Audibility thresholds of amp and DAC measurements

[Lambda]

That said, if SINAD is, say, 100, you’re not going to be hearing any issues.

Assuming frequency response is flat and everything else is fine.
Also if the N part is a -100dB (AM modulated) sine wave at 1khz you might hear a issue if you listening with 120dB peak level

There’s a great and understandable desire for some sort of unidimensional metric for ranking components. Unfortunately, that metric doesn’t exist,

Exactly!
Sadly many user here think so... why?

I’d appreciate a constructive answer.

If I’m wrong show me, thanks

So do i.
Why would you mix dB and %?
I also would also appreciate if you do your homework and don't expect to be spoon fed....
Lots of basics you don't seam to know have been answered many times before in this forum.

This isn’t helpful. I’m looking for a scientific answer, because I’m on a scientific website. If there isn’t one then there should be IMO.

Well. The answer is it depends.
it looks like your looking for this:

"This isn’t helpful. I’m looking for a scientific answer, because I’m on a scientific website."
Do your homework, learn the lingo, come back ask a "scientific" question.

statement it sounds like everything that is done here goes out the window.

Why?

can a trained listener say in his mid-20s, hear the difference between .1 THD and .0005 THD?

Now your going from mixing dB and % to not labeling it at all?! "I’m on a scientific website...."

With the right signal an the right type of distortion. Yes.

 

[pma]

I appreciate everyone’s input. Thanks.

To me, SINAD as a single number, without knowledge of the error spectrum, is a pointless number. Please read also a reply by @j_j here:

God of SINAD vs. reality we get from most available music files

9. The notion that you can pick any device ("CD player") you want and it would be transparent is just wrong. You need to measure to make sure this transparency is there. There are a ton of devices in the red category of my SINAD ratings which fall below your criteria: In normal listening...

www.audiosciencereview.com

With a single 1kHz sine, I can hear down to 0.1% H2 distortion in a DBT test. With music and mathematically added polynomial, frequency flat distortion of 5% with decaying energy of harmonics, I have big troubles to tell 5% FS distortion. There are no simple answers to simple questions. I know that you guys would like to have them, but it is not possible. Generally speaking, noises and non-harmonic distortions are more audible than harmonic distortion.

 

[Deleted member 50971]

To me, SINAD as a single number, without knowledge of the error spectrum, is a pointless number. Please read also a reply by @j_j here:

God of SINAD vs. reality we get from most available music files

9. The notion that you can pick any device ("CD player") you want and it would be transparent is just wrong. You need to measure to make sure this transparency is there. There are a ton of devices in the red category of my SINAD ratings which fall below your criteria: In normal listening...

www.audiosciencereview.com

With a single 1kHz sine, I can hear down to 0.1% H2 distortion in a DBT test. With music and mathematically added polynomial, frequency flat distortion of 5% with decaying energy of harmonics, I have big troubles to tell 5% FS distortion. There are no simple answers to simple questions. I know that you guys would like to have them, but it is not possible. Generally speaking, noises and non-harmonic distortions are more audible than harmonic distortion

That’s good info, thank you very much. I’m sure I’ll have more questions after I get a chance to read these replies better.

 

[Sokel]

To me, SINAD as a single number, without knowledge of the error spectrum, is a pointless number. Please read also a reply by @j_j here:

God of SINAD vs. reality we get from most available music files

9. The notion that you can pick any device ("CD player") you want and it would be transparent is just wrong. You need to measure to make sure this transparency is there. There are a ton of devices in the red category of my SINAD ratings which fall below your criteria: In normal listening...

www.audiosciencereview.com

With a single 1kHz sine, I can hear down to 0.1% H2 distortion in a DBT test. With music and mathematically added polynomial, frequency flat distortion of 5% with decaying energy of harmonics, I have big troubles to tell 5% FS distortion. There are no simple answers to simple questions. I know that you guys would like to have them, but it is not possible. Generally speaking, noises and non-harmonic distortions are more audible than harmonic distortion.

Just a note:the sawtooth you were talking about yesterday is ready.

 

[Deleted member 50971]

Here’s one of my main concerns, is the .1% total
harmonic distortion in my 1993 monoblocks an audible concern?

 

[Hayabusa]

So do i.
Why would you mix dB and %?
I also would also appreciate if you do your homework and don't expect to be spoon fed....
Lots of basics you don't seam to know have been answered many times before in this forum.

@Lambda Please stop being so annoying.

THD -> mostly expressed in %
SNR -> mostly expressed in dB

Maybe that's the reason?

 

[pma]

Here’s one of my main concerns, is the .1% total
harmonic distortion in my 1993 monoblocks an audible concern?

Can you post a link to a trusted set of measurements?

Assuming that the amplifier has flat frequency response that is not modulated by speaker complex impedance and does not have excess noise, the value of total harmonic distortion that you have mentioned should be inaudible with music signal.
In case it is a valve amplifier with high output impedance, other parameters than THD may become audible.

 

[sergeauckland]

Here’s one of my main concerns, is the .1% total
harmonic distortion in my 1993 monoblocks an audible concern?

If it's at all frequencies and levels up to clipping, with the load you present, then of no concern.
S

 

[Deleted member 50971]

If it's at all frequencies and levels up to clipping, with the load you present, then of no concern.
S

That’s good news, and there’s nothing worse than finding out one of your amplifiers that you think highly of has audible distortion.

To be honest I could not hear any difference between those and a new set of class D monoblocks that I have. Thanks,

 

[peng]

I’d appreciate a constructive answer.

If I’m wrong show me, thanks

THD in % = 20Xlog(THD%/100) THD in dB

There is nothing technically/mathematically wrong to quote distortions in either way but the dB format that is based on the Logarithmic scale - Wikipedia, that is probably preferred by more technically oriented people, and in my opinion, for good reason.

In your post, you were referencing to THD and SNR, and SNR is typically expressed in dB anyway, so all is good except if both are quoted in dB it may be less confusing to some readers.

Back to your question:

.1% TDH with an SNR of 100 dB

VS

.0005% THD with an SNR of 120 dB

.1% THD is likely audible especially if the distortions are dominated by the odd harmonics especially the higher order ones such as the 5th, 7th, 9th and 11th.

SNR of 100 dB could be audible in a very quiet room, especially if you listen to contents (such as classical music) that have a lot of quiet musical passages. How audible the noise is going to be would also depend on the spl you mostly listen to.

Such comparisons are not always cut in dry as there are many other factors, but in your example, I think it is a cut and dry thing because THD 0.0005% (assuming frequency dependence is not an issue) and SNR 120 dB are really great numbers that imply distortions and noise will not be an issue whereas THD .1% with SNR 100 dB may be audible whether the effects would bother you or not.

 

[peng]

The music, what about the distortion that’s inevitable in all music? I’ve heard a range from 5% to 10% distortion just in the music we listen to.

This is a great point, as others pointed out, if it is in the music itself it really is not consider distortions. However, if the distortions you referred to are introduced by poor quality electronics in the music productions such as amplifiers, and the recording/mastering process, then the final media contents may sound poor to you (to me anyway). Some of such distortions introduced may be intentional too, so it again, depends.....

In my experience, the quality of the media contents are more important than the overall distortions of my signal chain components as long as it is <0.01% or even 0.05%.

 

[Curvature]

odd harmonics

This really has no bearing on audibility if you consider how chaotic the spectrum looks when distortion is applied to music. What really matters is if the distortion falls into the same critical band or not, and how much energy it represents. The higher the order, the more likely distortion can be heard.

 

[RichB]

For amplifiers, reactive loads can affect the distortion levels.
SINAD is useful and I select amplifiers with SINAD headroom, where in the past I select amps with power headroom to limit distortion effects.

Anecdotally, two amplifiers with audible difference were observed in my brother in laws Revel F206 systems were the Pioneer SC-07 (ICE amps) and the Outlaw M200 amps.
To me, the SC-07 always sounded flat and the Outlaw M200 got harsh driving the F208s in the 80 to 90 dB range, when compared to the Parasound A21 in listening tests.

The Outlaw is a class-G and I suspect the music that crossed the rail switching was causing audible issues.

The SC-07 measures well at Sound and Vision:

Audibility thresholds of amp and DAC measurements

That said, if SINAD is, say, 100, you’re not going to be hearing any issues. Assuming frequency response is flat and everything else is fine. Also if the N part is a -100dB (AM modulated) sine wave at 1khz you might hear a issue if you listening with 120dB peak level There’s a great and...

www.audiosciencereview.com

Response from the multichannel input to the speaker output measures –0.07 decibels at 10 hertz, –0.01 dB at 20 Hz, +0.30 dB at 20 kilohertz, and –1.57 dB at 50 kHz. THD+N from the CD input to the speaker output was less than 0.006 percent at 1 kHz when driving 2.83 volts into an 8-ohm load. Crosstalk at 1 kHz driving 2.83 volts into an 8-ohm load was –100.72 dB left to right and –101.73 dB right to left. The signal-to-noise ratio with 2.83 volts driving an 8-ohm load from 10 Hz to 24 kHz with “A” weighting was –105.07 dBrA.

Yet, Audioholics found issues driving full bandwidth loads into 4 Ohms:

Pioneer Elite SC-07 A/V Receiver Review

The Pioneer SC-07 A/V receiver is armed with all of the latest decoding and processing, making it a capable product for next generation high definition audio and video.

www.audioholics.com

The ASR amplifier test power sweeps have been conducted at different frequencies yielding interesting results that would go completely unnoticed with THD measurements 1 kHz signals. In the future, tests may measure SINAD dynamically into reactive loads and I suspect, interesting results.

Stereophile sometimes include THD with full sweeps at selected power levels into 2, 4, and 8 Ohms which are interesting but not always included.
Curiously, their simulated speaker tests inconsistent scales, sometimes in .25 dB and sometimes in 1 dB increments.

- Rich

 

[pma]

0.1% THD is likely audible especially if the distortions are dominated by the odd harmonics especially the higher order ones such as the 5th, 7th, 9th and 11th.

Only very poor amplifier would have 5th - 11th harmonics of -70 to -60dB. The only way I can imagine is crossover distortion of worst kind.

Normally, an amplifier with -60dB H2 will have 5th - 11th between -110 to -80 dB. I can make you sure you will not hear it on music.

 

[Deleted member 50971]

@Lambda Please stop being so annoying.

THD -> mostly expressed in %
SNR -> mostly expressed in dB

Maybe that's the reason?

It was rather annoying, and I know my terms are not correct, but I didn’t need the door slammed in my face by anyone.

Anyway thank you to the ones that took the time to explain things to me. You guys were very helpful indeed!

Edit;

I’ve learned there is no simple answer, and it seems a lot simpler when Amir is doing the reviews.

 

[Deleted member 50971]

THD in % = 20Xlog(THD%/100) THD in dB

There is nothing technically/mathematically wrong to quote distortions in either way but the dB format that is based on the Logarithmic scale - Wikipedia, that is probably preferred by more technically oriented people, and in my opinion, for good reason.

In your post, you were referencing to THD and SNR, and SNR is typically expressed in dB anyway, so all is good except if both are quoted in dB it may be less confusing to some readers.

Back to your question:



.1% THD is likely audible especially if the distortions are dominated by the odd harmonics especially the higher order ones such as the 5th, 7th, 9th and 11th.

SNR of 100 dB could be audible in a very quiet room, especially if you listen to contents (such as classical music) that have a lot of quiet musical passages. How audible the noise is going to be would also depend on the spl you mostly listen to.

Such comparisons are not always cut in dry as there are many other factors, but in your example, I think it is a cut and dry thing because THD 0.0005% (assuming frequency dependence is not an issue) and SNR 120 dB are really great numbers that imply distortions and noise will not be an issue whereas THD .1% with SNR 100 dB may be audible whether the effects would bother you or not.

Excellent info thank you!

 

[Deleted member 50971]

If it's at all frequencies and levels up to clipping, with the load you present, then of no concern.
S

Great thank you.

 

[Lambda]

Sorry I’m not posting things the way you’re used to

This is the forum where we try to impress each other with how much we know about audio. Heavy use of technical jargon is required for participation. ​

Maybe start with learning what SINAD, THD, Noise, SNR, Dynamic means.

Before you "Demand" a "scientific answer" you would not understand anyways because of

This isn’t helpful. I’m looking for a scientific answer, because I’m on a scientific website. If there isn’t one then there should be IMO.

There is almost never a simple answer. it always is it depends.

If you would ask for the ""scientific answer" of the boiling point of Water. the answer would be it depends.
but if you don't seam to care or know or learn on what it depends. You seam to want the simple but wrong answer 212°F.
We don't know what you don't know but if you start mixing up degree Fahrenheit with degrees Celsius. sure it's not "wrong" they can be converted back and forth

But normally you would not mix them up in one sentence. unless you have a good reason or you don't know what you are talking about.
You could not tell the reason so i assume it's the later.

So is 0,1% total harmonic distortion + Noise good or audible?! it depends. its only the total harmonic distortion PLUS Noise.

Is 99,9% clean water good or save to drink? It depends on whats the other 0,1%
is 99,99% save or 99,999%?
And yes you can have water tarts 99,9% pure and save to drink and water that's 99,99% but will kill you.

 

[majingotan]

Seems like it bad at listening then.
I could tell the difference between many different products even with reasonably high performance.
Now recently I haven't really been able to tell this recent high measuring stuff apart, but that is to be expected when we hit the scientific limits of audible performance.
However below like 105 sinad? You should be able to tell lots of stuff apart with well recorded music.

It could be that. I compared my Chord Mojo 1 output against my tube headphone amp (SET with solid state rectification, guaranteed to have 30-50dB SINAD, typical for a simple SET amp without negative feedback on the outputs) and truth to be told, I seriously don't think I would be able to tell them apart on a DBT level matched test as long as the SPL is within 60-70 dB nominal and the recording is not overly quiet. However, the tube amp helps with Dolby Atmos Blu-ray movie playback with its very wide swing of dynamic range (SPL). With Mojo at 2 blue lights, when action scenes are in play, I can hear it run out of steam (gets fuzzy in bass and mids start to crackle) while the tube amp with plenty of headroom plays it back like how it felt when watching IMAX at the movie theater. Also, certain classical recordings that I want to crank up the volume to 70+ dB on the quiet parts with the crescendo part hitting 100+ dB SPLs, the tube amp certainly has much better control of the Aeon X drivers than the Mojo alone.

In those cases, we're now factoring power limits aside from just pure SINAD DBT volume matched test. I am still firmly confident that I won't be able to pass the strict DBT volume matched testing when the Aeon X is driven within the power requirements for both Mojo 1 and the tube amp (which I like since it has linear max power at 4W RMS per channel for both 32 ohm and 300 ohm headphone outputs)

 

[Jimster480]

It could be that. I compared my Chord Mojo 1 output against my tube headphone amp (SET with solid state rectification, guaranteed to have 30-50dB SINAD, typical for a simple SET amp without negative feedback on the outputs) and truth to be told, I seriously don't think I would be able to tell them apart on a DBT level matched test as long as the SPL is within 60-70 dB nominal and the recording is not overly quiet. However, the tube amp helps with Dolby Atmos Blu-ray movie playback with its very wide swing of dynamic range (SPL). With Mojo at 2 blue lights, when action scenes are in play, I can hear it run out of steam (gets fuzzy in bass and mids start to crackle) while the tube amp with plenty of headroom plays it back like how it felt when watching IMAX at the movie theater. Also, certain classical recordings that I want to crank up the volume to 70+ dB on the quiet parts with the crescendo part hitting 100+ dB SPLs, the tube amp certainly has much better control of the Aeon X drivers than the Mojo alone.

In those cases, we're now factoring power limits aside from just pure SINAD DBT volume matched test. I am still firmly confident that I won't be able to pass the strict DBT volume matched testing when the Aeon X is driven within the power requirements for both Mojo 1 and the tube amp (which I like since it has linear max power at 4W RMS per channel for both 32 ohm and 300 ohm headphone outputs)

View attachment 224336

Well the MOJO doesn't have enough current to drive the Aeon probably. This is what I have found out with my Aeon. Many things can drive it, but the sound isn't full without a really proper high current amplifier.
With the MOJO being a portable device; it likely lacks that current.
For the tube amp; micro details will be lost and everything would be smoother. Some people like this sound but still, it's lacking in detail..
I have noticed for the Aeon though that in order to have good lows and mids; current output needs to be high. Otherwise it will have screeching highs and no body in the mids unless the volume is up way too high.
I specifically use a THX789 as my daily amp for this reason. When I have even been testing with my Geshelli Archel2 pro this last week (due to testing new DACs) I can tell the difference between the Geshelli and the 789.
The difference isn't immediately obvious and it depends on the songs. But a song with a lot of low frequencies will make me turn the volume up and then back down for the highs (sometimes in the same song, which happen a often in trance music) because the lows will be so quiet. The 789 has a more "consistent" volume and I won't need to turn it up or down.
This doesn't happen with every headphone though and the overall resolution of details is good on both amps and dacs (D70S + 789) & (SU-8S + Archel2).
To me as sinad goes down, highs are cut off. Either with the highest frequencies not being rendered properly or being blended into other frequencies. Certain background details from songs are lost the lower you go, or may be harder to make out. I notice also a harshness which can come with some Cymbals or "Slaps" (from electronic music), same with certain violin notes (the violin IMHO is the destroyer of low SINAD devices, as it has looooong notes at sometimes higher frequencies).
Since you have an Aeon X; it should be very similar to my Aeon Flow Closed. Try out what I mentioned and see if you can hear a difference if you like.
There are some cheap switches on Amazon too, where you can switch two amps. However you will need some controlled listening first to be able to distinguish details to hear quality with before you attempt to abx your system.