre or other DS DAC chips with similar filter settings be told apart ?
Does one brand 'glare' more than the other ? Is Sabre the worst ?
Can Sabre sound be 'undone' in the analog stage or in the power supply ?
Can it be shown to exist in blind tests ?
Can it be shown to exist in measurements ?
Can it be demonstrated with nulls ?
Is it just 'bias' as peeps already know which DAC chip is used ?

Definitions of glare found on the web and (as far as I can couple the meaning of 'glare' to sound in a technical sense)
1: to be very bright and intense (poor quality and/or elevated treble)
2: to be too bright or showy (elevated treble or fake detail)
3: to shine with a harsh uncomfortably brilliant light (poor quality and/or elevated treble)
4: stand out, obtrude (fake detail or elevated treble and when it is obtrusive could be distracting/degrading the experience of listening to music)
5: cheap showy brilliance (exaggerated fake detail, usually caused by poor quality or elevated treble)

Seems like a word used by lack of technical descriptors on the 'reviewers' side. Seems related to extended or elevated treble, which I have never seen in any DAC measurement.
Rolled-off treble is seen a lot and does not seem coupled to 'glare' so rolled off may be the opposite of glare.

Was this a glaring moment ?

 

[confucius_zero]

to shine with a harsh uncomfortably brilliant light

Interestingly, I listened to the 400i (a headphone with a treble peak between 6 and 10K) on the atom and the headphone out of my Yamaha Rs202 speaker amp. While both could power the headphones well, I found the treble to be better controlled on the atom while the treble coming from the Yammy felt more wobbly and contributed to "glare".

I assume it's due to the difference in harmonic distorsion between an amp with better numbers (the atom) and one untested...?

 

[SIY]

No, nothing that funny. Just have a loved one do the AB switching a few times and see how consistent you are in your detection ability.

But that has to be done outside your presence lest Clever Hans inadvertently rear his head.

 

[March Audio]

I have seen/read many 'subjective' reports on devices using Sabre DACs yet some were reported to not have any 'glare' nor the typical 'Sabre sound' but the majority (even some expensive ones) were said to have the infamous 'glare' or 'Sabre sound'.
This confuses me ... does this mean the 'glare' is not coupled/inherent to Sabre DAC chips ?

Does 'Sabre sound' truly exist ?
Is it the opposite of 'R2R sound' ?
Do filter settings (if possible) affect the glare ?
Can Sabre or other DS DAC chips with similar filter settings be told apart ?
Does one brand 'glare' more than the other ? Is Sabre the worst ?
Can Sabre sound be 'undone' in the analog stage or in the power supply ?
Can it be shown to exist in blind tests ?
Can it be shown to exist in measurements ?
Can it be demonstrated with nulls ?
Is it just 'bias' as peeps already know which DAC chip is used ?

Definitions of glare found on the web and (as far as I can couple the meaning of 'glare' to sound in a technical sense)
1: to be very bright and intense (poor quality and/or elevated treble)
2: to be too bright or showy (elevated treble or fake detail)
3: to shine with a harsh uncomfortably brilliant light (poor quality and/or elevated treble)
4: stand out, obtrude (fake detail or elevated treble and when it is obtrusive could be distracting/degrading the experience of listening to music)
5: cheap showy brilliance (exaggerated fake detail, usually caused by poor quality or elevated treble)

Seems like a word used by lack of technical descriptors on the 'reviewers' side. Seems related to extended or elevated treble, which I have never seen in any DAC measurement.
Rolled-off treble is seen a lot and does not seem coupled to 'glare' so rolled off may be the opposite of glare.

Was this a glaring moment ?

Well I recently performed a blind test between a sabre dac and a soekris R2R with some Audiophile friends. Guess what, level matched and instant switching, they could not reliably tell which was which.

It wasn't an exhaustive test but still the difference was not enough to be "glare" ingly obvious.

It's an anecdote, make of it what you will, but my view is that most Audiophile reviews are woefully uncontrolled and cannot be trusted. So park reports of fundamental issues like glare in the bin.

 

[rebbiputzmaker]

Well I recently performed a blind test between a sabre dac and a soekris R2R with some Audiophile friends. Guess what, level matched and instant switching, they could not reliably tell which was which.

It wasn't an exhaustive test but still the difference was not enough to be "glare" ingly obvious.

It's an anecdote, make of it what you will, but my view is that most Audiophile reviews are woefully uncontrolled and cannot be trusted. So park reports of fundamental issues like glare in the bin.

The only trick to this is the quick switching with random music. No chance to pick out differences.

Have done the same thing many times and can pick differences Blind level matched testing is fine, but use music that is familiar and has artifacts that allows one the chance to pick out differences or deficiencies. A couple of quick examples, harmonic structure / overtones of bells, piano, or overload patterns of know voice mikes.

 

[andreasmaaan]

The only trick to this is the quick switching with random music. No chance to pick out differences.

Have done the same thing many times and can pick differences Blind level matched testing is fine, but use music that is familiar and has artifacts that allows one the chance to pick out differences or deficiencies. A couple of quick examples, harmonic structure / overtones of bells, piano, or overload patterns of know voice mikes.

What are some examples of DACs that you've been able to differentiate between using your method?

 

[Frank Dernie]

I have quite a few CDs which have a harsh sound I could call glare, otoh most of my CDs, played through the same equipment, are not harsh and do not glare, so the recording itself is the culprit in my system. Nothing to do with being digital imo.

 

[rebbiputzmaker]

What are some examples of DACs that you've been able to differentiate between using your method?

Over the years various dacs with many different designs. Some converters such as; pm63, pcm1704 (still own one, Lite tube dac), tda1541a, wm8741, ad1865, various ess and akms, (currently have an ak4495 and es9038). I am not saying that the dac chip alone is responsible for the sound, but imo there can be determinable characteristics of dacs, d/s vs multibit.

 

[rebbiputzmaker]

I have quite a few CDs which have a harsh sound I could call glare, otoh most of my CDs, played through the same equipment, are not harsh and do not glare, so the recording itself is the culprit in my system. Nothing to do with being digital imo.

Yes poor recordings are also a problem

 

[andreasmaaan]

Over the years various dacs with many different designs. Some converters such as; pm63, pcm1704 (still own one, Lite tube dac), tda1541a, wm8741, ad1865, various ess and akms, (currently have an ak4495 and es9038). I am not saying that the dac chip alone is responsible for the sound, but imo there can be determinable characteristics of dacs, d/s vs multibit.

Could you give some examples of pairs of specific complete DAC units as opposed to just the chips perhaps?

 

[rebbiputzmaker]

Could you give some examples of pairs of specific complete DAC units as opposed to just the chips perhaps?

Two examples.

Recent , Aune s16- Pacific Valve 62 - Smsl su8.

Older, Adcom 575 - Eastern Electric mini max. - Mhdt labs.

imo the newer ess is better than the older one.

I left out different Ifi dacs, micro and idac2, all with the chip that Thorsten uses. I tend to like them better than Ess dacs.

ymmv