SINAD Measurements

[scooter]

Wow I'm really surprised and disappointed to discover that almost all my home audio equipment in which I've invested my hard earned cash are actually operating at their clipping limits!
This time I looked down at the specs of my Yamaha RX-V583 AVR. Maximum analog line level input signal is stated as 2.3V, meaning around +9.45dBu (measured at 1kHz, 0.5% THD).
Let's say I want to bypass an internal DAC and use an external one with nominal 2V output, I know it won't be overloading an input channel but is it desired to feed the channels with signal levels near to clipping points? Even if THD will fall to 0.1% due to a 0.3V less from maximum voltage applied to it, I will be still amplifying that noice and distortion from the beginning!
I don't need 0.1% distortion when listening at moderate levels, I want to see this number as low as 0.001 or similar.
Next time I'll think twice before making any purchase in audio equipment, analyze the specs and only after go ahead. It is so important when all interconnected devices are operating within nominal ranges and do not stress each other causing to further worsen their performance...

@amirm if it won't be difficult or much time consuming for you, could you test SINAD for various line out levels in the future? For example at 4, 6 and 8dBu? This will help us to identify how they perform in these regions, and accordingly choose them to adapt to our existing gear.

 

[scooter]

Now back to my Yamaha MG06 mixer. As previously noted, it doesn't have much headroom on analog line input channels (3/4, 5/6), clipping at just +4 dBu (1.258 V).
I'm planning to use them with SMSL M100 which has 2 V output as measured here.
I thought instead of purchasing and using signal attenuators, I could just add 70 kOhm resistors in series with each RCA channels. Cables are cheap, easy to cut somewhere in the middle and solder, isolate and hide behind tv stand.
Considering 10 kOhm impedance of line inputs, this could create 250 mV voltage drop on it - nominal input level for MG06 (-10 dBu).
If I go this route, will there be any chance of signal quality degradation? Pure resistive load resistance does not change according to the signal frequency and it shouldn't modify any sound signature, in theory at least.

 

[RayDunzl]

I thought instead of purchasing and using signal attenuators, I could just add 70 kOhm resistors in series with each RCA channels. Cables are cheap, easy to cut somewhere in the middle and solder, isolate and hide behind tv stand.

If you're going to DIY, I'd use potentiometers, as they are adjustable.

For years I used a pair of 10k pots between the CD player and the power amp, one per channel, to avoid ganged pot volume imbalance.

 

[pma]

Wow I'm really surprised and disappointed to discover that almost all my home audio equipment in which I've invested my hard earned cash are actually operating at their clipping limits!
This time I looked down at the specs of my Yamaha RX-V583 AVR. Maximum analog line level input signal is stated as 2.3V, meaning around +9.45dBu (measured at 1kHz, 0.5% THD).

This is a real issue and I have measured many times that the power amp or AVR was clipping not at the output, but just because it was not able to accept signal level from CD. This is normally 2V, but often 2.3V (as you mentioned) or even more, and many amplifiers and AVRs start to clip when their CD input level is at 2V. I do not understand why producers do not fix it, SE maximum input level should be at least 4V without input clipping. There is an easy cure - build a simple voltage divider and put it in a tiny box before the amplifier.

 

[scooter]

If you're going to DIY, I'd use potentiometers, as they are adjustable.

For years I used a pair of 10k pots between the CD player and the power amp, one per channel, to avoid ganged pot volume imbalance.

Simply I don't want to raise a price. Some fine tuning and precise potentiometers probably won't come cheap, also more complicated to hide.
I've calculated an exact value of resistance needed, so I think no need for potentiometers for now (for future proofing maybe).
So based on your comment I shouldn't worry about quality issues, right?

 

[scooter]

This is normally 2V, but often 2.3V (as you mentioned) or even more, and many amplifiers and AVRs start to clip when their CD input level is at 2V. I do not understand why producers do not fix it

Especially when their own line of CD players have an output level of 2 ± 0.3 V (1 kHz, 0 dB). CD-C600 model for example.

 

[signalpath]

SINAD is fine, but don't overweight THD as an indicator of "audio quality." Noise is by far the more important metric in DACs, etc.. It's been proven (ABX testing) that THD below a certain level is inaudible. I don't remember the exact numbers, but it's somewhere around 0.1% or maybe 0.05%? I think it's a bit anti-scientific for ASR to assume that lower THD assumes "better engineering" or "better sound" or "better" in any manner. In most cases, today's ultra-low THD is achieved by modern IC's using gobs and gobs of intrernal feedback, which (in my experience) often negatively impacts perceptual sonic performance.

This is why THD numbers can be counter-intuitive. For decades, we've used our own bespoke 50V discrete JFET amplifiers (shipped 1,000's). These are straight-ahead J74/K170 class-A discrete amps. THD on these beautiful devices is typically in the .02% range, rising to .08% at extremes. These amplifiers sound better than today's IC opamps which perform in the 0.0001% range (as our pro engineering clients will attest). And yet, we would fall quite low in ASR's THD rankings. That's not helpful, nor is such "ranking" based in any kind of settled audio science. Beyond .05%, it's just a number that doesn't represent anything of audio importance. Wildly low THD numbers simply tell us that the designer is using IC's with massive internal feedback. Big deal.

Anyway, just an observation: suggest that ASR not "rank" (color-code, left-better / right-worse, etc.) performance by THD, at least not below 0.01%. It creates a false sense of qualitative rank for readers who may not understand the deeper issues involved. Stick with the audio metric rankings that actually can make scientifically (ABX) perceptual audio differences, such as noise, linearity, jitter, maybe the 32-tone test, filter responses, etc.. Thanks.

 

[amirm]

SINAD is fine, but don't overweight THD as an indicator of "audio quality." Noise is by far the more important metric in DACs, etc.. It's been proven (ABX testing) that THD below a certain level is inaudible. I don't remember the exact numbers, but it's somewhere around 0.1% or maybe 0.05%? I think it's a bit anti-scientific for ASR to assume that lower THD assumes "better engineering" or "better sound" or "better" in any manner. In most cases, today's ultra-low THD is achieved by modern IC's using gobs and gobs of intrernal feedback, which (in my experience) often negatively impacts perceptual sonic performance.

SINAD is the ratio of signal to noise+THD. In highest performing DACs, it is almost entirely determined by noise, not THD. Often distortion products are at or below -130 dB.

As to your last statement, is there a study backing that? Because I am not aware of any, nor does it match my experience.

 

[signalpath]

is there a study backing that? Because I am not aware of any, nor does it match my experience.

Yes, it was an AES paper years ago, maybe more than one ABX study. JJ might remember.

Can assure you, after 30 years of designing and blind AB listening to world-class audio circuits (over 50,000 channels in use at the highest levels of professional audio, film scoring, classical music, post production, archiving, etc.), THD below a certain threshold in no way correlates with better perceptual audio quality. None. In fact, in many cases, it's just the opposite (see my prior post).

I would encourage ASR to not "qualitatively rank" audio gear based on THD. Giving performance numbers is fine, but use care in making scientifically qualitative inferences that have no basis in science.

 

[digicidal]

At which point does ASR then, in the interest of science, effectively lose the ability to provide any meaningful data for casually interested parties? That's not a snide comment, it's a serious question in this light. I understand the argument and agree for the most part... but we already have a fairly consistent number of "but just tell me which one to buy" kind of posts as it is.

There are a number of discussions revolving around the problems with using SINAD as the primary criteria, but (to my knowledge at least) there hasn't been a viable alternative presented - at least not to any level of consensus. @amirm measures and posts many other metrics - and compares and elaborates upon their audibility in every review so it more comes down to needing a starting point in researching a purchase for most members I think. If nothing is used to separate, then ASR will quickly become simply a pile of data with little use to anyone but the products' engineers themselves... and most of them have "pure specifications" on their own websites.

Just my $0.02... and I probably should discount that some.

 

[DonH56]

Stick with the audio metric rankings that actually can make scientifically (ABX) perceptual audio differences, such as noise, linearity, jitter, maybe the 32-tone test, filter responses, etc.. Thanks.

I followed until this last bit, though would like to hear how negative feedback correlates to worse sonic performance... THD is a measure of linearity, and of course SINAD includes noise, so I am not sure exactly what you are requesting we (that is, @amirm ) to use instead of THD and SINAD? The 32-tone test is also a measure of linearity, and I did not think jitter below a reasonable level correlated well as a distinguishing metric either?

The papers I recall relating feedback to worse performance revolved mainly around inadequate loop bandwidth or poor stability.

Confused - Don

 

[majingotan]

though would like to hear how negative feedback correlates to worse sonic performance...

Global Negative Feedback actually improves SINAD and in no way would cause worse sonic performance. In fact, feedback is used to control the gain and make the amp even more stable (if engineered well)

https://www.vtadiy.com/book/chapter...-tube-amplifier/4-4-global-negative-feedback/

 

[Rja4000]

These amplifiers sound better than today's IC opamps which perform in the 0.0001% range (as our pro engineering clients will attest). And yet, we would fall quite low in ASR's THD rankings.

Hi
Just curious: who are "we" ?

 

[anmpr1]

1) ...THD below a certain level is inaudible. I don't remember the exact numbers, but it's somewhere around 0.1% or maybe 0.05%?

2) I think it's a bit anti-scientific for ASR to assume that lower THD assumes "better engineering" or "better sound" or "better" in any manner.

3) In most cases, today's ultra-low THD is achieved by modern IC's using gobs and gobs of intrernal feedback, which (in my experience) often negatively impacts perceptual sonic performance. Big deal.

4) ...suggest that ASR not "rank" color-code... It creates a false sense of qualitative rank for readers who may not understand the deeper issues involved.

I've labeled your points to make it easier to respond. I hope my highlighting and editing for space has not resulted in misinterpreting your words.

1) On pure tones it's in the 0.1 to 0.2% range. On music it is in the 2-3% (or more range). I've cited tests from the '70s in various places (Stereo Review-Bob Carver experiment and Boston Audio Society report on an Acoustic Research facility demonstration).

2) I've not read on ASR the claim that THD in and of itself (at least at these thresholds) is determinate of sound quality. Is anyone making that claim? On the other hand, low distortion is one aspect of excellence in engineering. As long as the means to achieve it does not cause adverse effects.

3) You give yourself some wiggle room by writing 'in most cases'. Would you say that the low THD Benchmark amplifier negatively impacts its perceptual sonic performance? The amplifier has the lowest distortion of any commercial amp I know about, and outside of a few blog comments here and there, no one of any import has claimed poor sonic performance as one of its attributes. Your use of the word 'big deal' to describe low distortion is meant to cast an aspersion or somehow trivialize the matter. Correct? Yet in engineering circles, the low distortion Benchmark amplifier is considered a 'big deal'.

4) If someone is just going to look at a chart, then they will miss other important stuff. That is true. On the other hand, anyone buying something in the 'green' section will not likely have to worry about sonic performance, although there might be other issues such as availability, build quality and support they will have to consider.

 

[scott wurcer]

I wonder how a "feedback goes round and round" thread would play out here?

 

[signalpath]

Presenting THD/IM/SINAD data is good. We want that. More!

The problem is in using SINAD or THD or IM (beyond a known threshold) as a qualitative measure or "ranking." ASR uses these charts that move from left to right, with left being the ultimate performers (in blue-green) and right being poor performers (in red),

The SINAD chart says "higher is better."

No.

There is no perceptual basis to infer that lower THD (below a known threshold) delivers improved perceptual audio quality. The idea that an audio device with .0001% THD is in any qualitative way "better" than a device with .01% THD is scientifically fallacious. Bottom line: saying "higher SINAD is better" will confuse a large population of ASR readers into thinking (1) it will sound better, or (2) it is better engineered, or (3) it is preferred (etc). None of those statements are necessarily true.

I'm simply asking ASR, for the sake of good science and honest reporting, to represent distortion measurements without qualitative editorial. In fact, this is a a great opportunity to educate the consumer audio public on the limits of what distortion numbers can tell us. Beyond a certain threshold, distortion numbers tell us nothing of qualitative value. My advice would simply be to remove all inferences of "better" (color, ranking, etc.) from distortion measurements. Simple. Honest. Better science.

 

[March Audio]

I don't think Amir suggests higher ranking equals better perceived quality. He is very down to earth on what may or may not be audible and often makes comment to that effect in the reviews. His background has very much been involved with subjective assessment.

He usually refers to the rankings as an indication of the engineering quality.

People do need to read the full text and not just make assumptions from the ranking charts.

 

[signalpath]

I've labeled your points to make it easier to respond. I hope my highlighting and editing for space has not resulted in misinterpreting your words.

1) On pure tones it's in the 0.1 to 0.2% range. On music it is in the 2-3% (or more range). I've cited tests from the '70s in various places (Stereo Review-Bob Carver experiment and Boston Audio Society report on an Acoustic Research facility demonstration).

2) I've not read on ASR the claim that THD in and of itself (at least at these thresholds) is determinate of sound quality. Is anyone making that claim? On the other hand, low distortion is one aspect of excellence in engineering. As long as the means to achieve it does not cause adverse effects.

3) You give yourself some wiggle room by writing 'in most cases'. Would you say that the low THD Benchmark amplifier negatively impacts its perceptual sonic performance? The amplifier has the lowest distortion of any commercial amp I know about, and outside of a few blog comments here and there, no one of any import has claimed poor sonic performance as one of its attributes. Your use of the word 'big deal' to describe low distortion is meant to cast an aspersion or somehow trivialize the matter. Correct? Yet in engineering circles, the low distortion Benchmark amplifier is considered a 'big deal'.

4) If someone is just going to look at a chart, then they will miss other important stuff. That is true. On the other hand, anyone buying something in the 'green' section will not likely have to worry about sonic performance, although there might be other issues such as availability, build quality and support they will have to consider.

1.) Yes, THD perception tests from long ago. Thanks! An old-timer. I'm pretty sure there are also AES papers on the topic. JJ could help here.
2.) See my prior comment. Substituting a .0001% IC opamp for a 0.01% well-engineered class-A discrete FET amplifier is not necessarily an indication of "good engineering." Good engineering uses heroic measures (endless blind AB circuit listening tests) to determine "better" -- not THD numbers.
3.) THD and sound quality are not necessarily correlated. There are high-THD audio devices (say 0.1%) that perform with astonishing transparency and dynamic uniformity, but in ASR's ranking they would be inferred as being in the red zone "poor." I'm simply asking for ASR to eliminate any sort of perceived "ranking" (better / worse) from distortion-related measurements -- and perhaps to use this as a teachable moment to help educate their audio public on the weak bonds between "lower THD" and audio performance.

 

[signalpath]

I don't think Amir suggests higher ranking equals better perceived quality. He is very down to earth on what may or may not be audible. His background has very much been involved with subjective assessment. He usually refers to it as an indication of the engineering quality.

And using THD as an indication of "engineering quality" would be misleading. As if a 0.0001% THD product is necessarily better-engineered than a 0.01% product. No. That's not right.

As I just wrote to anmpr1, substituting a 0.0001% IC opamp ($3) for a 0.01% well-engineered class-A discrete FET amplifier ($50) is not necessarily an indication of "good engineering." Good engineering uses heroic measures (endless blind AB circuit listening tests) to determine "better" -- not THD numbers. I remain that ASR's "qualitative ranking" presentation of distortion numbers should be modified.

And, yes, I do think a number of test metrics SHOULD be presented as scientific qualitative statements: noise, linearity, jitter, maybe the 32-tone test, filter responses, etc.. By all means, rank these by color and magnitude, and represent the green-blue candidates as "better."

Thanks.

 

[daftcombo]

THD below a certain threshold in no way correlates with better perceptual audio quality. None. In fact, in many cases, it's just the opposite (see my prior post).

Do you mean that
A) if THD is too low, sound quality begins to worsen
B) many devices which have very low THD happen to have defects that make them sound worse than others
?