Audioholics - Stop Chasing SINAD Distortion in Audio Amplifiers

[Music1969]

Funny title of his video because he goes on to say and show the amp that sounded less good had worse N+D performance

 

[restorer-john]

Amateurish video, with poor sound, lighting, and content.

It's becoming a wasteland on Youtube with one guy attacking another, who attacks another etc. Yawn.

 

[dfuller]

View attachment 311965

No, this is true. Beyond a certain point everything is going to be clean to the point of inaudible, so there's little reason to reduce distortion further.

 

[jhaider]

Gene's one of the good guys...

I haven't watched the video and I'm not about to spend my time watching a video about boring solved problems such as audio amplification.

However, if Gene is perpetuating the myth here of amplifier "sound" in reasonably designed amplifiers used within their design limits - as this discussion implies, and as he did previously in a whiny video complaining about how people didn’t like expensive audio stuff or something like that - then that casts some doubt on the above. Not wholly, mind. To state the obvious, he sends James Larson a platform and steady stream of speakers to measure.

 

[amirm]

Speaking of James Larson, he was using a dirt cheap class D amp from ebay/Amazon to drive the speaker for Audiohlics speaker measurements. Turns out the thing has load dependency that was corrupting the high frequency response of the speakers he was measuring! And he has been doing this for a long time until I think last year!!!

 

[Steve81]

Speaking of James Larson, he was using a dirt cheap class D amp from ebay/Amazon to drive the speaker for Audiohlics speaker measurements. Turns out the thing has load dependency that was corrupting the high frequency response of the speakers he was measuring! And he has been doing this for a long time until I think last year!!!

Interesting, though not surprising. IME, Gene never made investments regarding staff measurement gear. It was just whatever we happened to have on hand (if anything at all).

 

[JSmith]

Just leave me alone with my SINAD...

JSmith

 

[Albert Dagger]

Here is our rank order of AVRs:

Notice where Marantz is, and where sister brand Denon is. Entirely different class. So much performance left on the table.

It seems the Marantz AV10 is an exception to this trend, and they have been more transparent than usual with the AV10 and the Denon A1H in sharing their measurements, so that seems to be a good sign for the future! I really enjoyed watching the Masimo videos discussing the AV10 and the A1H.

However, I'm not happy that both brands have only been fully transparent with the measurements of their flagship models. They should have shared YouTube videos or slides with this level of detail for all products as a default. I'm also not sure why they use a-weighted measurements (and Gene's does that too).

Always been a Denon fan, but this time I am seriously considering the AV10 because of its measured performance (and reduced weight compared to the A1H, which is impossible to move around).

 

[SCG]

First off I have great admiration for Amir and this site as he tests so much more that SINAD and I have learned so much. As the OP my intent wasn't to start a ASR vs AH debate or conflict. Prior to knowing about ASR; AH (via measurements) introduced me to some good value brands such as Emotiva, Outlaw, and Monolith that I still have in my systems today.

Thru ASR (via measurements) I learn about good value brands such as Topping, SLSH, and Schiit that I also have in my systems today. I also learned that the Denons AVR that I really like for a number or reasons as a pre-amp, measures good and that their sister company Marantz, who’s AVR that I was thinking about switching to as pre-amp because of XLR in/out (that would help with the signal distance between my pre-amp, amps and subs due to AC power balancing distribution) - measures poorer and is to be avoided with the higher costs.

Amir, John Atkinson, and Gene have all discovered expense gear that didn't measure well and all have exposed high price snake oil hiding in plain sight. I’ve also seen examples of how individually, ASR and AH have gotten manufactures to improve their products by posting their poor measurements. So I never got from the video that measurements are not important, in fact just the opposite.

It also has a good overview of how SINAD is the same as % THD+N and how they correlate, which I don't think most newcomers understand. It also states that he thinks most consumers understand the concept of % THD+N better but he gives both % THD+N and SINAD so people can compare directly 'apples to apples' with ASR measurements.

In addition, it states that past a certain point of SINAD he believes that other things come into play like power, features, warranty, returns, customer service, design etc. - vs getting even lower SINAD.

Case in point, I was looking for a phono pre-amp with a rumble filter that wouldn't add any extra noise/distortion, had a good RIAA curve, and was cheaper than the Fluance turntable I had given them.

After reading Amir's reviews I pick a phono preamp that meet all those requirements but it wasn't one of the highest SINAD scores, as some of those had worst RIAA curves. I later saw someone ask Amir "what was the highest SINAD phono preamp under $100" and below is his response:

SINAD is not very instructive for these amps since it is dominated by noise, not distortion. I say best measurement to look at is the frequency response. That determines coloration it may add to everything.

Which I totally agree with and how I made my prior selection. Though, phono pre-amps are ranked and color coded by SINAD, this is not a criticism as I believe this is for overall website constancy and is easier than RIAA curve accuracy ranking - though a newbie might not know or fully understand that, hence the question above.

I also posted in my OP “Best reply in AH's comment section is this one: "In a recent video Amir talks about how SINAD isn't the end all be all but a good single indicator of quality engineering and thus good amp perform (correlation between the two not causation)." Which I believe AH also stated in his video.

So I think there’s a fair bit in common between ASR and AH and don’t think it’s an ASR vs AH. It's just a reminder to newcomers to look at the overall picture when evaluating equipment and reviews - and that info is at ASR, AH, and at other places on the net.

 

[JSmith]

I'm also not sure why they use a-weighted measurements

It's for the improved numbers most of the time.

A-weighting is a pretty rough way to estimate the frequency sensitivity of every persons hearing as it doesn't take into account the non-linear behaviour of the ear.

JSmith

 

[Overseas]

If the piece of equipment stops working (prematurely) the audio quality drops to zero.

You know, you cannot measure the reliability factor the way you measure.... uhhh, SINAD. Then, what you cannot measure does not affect you, right!?

 

[waynel]

Don‘t care what he says , I’m still chasing it.

Chasing SINAD

Thought I’d share some photos of my main system In its latest iteration. I appreciate good engineering and have strived to optimize measurable performance hence the SINAD chasing. Source : Yamaha WXC-50 in player mode with SPDIF out ( I tried Volumio in the SHD but think it’s not ready for...

www.audiosciencereview.com

 

[GXAlan]

Notice where Marantz is, and where sister brand Denon is. Entirely different class. So much performance left on the table.

Exactly! And that has led Marantz to develop new products like the AV10 which offers more performance too.

Marantz has opinions about the digital filter. If they think that their less transparent approach sounds better, go ahead and give user options. Which they have.

Win-win.

No, this is true. Beyond a certain point everything is going to be clean to the point of inaudible, so there's little reason to reduce distortion further.

Beyond a certain point, everything is inaudible.
But we are only testing 1 kHz at 2 or 4V output voltage for line level products at 5W into 4 ohms for amplifiers.

We really care about 100mV or 500 mV for DACs and the “first watt” is most critical for amplifiers.

Here’s the thing.
1) At these lower voltages, there may be audible differences.
2) You DO need to look at the details, but ON AVERAGE, the gear that does well at 2V or 4V or 5W ALSO does well at the very low voltages or wattage.

So the ranking still works somewhat. HTP-1 being one exception to the trend where it peaks below 4v. The grouping still helps to makes a difference.

 

[Deleted member 21219]

1) SINAD is just one of many parameters that indicate how well (or poorly) an item functions. Each characteristic tells a story by itself, but only by putting them all together can you see the whole story.

2) "Chasing" something in engineering is the way we push into new territory. It's how we advance. The supersonic aircraft of the '50s and '60s, the rise of radical racecars, the experimentation into new alloys in general, the constant experimentation into miniaturization .... these have all led to better products for the common man. No one could predict 60 years ago all the things that we have now, but boundaries were pushed and broken, and pushed and broken again. Now here we are, and I think it's great!

The attitude of those who utilize science is to defy limits. Resting on one's laurels produces nothing.

Jim

 

[MaxwellsEq]

It is also absurd that SINAD doesn't matter. Once you get north of 100 dB, distortion becomes very small and noise dominates. And noise is absolutely a problem

+1 This is an important point. For those who think SINAD of 80 or greater is sufficient, noise starts becoming the dominant characteristic to track and low noise becomes more critical than low distortion.

 

[MaxBuck]

1) SINAD is just one of many parameters that indicate how well (or poorly) an item functions. Each characteristic tells a story by itself, but only by putting them all together can you see the whole story.

2) "Chasing" something in engineering is the way we push into new territory. It's how we advance. The supersonic aircraft of the '50s and '60s, the rise of radical racecars, the experimentation into new alloys in general, the constant experimentation into miniaturization .... these have all led to better products for the common man. No one could predict 60 years ago all the things that we have now, but boundaries were pushed and broken, and pushed and broken again. Now here we are, and I think it's great!

The attitude of those who utilize science is to defy limits. Resting on one's laurels produces nothing.

Jim

All this being the case, perhaps looking separately at THD and S/N, as we all used to do, is the way to go.

 

[RDoc]

I think there might be an argument for an amplifier measure analogous to Olive's preference rating for speakers. While I'm not a tuber (nor cruciform ) it probably would weight odd harmonics more heavily than even since most people seem much more annoyed by odd than even and noise levels would be non-linearly weighted so low levels of noise would be weighted proportionally less than higher levels.
I'm not at all sure what the actual audible effect of amplifier non-linearity vs speaker impedance is but, if significant, that could be added as well.
To be real, this would require some blinded listening tests I imagine.

 

[jhaider]

+1 This is an important point. For those who think SINAD of 80 or greater is sufficient, noise starts becoming the dominant characteristic to track and low noise becomes more critical than low distortion.

That begs the question, shouldn’t SNR be the more closely followed metric? Why pollute what matters (SNR) with what is largely irrelevant in audio electronics except in edge cases (distortion).

 

[egellings]

I got to -120db on dac and pre, - 100db on power amps and gave up wasting time.

After SINAD gets so good that an improvement is no longer audible, I see no point in chasing after better results for a home audio application.

 

[Sokel]

I would argue that SINAD not only counts but that it should also come verified and signed by the manufacturers for each single device.
We've seen enough so to trust no one,the (whatever) advertised specs should follow each device.

If not and the famous "sample to sample variation "excuse is at play then we're good with 20 yo stuff cause that's what some variations do.
Bad news is that for this to happen components,thermals,etc must have very-very small tolerances.

So,measure,measure,measure.