A SINAD of 80 or SINAD of 100 Can You Really Tell The Difference?

[Sokel]

I understand scepticism when it comes to audio claims however you have already been provided the measurements that determine differences in between the products. I am not trying to push any particular point of view, only relate my own experience . I am one of the people who who look for 3rd party verified measurements to distinguish between products, which I believe is one of the founding purposes of the Audio Science Forum. In this particular case the measurements came after my listening experience which helped explain the differences and reinforced what I had heard.

The problem with your request to do ABX testing is that apart from the difficulty in arranging it, I haven't seen an ABX test yet where all parties were happy with its parameters or conclusions. And on that subject, it is not for me to prove that item X sounds better than item Y due to Sinad differences . The proof is there in black and white that there are sonic differences. Have you done ABX testing that proves differences in SINAD is inaudible?

Most of the people here have done (I believe and hope) the Klippel test.
It's easy and gives peace of mind.

No need for ABX.

 

[Astrofly]

I understand scepticism when it comes to audio claims however you have already been provided the measurements that determine differences in between the products. I am not trying to push any particular point of view, only relate my own experience . I am one of the people who who look for 3rd party verified measurements to distinguish between products, which I believe is one of the founding purposes of the Audio Science Forum. In this particular case the measurements came after my listening experience which helped explain the differences and reinforced what I had heard.

The problem with your request to do ABX testing is that apart from the difficulty in arranging it, I haven't seen an ABX test yet where all parties were happy with its parameters or conclusions. And on that subject, it is not for me to prove that item X sounds better than item Y due to Sinad differences . The proof is there in black and white that there are sonic differences. Have you done ABX testing that proves differences in SINAD is inaudible?

My question wasn't directed at you personally. It was a general observation that ABX testing is the solution given for many other types of audio issues discussed here on ASR, so why not SINAD, particularly since SINAD is seemingly the one measurement discussed more than any other here? If as you say, ABX testing is difficult to arrange and by implication it is unreasonable to request of someone, how is that any different than requesting ABX tests of someone for DACs, amplifiers, etc? With some of those other audio areas there are existing ABX tests. Are there none here?

In answering your question to me, no, I have not done ABX tests for SINAD because as you mention, it would be difficult to arrange and set up, or at least more difficult to do than I am willing to try and take on, and of course this applies to any other ABX test. However, were I to make observations or claims of fact on audio audibility in certain areas without that data, I would assume that I'd be told I need to do them to verify my claims. Why is that fair to request in those instances and not here?

The larger point I'm getting at is either SINAD is a solved issue or it is not. Either a person can point to a specific db measurement and say, "That's sufficient for human audibility," and then move on, or it's an ongoing technical challenge yet to be solved and new technologies with ever higher SINAD ratings are striving to one day get us there.

My suspicion is that it's a solved issue in terms of human audibility and that obsessing over it is another way guys like to have fun discussing technology and can point to higher SINAD numbers in Product A over Product B as a point of pride in ownership. If one doesn't care about such things though and just wants the most affordable, competent audio product, then it seems focusing on SINAD beyond or even well beyond the point of human audibility and acting like it matters for human audibility is a disservice to that person here on ASR looking for reliable information.

 

[GaryH]

Most of the people here have done (I believe and hope) the Klippel test.
It's easy and gives peace of mind.

No need for ABX.

Nope, that test is invalid. Pkane's Distort + Foobar's ABX Comparator (with verifiable checksum) is the way to go.

 

[Sokel]

Nope, that test is invalid. Pkane's Distort + Foobar's ABX Comparator (with verifiable checksum) is the way to go.

If think I'm well aware of both

 

[fpitas]

As a practical matter, the SINAD/THD+N meter doesn't care if there is or is not a signal. It simply subtracts a tone that you give it and what remains, which in this case is all noise, is reported as THD+N relative to whatever signal level you would have played, but didn't.

Point taken. Although as an example, when you take data on SINAD it is relative to the power output. I'm not sure how you extrapolate that graph to 0 power.

 

[tp1]

One point I should mention about my observations of differences between equipment (which I put down to Sinad differences) is that I am running a 12 channel system which means that most channels are playing low level sounds for most of the time. Low level sounds such ambient sounds for surround effects. It is at this level that I heard differences what I believed to be due at least in part to Sinad differences. Specifically, more detailed low level sounds which when combined as a whole was pretty hard to miss with a system and room I am very familiar with.

I listen to 2 channel music on another system but I suspect the challenge would a lot harder where you would need to try to spot tiny differences during quieter passages .

 

[lashto]

..
Can anyone with excellent hearing distinguish the difference between SINAD 80 and SINAD 100? I don’t want to buy something only to find it is no improvement.
..

almost every hour there is a post like this, where someone asks about the audibility or likeability of THD numbers. And the approx same responses and discussions follow, with all sorts of funny numbers and oppinions.

The Geddes slides linked somewhere above do state very clearly:

THD and IMD have no correlation to the perception of the distortion that they are intended to represent.

Plus many other studies with the same conclusions, starting from ~1930 when the THD measurements were introduced and up to Geddes/Toole/etc ~2010.
There is no magic THD threshold-number where distortion becomes audible or non-audible, there is no THD number that sounds better/good/bad/same/different .. there is only no-correlation.
And any question or 'answer' that contains a correlation is not even wrong.

P.S.
Maybe the ASR website should display a giant header like
"The THD numbers do not corelate with either audibility or preference".
... or maybe put it on the SINAD table .. or maybe someone has a better idea of how to stop those annoyingly repetitive and wasteful 'discussions'

 

[Hayabusa]

almost every hour there is a post like this, where someone asks about the audibility or likeability of THD numbers. And the approx same responses and discussions follow, with all sorts of funny numbers and oppinions.

The Geddes slides linked somewhere above do state very clearly:

Plus many other studies with the same conclusions, starting from ~1930 when the THD measurements were introduced and up to Geddes/Toole/etc ~2010.
There is no magic THD threshold-number where distortion becomes audible or non-audible, there is no THD number that sounds better/good/bad/same/different .. there is only no-correlation.
And any question or 'answer' that contains a correlation is not even wrong.

P.S.
Maybe the ASR website should display a giant header like
"The THD numbers do not corelate with either audibility or preference".
... or maybe put it on the SINAD table .. or maybe someone has a better idea of how to stop those annoyingly repetitive and wasteful 'discussions'

May be I miss the point but if I do the audibility of THD test on the klippel site I can clearly correlate THD numbers with audibility...

 

[Blumlein 88]

almost every hour there is a post like this, where someone asks about the audibility or likeability of THD numbers. And the approx same responses and discussions follow, with all sorts of funny numbers and oppinions.

The Geddes slides linked somewhere above do state very clearly:

Plus many other studies with the same conclusions, starting from ~1930 when the THD measurements were introduced and up to Geddes/Toole/etc ~2010.
There is no magic THD threshold-number where distortion becomes audible or non-audible, there is no THD number that sounds better/good/bad/same/different .. there is only no-correlation.
And any question or 'answer' that contains a correlation is not even wrong.

P.S.
Maybe the ASR website should display a giant header like
"The THD numbers do not corelate with either audibility or preference".
... or maybe put it on the SINAD table .. or maybe someone has a better idea of how to stop those annoyingly repetitive and wasteful 'discussions'

10% THD is audible. 1% is obviously less distorted. Now do THD numbers correlate linearly as in half the THD is heard as half the distortion? No. Nor is one simple number enough. Sometimes that is the breaks. Is .0001% THD audible? No.

You've read the various numbers and parameters. You are going too far with this I do think. Below .1% is almost surely inaudible and even more surely with music. It is about sloppy semantics. Is THD of .0001% better than .01%? Well it is technically less distorted and represents better performance. Neither is audible however. So you could say practically speaking both numbers are perfect. There are lots of fat gray lines in audio.

 

[JSmith]

SINAD and other measurements are about verifying good engineering, not "tell the difference".

If one had a choice of a well measuring device and a rather poor measuring device... why would one purchase the pooly measuring device, regardless of any audible "difference"? Wouldn't one want the best, i.e. minimal distortion and a flat FR in the audible band, assuming functionality is the same?

It is a misnomer IMO to say " if I can't hear it it doesn't matter"... it does matter. It's a poorly designed product... that can't be defended from a design and engineering perspective.


JSmith

 

[beefkabob]

I listened to a DAC that gets a SINAD of 68, dominated by distortion. I don't know what it was actually producing SINAD and distortion wise, as it gets worse when pushed, but I found lyrics almost unintelligible. That was before I knew anything about ASR. I listened to a tube amp whose cousins measure about the same. Again, the singing felt all wrong, hard to understand. The guitar did sound nice, though. SINAD of 80? 100? No idea. Below 70? That's problematic for me, at least with the specific components I listened to. The problem with SINAD is that one SINAD doesn't have the same issues as another SINAD.

 

[007Shorty]

Suppose you have a playback chain consisting of a cd player, amplifier and loudspeakers. The cd player and amplifier each have a SINAD of -80 dB. What would be the total SINAD at the amplifier output? Is it even possible to say without further information?

 

[Blumlein 88]

Suppose you have a playback chain consisting of a cd player, amplifier and loudspeakers. The cd player and amplifier each have a SINAD of 80 dB. What would be the total SINAD at the amplifier output? Is it even possible to say without further information?

Not possible without further info. The biggest problem with SINAD is knowing it is noise and distortion. You need the two separated to have any chance of answering your question.

The main thing is don't worry about this scenario. You are hard pressed not to find a CD player or DAC that isn't around -100 db SINAD. Ditto for many amplifiers. So just don't use those lesser gears in the modern world. That leaves speakers only to worry about. If there is one thing testing shows is good gear means all of it is a solved problem except speakers (and microphones). You don't have to choose, you don't have to wonder, you don't have to do some magic system synergy matching. Just buy good performing stuff and don't sweat anything except the speakers.

 

[antcollinet]

I don't know what it was actually producing SINAD and distortion wise, as it gets worse when pushed, but I found lyrics almost unintelligible.

And that was almost certainly nothing to do with the actual SINAD figure. Much more likely to be related to in room effects, or other frequency response issues.

 

[lashto]

May be I miss the point but if I do the audibility of THD test on the klippel site I can clearly correlate THD numbers with audibility...

that is just one THD number which refers to just one very particular kind of THD (and not even sure if that the Klippel test is only pure THD).

One big problem is that any THD X number can (theoretically) correspond to an infinite number of HD spectra. Different HD spectra. All of them may measure with the same THD number but some will be audible, some not, some will sound 'good', some 'bad' etc...
THD 60 from a speaker is usually very different than THD 60 from a DAC. And two amps with same THD 60 may sound the same .. or not.

The THD number is an (over) simplification of the HD measurement, which 'wastes' almost all the properties of a HD-spectra that may tell you how it sounds. That number simply does not contain "sound quality" info, just "sound accuracy". Kind of like the size of an apple does not contain any taste information .. or how the max-speed of a car does not correlate with its confort.

Hope it's more clear...

 

[lashto]

Not possible without further info. The biggest problem with SINAD is knowing it is noise and distortion. You need the two separated to have any chance of answering your question.

The main thing is don't worry about this scenario. You are hard pressed not to find a CD player or DAC that isn't around -100 db SINAD. Ditto for many amplifiers. So just don't use those lesser gears in the modern world. That leaves speakers only to worry about. If there is one thing testing shows is good gear means all of it is a solved problem except speakers (and microphones). You don't have to choose, you don't have to wonder, you don't have to do some magic system synergy matching. Just buy good performing stuff and don't sweat anything except the speakers.

10% THD is audible. 1% is obviously less distorted. Now do THD numbers correlate linearly as in half the THD is heard as half the distortion? No. Nor is one simple number enough. Sometimes that is the breaks. Is .0001% THD audible? No.

You've read the various numbers and parameters. You are going too far with this I do think. Below .1% is almost surely inaudible and even more surely with music. It is about sloppy semantics. Is THD of .0001% better than .01%? Well it is technically less distorted and represents better performance. Neither is audible however. So you could say practically speaking both numbers are perfect. There are lots of fat gray lines in audio.

was only talking about THD, no noise.
And I wouldn't bet on 10% THD being (always) audible. Also wouldn't bet on -120 THD being (always) inaudible. Those numbers just do not contain 'audibility' information.

Yes, mine is also a bit of an exageration/simplification. But a much safer one and much easier to communicate and to understand for 'everyone'.
We can all just use the THD number for what it is: an engineering metric that measures engineering quality / accuracy ... nothing else should be 'speculated' out of it, that's just a source of neverending 'noise'.

And yes many do sweat too much about those THD numbers. And for the wrong reasons too, it's quite obvious that many/most do not even understand em.
But that is exactly what happens when you give people a top10 chart

 

[lashto]

SINAD and other measurements are about verifying good engineering, not "tell the difference".

yes

If one had a choice of a well measuring device and a rather poor measuring device... why would one purchase the pooly measuring device, regardless of any audible "difference"? Wouldn't one want the best, i.e. minimal distortion and a flat FR in the audible band, assuming functionality is the same?

no

It is a misnomer IMO to say " if I can't hear it it doesn't matter"... it does matter. It's a poorly designed product... that can't be defended from a design and engineering perspective.

fully agree

 

[ItSoundsNatural01]

yes

no

fully agree

why "no" ?

 

[DonH56]

Suppose you have a playback chain consisting of a cd player, amplifier and loudspeakers. The cd player and amplifier each have a SINAD of -80 dB. What would be the total SINAD at the amplifier output? Is it even possible to say without further information?

If you want to know a little more about how to generate the numbers for a chain of components look here:

The Signal Chain: How do noise and distortion propagate through my system?

This article is to explore how noise and distortion change through our system, and how different components can influence the final output. A common question is “If I have DAC A, preamp B, and amp C, where should I set the volume and gain controls for the best performance?” Usually this...

www.audiosciencereview.com

Set aside some time...

 

[antcollinet]

Also wouldn't bet on -120 THD being (always) inaudible.

You'd be in a tiny minority here. I'm pretty sure there is lots of research showing it is.

Even if we can't find it : If you were listening at 100dB (Ear damaging levels) then -120dB would be 20dB below the human audibiity threshold for pretty much anything - even if it isn't totally swamped by the actual music. It'd also be 50dB below the level of your room's noise floor, so probably masked by that alone.

It's like being in a club with music so loud you can't hear your mate shouting right into your ear, and then expecting to hear a down feather gently landing on the back of a kitten.


EDIT - and we have a hypothesis for audibility thresholds here:

Audibility thresholds of amp and DAC measurements

Introduction A recurring theme on ASR is whether or not the various measured qualities of the devices are audible. In this post, I'll present some clear and visual thresholds for when those imperfections can be considered a potentially audible concern. I will not explain the basics of amp/DAC...

www.audiosciencereview.com