Is it audiophoolia to care about SINAD differences which have no correlation in blind listening tests? H2/H3 distortion 'enriches the sound'?

[MAB]

This is an oddly pedantic reply in about three different ways. It's a quote from a text by Geddes with the date listed. The citation is from MarnixM where I read it. There's no chance of 'misleading' people to think it was an academic peer-reviewed text as part of the quote is written in all CAPs, just missing some emojis. It's not less impactful for that as he's explaining his findings. It's pertinent to the debate as lower order harmonics won't sound different if they were added by a speaker, a DAC, an amplifier or the source. Just a DAC and a solid state amplifier below clipping are not adding any which could be audible, a speaker and not inconceivably certain tube amps could add enough to be audible.

Putting aside your post-editing, odd-citing, and desire to play the eternal devil's advocate in your interactions here.

Good engineering is the goal, even if at times it seems to you it's just for the sake of good engineering. Most of the other audio forums value products that often have poor engineering, relying on the consumer not being able to discern with their ears.

Good engineering actually matters, I just posted noise measurements of several amplifiers with a compression driver, and the measured differences are large and these differences are audible. Two of the amps are modern Class D. Two are good Class A/B. One is a 40+ year old Aiwa micro amp.

Topping PA5 II Stereo Amplifier Review

Please report on your findings. Perhaps if you use the stock psu with a regulator this may help, with sufficient capacitance to prevent voltage sag at high volumes 1 - what makes you think there is voltage sag? 2 - The stock PSU is probably already regulated. 3 - Even if there were voltage sag...

www.audiosciencereview.com

The Purfi and NCore are both inaudible with the passive network. The Aiwa is audible for me at 5cm with the passive network, and audible at over a meter with no network. It's a tweeter, I am 59, YMMV.
Good engineering matters, and is audible in many applications, and driving an efficient driver is just one example.

The problem is you are the one focusing on SINAD. And just the distortion part of it. You ignore or don't realize noise as you mention distortion over and over. You end up with an argument against good engineering, on a site that is trying to steer the industry away from purely superfluous and overpriced engineering. My driver example is real world, and a problem people regularly experience on active speakers.

I also don't accept your position if applied to quality, but that's another topic I think.

 

[BobbyTimmons]

Putting aside your post-editing, odd-citing, and desire to play the eternal devil's advocate in your interactions here.

Good engineering is the goal, even if at times it seems to you it's just for the sake of good engineering. Most of the other audio forums value products that often have poor engineering, relying on the consumer not being able to discern with their ears.

There was no 'odd citing or post editing'. I wrote a quotation, wrote the name of the person I quoted in brackets and even the date of the quote. It's strange anyone would have a problem with quoting someone and writing the name and author of the quote in brackets. It seems you also have no experience with academia or citations? This is all standard practice when you quote someone.

Honestly I think some people here don't understand how quotations work. I didn't say 'Recent studies have clearly shown'. I wrote a quote which says 'Recent studies have clearly shown the human' and attributed the quote to its author and its date of authorship. That's a direct quote from 'gedlee'. I'm not citing any studies, I'm citing the quote that says 'Recent studies'.

I'm talking about SINAD differences. I wrote it at the top of the post. Would you say inaudible differences in SINAD are important in music reproduction equipment because of 'good engineering'? Engineering serves a purpose. In this case it's audibility, music reproduction. If you want to judge engineering in terms of SINAD criteria which don't influence the performance of the product that is like saying measuring the interior decoration of speaker cabinets is measuring 'good engineering'. The main consumer part of the industry settled on good engineering with SINAD in DACs and amplifiers years ago. Distortion fell far below audibility.

 

[DanielT]

The best sound source I have ever heard was the mastertape from a reel to reel taperecorder from Revox.

According to Amirms reviews, such a recorder ( of similar quality ) has a SINAD about 50 .
A turntable has a SINAD of 30 , and many people like the sound from a good vinyl record more than a high res 24 bit 96 kHz recording.

So, its clear that the SINAD has no correlation at all regarding the perceived soundquality, and we seems to like distortion .

Edit: the noise floor from many good microphones are about -80 dB , and theres many of them used in the studio.

Perhaps it was a good recording with a lot of dynamics that is the explanation?
It doesn't matter if you have the absolute best lossless streaming if the recording looks like a compressed sausage. Incidentally, one reason why I sometimes look for old CD recordings when I visit flea markets. Those that are not exposed to lodness war. Now, I'm not the first to look for these nuggets, so they're hard to find these days.

Otherwise, I think you are on the right track. SINAD 45, around that I think the limit of audibility lies. It depends on the music. Incidentally, about where many speakers lie in distortion, around 0.5%.
Perhaps it would be easier to detect distortion with ultra-low distortion speakers, such as electrostats for example?
It also varies a lot from person to person. 50% distortion at 15 kHz for a 20 year old vs the same for a 50-60 year old and the results would differ, just to show an extreme.
A person with long experience working professionally with detecting differences, (designer of amplifiers/speakers), may, on the other hand, be able to easily identify 45 SINAD. That's my guess anyway.


Edit:
Audible noise, grounding problems, humming transformers, amps driven into clipping for example are , on the other hand, something that is of course damn annoying.
It can also be annoying for those not interested in HiFi, if it becomes clearly audible. Think old crap low powerd car stereo where the volume is turned up really high so the amp is clipping so the sound cuts, and screams your ears. Nobody likes that. The same sound level with good speakers and amp with low distortion and enough power can, however, vs the mentioned car stereo be a really pleasant experience.

 

[MaxwellsEq]

As others have said, focus on minimising noise. A poor SINAD score may reflect higher distortion or higher noise.

Whilst SINAD is convenient as a ranking tool, I look at the noise behaviour.

 

[Sokel]

I would separate SINAD to it's elements.
Care about distortion?Depends where is that,if it's in the source or early in the chain and will be amplified several times yes,I do care.
Noise on the other hand,I do care,a lot,SNR is the first I look at charts.

What I absolutely care for more than the single elements is the overall performance with everything as sits in the rack and for the shake of sanity I always measure it that way as I have seen snakes coming out with stuff I wouldn't suspect otherwise.

So,it depends,yes,it shows good engineering but what I consider really good engineering is to have the same performance under various conditions (I have learned how to "play" measurements as everyone else who involved in the learning curve so I'm always suspicious ) and most importantly for a long time.
For example I consider good engineering an Alps pot,which I measure and has no difference between channels down to the last mV.

We,older ones are spoiled by devices working nicely for 30 and 40 years,so...

 

[BobbyTimmons]

I would separate SINAD to it's elements.
Care about distortion?Depends where is that,if it's in the source or early in the chain and will be amplified several times yes,I do care.

The SINAD in the products is so good the distortion could be multiplied many times and still wouldn't be audible to young people with sensitive hearing (most of the products had THD less than 0.01%).

 

[Sokel]

The SINAD in the products is so good the distortion could be multiplied many times and still wouldn't be audible to young people with sensitive hearing (most of the products had THD less than 0.01%).

I can always put a 40db SINAD source on the start,you know where that will end?
I can show you when I get back on my rig later today.

 

[Keith_W]

I would separate SINAD to it's elements.
Care about distortion?Depends where is that,if it's in the source or early in the chain and will be amplified several times yes,I do care.
Noise on the other hand,I do care,a lot,SNR is the first I look at charts.

I agree If the SINAD is low enough, I don't care. But if it's high, then I want to know whether it is noise or distortion. If it's high noise, it gets rejected immediately. If it's low noise + high distortion, then I want to know which distortion products. If it's low order (e.g. 2nd, 3rd) it matters less because those are masked and probably inaudible.

 

[MAB]

I'm not sure what you are saying here. You believe inaudible things are important in music reproduction equipment because of 'good engineering'? We are talking about music reproduction.

Sorry you didn't follow.
The amps with lower noise are silent. The amps with higher noise are not silent, they had audible noise. I don't own any extremely noisy amps, but even the Aiwa is unacceptable. Even with the filter, there is audible noise. I showed measurements. Lots of more sensitive drivers than this exist, and worse performing amps.
You should get a sensitive woofer and try a few amps with different noise and see if the pursuit of low noise is unimportant.
Most speaker/amp combos don't really care for sure, but lots do, including the active designs. And many active are noisier than my setup.

The other point was to wonder why you keep talking about one thing, distortion. I don't even know what type you are talking about, like distortion that shows up in TIM measurement from certain amps is really audible, thank god we seem to have left that era behind through this pursuit of low distortion. Or the audible distortion in speakers with high BL non-linearities, for example:

Doppler distortion vs IMD?

purifi-audio.com

Other distortion we can hardly hear.

I guess I don't agree with your characterization or focus on distortion, when many bad amps' noise performance ends up limiting the SINAD score, for what it's worth. Noise is one of the most distinguishing features of some of the really poorly measured amps here. And I showed you with measurements the noise really matters.

 

[BobbyTimmons]

Most speaker/amp combos don't really care for sure, but lots do, including the active designs. And many active are noisier than my setup.

Ok they are run at high enough gain. Most manufacturers of studio monitors don't seem to focus on the hiss as their customers don't seem to care. When actives like Genelecs are repurposed for home listening it becomes more of a problem.

The other point was to wonder why you keep talking about one thing, distortion. I don't even know what type you are talking about, like distortion that shows up in TIM measurement from certain amps is really audible, thank god we seem to have left that era behind through this pursuit of low distortion. Or the audible distortion in speakers with high BL non-linearities, for example:

Doppler distortion vs IMD?

purifi-audio.com

Other distortion we can hardly hear.

I guess I don't agree with your characterization or focus on distortion, when many bad amps' noise performance ends up limiting the SINAD score, for what it's worth. Noise is one of the most distinguishing features of some of the really poorly measured amps here. And I showed you with measurements the noise really matters.

Thanks. My question wasn't clear. I opened the topic saying 'most of the variability found in DACs and amplifiers wouldn't correlate with findings in blind listening tests'. I confused the issue mentioning preferences for low order distortion which also lowers the audibility of some of the components that are measured in THD+N.

Amplifiers. Most of the amplifiers have THD+N less than 0.01%. 'Most of the variability found in amplifiers' is below 0.01%.

DACs. Very few have THD+N more than 0.01%. Almost all of 'the variability found in DACs' is below 0.01%.

We all agree THD+N less than 0.01% is not audible at normal listening levels with normal gain levels. My question 'Is it audiophoolia to care about SINAD differences which have no correlation in blind listening tests?'

If you want my opinion. Our focus could be more on things we know 'correlate with findings in blind listening tests'. For the Fosi ZA3 the THD+N is well below audibility like most modern amps when run at normal gain levels. The more probable thing young people with sensitive hearing might perceive is the load dependency of the frequency response.

 

[daftcombo]

Does Pixelization enrich an image?

 

[JSmith]

As we can see, it doesn't cost more to ensure a device is Hi-Fi/transparent.

Those that try and convince others good engineering is not desirable based on some arbitrary concept of audibility are being disingenuous and likely have some kind of agenda, either commercial or personal.


JSmith

 

[BobbyTimmons]

Does Pixelization enrich an image?

Low order harmonic distortion might be more compared to outlines highlighting an image. Much of Western music is built on harmonics, they sound 'harmonious'.

 

[DanielT]

If you want my opinion. Our focus could be more on things we know 'correlate with findings in blind listening tests'. For the Fosi ZA3 the THD+N is well below audibility like most modern amps when run at normal gain levels. The more probable thing young people with sensitive hearing might perceive is the load dependency of the frequency response.

Deviations from FR , there we don't need much for us to detect that. You can test what "much" here:

Blind Testing a 1 dB Level Difference

Discover the smallest difference in level (dB) you can reliably hear.

www.audiocheck.net

I would never buy an amp that doesn't have pancake flat FR. Also, what happens if you plug a speaker that has a peak of 16 Ohms at 10 kHz into that Fosi ZA3? Will there then be 1 dB deviation from FR in that frequency range? Or more, less? Like I said, I would never buy such an amp. Then I'd rather buy a used class AB amp and live with the fact that it's physically bigger.

 

[DSJR]

The best sound source I have ever heard was the mastertape from a reel to reel taperecorder from Revox.

According to Amirms reviews, such a recorder ( of similar quality ) has a SINAD about 50 .
A turntable has a SINAD of 30 , and many people like the sound from a good vinyl record more than a high res 24 bit 96 kHz recording.

So, its clear that the SINAD has no correlation at all regarding the perceived soundquality, and we seems to like distortion .

Edit: the noise floor from many good microphones are about -80 dB , and theres many of them used in the studio.

Hmm, depends on the generation of the tape played - my CD's actually sounded better/cleaner than the master copies of tracks I played at 15IPS on the Revox B77 I had for a few years decades ago - FAR too many generalisations and time passing now, but it's what I felt at the time.

We loved live UK 'classical' Radio 3 concerts sent via 13 bit FM radio brick-walled at 15khz (noise floor at -70dB or so?) and when the Beeb are on form, it can still 'sound' very good although I'm told the FM feed now comes off the sometimes compromised DAB one (citation needed here - @sergeauckland ?).

Not so much distortion thresholds (so many valve loving 'audiophiles' put up with sinads estimated at 40 - 50dB in their amps) but one thing I'm noticing is that some A-D transfers of favourite recordings haven't set the Dolby A's 'quite right' and pumping occurs at the very tail end of fade-outs, which are clearly heard on headphones or earphones such as I use so much nowadays. Maybe these errors were always there and all but buried in the noise floor? 'Digital Silence' can be discomforting sometimes and I remember my ex-Decca mastering engineer telling me that often, venue 'ambience' was often added (analogue or digital sourced) in between orchestral 'movements' to save a dive into total silence as the tape leader went through to the next section.

Apologies for going off topic...

 

[BobbyTimmons]

As we can see, it doesn't cost more to ensure a device is Hi-Fi/transparent.

Those that try and convince others good engineering is not desirable based on some arbitrary concept of audibility are being disingenuous and likely have some kind of agenda, either commercial or personal.

Ignoring a couple at the bottom the DACs and amplifiers (below clipping) are measuring as transparent. Most of the variability in the SINAD measurements of these products is happening far above transparency.

 

[JSmith]

Most of the variability in the SINAD measurements of these products is happening far above transparency.

Audibility... not transparency to the source in comparison to better measuring devices. Respectfully, this type of discussion has been done to death.


JSmith

 

[Chr1]

I think that the subject of the OP, and as discussed in another thread : Whether some harmonics are actually preferred (double blind), is way more interesting personally. I also suspect that it could also therefore be quite likely that many people that poo poo valve amps, may inadvertently find that they actually prefer them.
Again double blind.
Heresy perhaps!
And just a theory obviously...

 

[Tangband]

Hmm, depends on the generation of the tape played - my CD's actually sounded better/cleaner than the master copies of tracks I played at 15IPS on the Revox B77 I had for a few years decades ago - FAR too many generalisations and time passing now, but it's what I felt at the time.

We loved live UK 'classical' Radio 3 concerts sent via 13 bit FM radio brick-walled at 15khz (noise floor at -70dB or so?) and when the Beeb are on form, it can still 'sound' very good although I'm told the FM feed now comes off the sometimes compromised DAB one (citation needed here - @sergeauckland ?).

Not so much distortion thresholds (so many valve loving 'audiophiles' put up with sinads estimated at 40 - 50dB in their amps) but one thing I'm noticing is that some A-D transfers of favourite recordings haven't set the Dolby A's 'quite right' and pumping occurs at the very tail end of fade-outs, which are clearly heard on headphones or earphones such as I use so much nowadays. Maybe these errors were always there and all but buried in the noise floor? 'Digital Silence' can be discomforting sometimes and I remember my ex-Decca mastering engineer telling me that often, venue 'ambience' was often added (analogue or digital sourced) in between orchestral 'movements' to save a dive into total silence as the tape leader went through to the next section.

Apologies for going off topic...

DSJR - your writings are always relevant and interesting because you know how good it can sound from 2 channels.

Im convinced that our ears like a sound that slightly distorted and noisy, just because of the flawed stereo system . With two channels, two ears and a brain, there is no such thing as accuracy.

Just look at the original Linn klimax DS player that have excellent lundahl transformers on the low distortion lm4562 outputs, making the sound distort even worse than the Rega Dac with its ” discreete ” analog output….. still the sound is considered excellent by most people who hear the player.

 

[antcollinet]

The idea of a 'cleaner' signal is entrenched in audiophile culture. Fundamentally people should do what they enjoy including chasing inaudible distortion measurements in electronic equipment. Just if it's not perceivable is it more rational than fussing about cable lifters? We're talking about things which people enjoy doing just they have no influence on the sound.

From my point of view I want to be certain any noise/distortion is inaudible. I'm confident that (at least for me) distortion anything below around -50dB is.

So to create some significant margin for error - and accounting for the fact that noise and distortion is additive through the chain AND that noise is a factor on top of distortion, I've set my own limit at -80dB.

How this works in reality? I look at the SINAD charts here, and select viable options from the blue and green sections (amps) or blue/green/orange Dacs), and beyond that I don't care about the Sinad "Score". Selection then is based on price, features, brand, after sales, etc etc.