At what point would improvements in distortion specs not be worth paying for?

[dfuller]

Harmonic distortion loudspeaker measurements give only a limited insight into audible problems as only one frequency is excited at one time, multitone distortion gives some more insight and can show the audibly better behaviour of a loudspeaker with more ways when for example deep bass is played next to some vocals.

(I agree, but I only had one of those with MTD)

 

[LtMandella]

You are hardly paying for it in some categories. All you have to do is choose well engineered products. Here is an $80 DAC:



Distortion is already below threshold of hearing.

For a few hundred dollars, you can have incredibly low noise and distortion electronics. So I don't see this having much to do with cost than knowing what you buy based on objective data.

Yes, agreed and that is kind of the point of my question. I bought two of SU1 units based on your reviews and they are my favorite Dacs (and best value in hifi I have ever experienced). And for my balanced preamp and amp just ordered a Topping E70 Velvet, again based on your review.

I guess transducers are the big challenge now? Speaker systems with integral and automated room correction?

 

[ElJaimito]

Will there be a point when reductions in levels of distortion artifacts in various audio products be essentially meaningless? Where improvements will be totally inaudible, almost or completely immeasurable by the best test instruments, and not worth paying for?
And are we already there or close to it?
Probably for transducers we are not close. But for electronics?

For amplifiers, I would argue that point was reached with the Leak Point One long ago. In normal listening, and IMHO (and that of many others), harmonic distortion products less than 0.01% - even 0.1% are totally inaudible. Some transducers are approaching the 0.1% harmonic level at normal listening levels. Unfortunately, the "yuf of today" are being conditioned into the belief that high levels of distortion are normal in reproduction (it's called "saturation", "density". "excitation","tape saturation", "low-fi" and other things). The cure is live concerts with orchestral instruments, or just playing an acoustic instrument and listening! The ear mechanism itself of course produces high levels of distortion at high levels of sound pressure, so it all gets entangled. Too much saturation is like adding sugar to the sacharin to the honey to the xylitol... you get the idea.

 

[Oddball]

All these numbers could make anybody's head spin. Kippel test being as limited as it is - is still indicative of how arrogant we are in asserting our hearing capabilities.

As I understand it though it is the Monte Carlo model so it multiplies the across the chain which could result in some ill numbers if the starting point is too low and multipliers bring it down?

 

[dfuller]

Unfortunately, the "yuf of today" are being conditioned into the belief that high levels of distortion are normal in reproduction (it's called "saturation", "density". "excitation","tape saturation", "low-fi" and other things).

I mean, that's all production side stuff. What happens on the production side doesn't matter if it gets the record where it needs to go. Reproduction is something else entirely.

 

[LtMandella]

For amplifiers, I would argue that point was reached with the Leak Point One long ago. In normal listening, and IMHO (and that of many others), harmonic distortion products less than 0.01% - even 0.1% are totally inaudible. Some transducers are approaching the 0.1% harmonic level at normal listening levels. Unfortunately, the "yuf of today" are being conditioned into the belief that high levels of distortion are normal in reproduction (it's called "saturation", "density". "excitation","tape saturation", "low-fi" and other things). The cure is live concerts with orchestral instruments, or just playing an acoustic instrument and listening! The ear mechanism itself of course produces high levels of distortion at high levels of sound pressure, so it all gets entangled. Too much saturation is like adding sugar to the sacharin to the honey to the xylitol... you get the idea.

I don't think I agree about amp performance. Here is one example: How many amps double there power when impedance is halved from 8 to 4 then 2 ohms? Very, very few can do that. I would bet a paycheck that Leak amp you mention cannot. Yet most speakers exhibit impedance plots with such numbers. And without that ability, FR will not be flat.

In other words, distortion measurements with what parameters? Not with parameters exhibited with actual music playback...

 

[a4eaudio]

...How many amps double there power when impedance is halved from 8 to 4 then 2 ohms? Very, very few can do that. I would bet a paycheck that Leak amp you mention cannot. Yet most speakers exhibit impedance plots with such numbers. ...

I don't know about amps ability to double power, I'll assume you are correct, but...

...."most" implies over 50% of speakers have impedance levels of 8, 4 and then 2 ohms? I'm quite skeptical. I went to the Speaker Reviews, Measurements, and Discussion subforum and went through the first two pages and glanced at the reviews and only one went below 4 ohms, and that one was a min of 3.73 ohms. And the fact that it "touches" 3.73 ohms for a fraction of an octave doesn't seem very concerning.

 

[antcollinet]

I don't think I agree about amp performance. Here is one example: How many amps double there power when impedance is halved from 8 to 4 then 2 ohms? Very, very few can do that. I would bet a paycheck that Leak amp you mention cannot. Yet most speakers exhibit impedance plots with such numbers. And without that ability, FR will not be flat.

In other words, distortion measurements with what parameters? Not with parameters exhibited with actual music playback...

How many amps need to?

• First, we rarely run amps even close to maximum power - typical average might be as high as 1/10th peak power. Pretty much all amps can double from 1/10th rated power to 2/10th rated power .
• Then speakers only dip to very low impedance for narrow frequency ranges. Doubling of power would only be needed if all the spectral energy in the music were focussed on those frequencies. In reality a quite small percentage of total spectral energy is - so nothing like a doubling of power is needed.
• And very few speakers actually dip much below 3 ohms, and even then that should only be from the speakers rated as 4 ohm. 8ohm rated speakers should never drop that low.

So "actual music playback" is far less likely to drive an amp into clipping at those low impedance points - than a swept full power sine wave would.

Finally - one of the lowest cost amps on the market. If you want an amp that can close to double power - even from 4 to 2 ohm - you don't have to pay much:

 

[Heinrich]

In other words, distortion measurements with what parameters? Not with parameters exhibited with actual music playback...

You‘re simply talking about clipping current. We‘ve seen it many times, just before clipping amps work best in regards to SINAD. Back in time clipping detection, indicating it via an LED wasn’t that uncommon. Today we‘ve got so much power available that such is deemed to be unnecessary. And in case, rare overdrive isn‘t that bad either.

Distortion is dominated by speakers. Exact numbers depend on the specialities of room integration. How do you define HD in regard to an impulse, represented by a continuous spectrum? Just curious how you think about it.

 

[LtMandella]

How many amps need to?

• First, we rarely run amps even close to maximum power - typical average might be as high as 1/10th peak power. Pretty much all amps can double from 1/10th rated power to 2/10th rated power .
• Then speakers only dip to very low impedance for narrow frequency ranges. Doubling of power would only be needed if all the spectral energy in the music were focussed on those frequencies. In reality a quite small percentage of total spectral energy is - so nothing like a doubling of power is needed.
• And very few speakers actually dip much below 3 ohms, and even then that should only be from the speakers rated as 4 ohm. 8ohm rated speakers should never drop that low.

So "actual music playback" is far less likely to drive an amp into clipping at those low impedance points - than a swept full power sine wave would.

Finally - one of the lowest cost amps on the market. If you want an amp that can close to double power - even from 4 to 2 ohm - you don't have to pay much:

View attachment 447251

Agreed that with the best new chip amps per your example things are better with respect to power output. It is definitely a brave new world with Class D. Although to be fair, you need to buy two of those amps unless playing very old mono lps
I did buy a V3 based on Amir's review, but not the mono version

And in case, rare overdrive isn‘t that bad either.

Testing a mono pair of newly built PASS ACA amps a couple weeks ago, I hit clipping once (and not playing all that loudly). It sounded like someone was playing a guido in the background. Still hurts me ears even now just thinking about it. Too many years of fighting grit and glare in hifi I guess...

 

[Heinrich]

Agreed that with the best new chip amps per your example things are better with respect to power output. It is definitely a brave new world with Class D. Although to be fair, you need to buy two of those amps unless playing very old mono lps
I did buy a V3 based on Amir's review, but not the mono version

Testing a mono pair of newly built PASS ACA amps a couple weeks ago, I hit clipping once (and not playing all that loudly). It sounded like someone was playing a guido in the background. Still hurts me ears even now just thinking about it. Too many years of fighting grit and glare in hifi I guess...

Pass ACA, c‘mon: https://www.audiosciencereview.com/...pass-aca-class-a-power-amplifier-review.9741/

 

[antcollinet]

Testing a mono pair of newly built PASS ACA amps a couple weeks ago, I hit clipping once (and not playing all that loudly). It sounded like someone was playing a guido in the background. Still hurts me ears even now just thinking about it. Too many years of fighting grit and glare in hifi I guess...

Amir measured that amp at around 7W for both 4 and 8 ohm. Not really that surprising you hit clipping.


Not exactly representative of modern solid state designs.

Oh - it also performs horribly in almost every other metric.

 

[LtMandella]

Amir measured that amp at around 7W for both 4 and 8 ohm. Not really that surprising you hit clipping.


Not exactly representative of modern solid state designs.

Oh - it also performs horribly in almost every other metric.

running a mono pair. but not a high power Class D that's for sure.

 

[ElJaimito]

I mean, that's all production side stuff. What happens on the production side doesn't matter if it gets the record where it needs to go. Reproduction is something else entirely.

Cordially, I disagree. Not that it's production side - of course. But training ears is the critical aspect to audio discrimination. It's hard, ears need all the help and support they can get to discriminate fine detail. The hearing system is highly suggestible, as with most faculties. And training an entire generation to expect distortion as a norm does not help with such discrimination -- IMHO, of course.

 

[antcollinet]

But training ears is the critical aspect to audio discrimination. It's hard, ears need all the help and support they can get to discriminate fine detail. The hearing system is highly suggestible, as with most faculties.

Which all is precisely why we should not be using them to attempt to determine the quality of audio reproduction gear.

Listening to music - sure - that is what they are for, and together with the brain perform a wonderful job.

But measuring stuff - or consistently evaluating small details and differences - no, we should leave that to the measuring instruments, like we've been doing for literal millennia, starting with things like the Nuppur Cubit. Our senses are just no good at performing measurements accurately and consistently.

 

[Heinrich]

Our senses are just no good at performing measurements accurately and consistently.

What about the sabre-tooth cat? Humanity developed the hearing to what it is for a reason. You need discriminative powers for survival. Don‘t feel differences in stereo reproduction, especially imaging, and your genes will eventually perish.

That is what they say. I would say, because the hearing is suggestive, as s/o called out correctly, we simply do not need stereo to be perfect. Only that the imagination should try to make sense of the (musical) content, rather than justifying some lunatic price of a component.

I could, reiterated, design a fine plausibility argument why contemporary amps and DACs are outright perfect for the job, Pass amps excluded. With (scientific) perturbance theory and all, but who would listen (sic!)? Seems to be a strawman argument altogether anyway, to say the system ain’t right. Guys, sometimes I simply don‘t feel to listen to music. In case my system remains cold.

 

[antcollinet]

What about the sabre-tooth cat? Humanity developed the hearing to what it is for a reason. You need discriminative powers for survival. Don‘t feel differences in stereo reproduction, especially imaging, and your genes will eventually perish.

That is what they say. I would say, because the hearing is suggestive, as s/o called out correctly, we simply do not need stereo to be perfect. Only that the imagination should try to make sense of the (musical) content, rather than justifying some lunatic price of a component.

I could, reiterated, design a fine plausibility argument why contemporary amps and DACs are outright perfect for the job, Pass amps excluded. With (scientific) perturbance theory and all, but who would listen (sic!)? Seems to be a strawman argument altogether anyway, to say the system ain’t right. Guys, sometimes I simply don‘t feel to listen to music. In case my system remains cold.

You don't need to be able to hear -80dB of distortion in order to hear a big cat of any kind.

You've effectively made my point for me. Our senses evolved to help us find food and detect predators. None of that has needed the ability to accurately measure stuff.

We can't even eyball something and say (with any accuracy) how big it is.

We can hear something coming, turn to look in the right direction and see how fast it is moving. We can have a good idea how long it is going to take to reach us - and if we have a chance of outrunning it or if we should climb a tree instead. But none of that requires detecting minuscule levels of distortion, noise, or frequency response variation - which are the only things that differentiate half way competent audio gear.

 

[Heinrich]

Our senses evolved to help us find food and detect predators.

I don't feel so. The auditory systems is made to listen to voices, in order to understand language.

But none of that requires detecting minuscule levels of distortion, noise, or frequency response variation - which are the only things that differentiate half way competent audio gear.

My take goes like this.
distortion: added harmonics are overwhelmed by the instruments' overtones**, intermodulation products are alarming, as they point to a secondary process, source unknown
noise: neglectable within a second if random
frequency response variation: expectable, automatically adjusted in-ear

**not so much for percussive instruments with irregular harmonics, but always impulsive, hence my question how to measure HD or an equivalent with impulsive (=broad spectrum) noises