Audibility thresholds of S/N - signal to noise - test

[voodooless]

all other aspects being equal?

But obviously, that is never the case. The point is more that these specs should not be a big consideration when choosing equipment.

 

[olieb]

When there are 2 devices to choose from, 1 with more noise and distortion than the other... why choose the 1 with more noise and distortion, all other aspects being equal?

For me the reason would be, that "all other things being equal" does not exist. Once SNR reaches i.e. 120dB other things might become (much) more relevant than improving to 130dB. For a dac: Price, design, firmware, weight, size, case construction, sockets, availability, seller, and so on (the list is very long).

 

[amirm]

An SNR test without the S, is not an SNR test, it would just ne an N-test... they can both be valid tests, depending on what exactly you want to show.

Very little is learned with this test as it only represents the situation for a single pure tone at 200 Hz. Perceptual masking is highly frequency dependent:

Since our hearing is most sensitive in the 2 to 5 kHz, test tones that cover that range will produce the highest threshold for the noise. Yet, that is not what we typically have in music (which is mostly bass heavy).

One simply can't extrapolate from a single pure tone test at 200 Hz, what the requirement would be for audio equipment as claimed by OP. The test I explained is the gold standard for that analysis, establishing the channel noise we can tolerate in all situations for all content.

 

[DavidEdwinAston]

SOTA electronic devices like DACs, preamplifiers and amplifiers have now incredibly high S/N ratio. The limiting factors are in fact S/N of recordings and ambient noise in a listening room. Would you like to check how much S/N you can reliably hear in your listening conditions? Below is a link to test files. They contain a sine 200Hz signal and the same 200Hz sine with added white noise, with rms level of -20dB, -30dB, -40dB, -50dB, -60dB and -70dB relative to 200Hz sine rms level.

https://pmacura.cz/testsignaltonoise.zip

In case you claim to hear -70dB S/N, please add your ABX record.

Forgive me pma. Downloading that zip produces a short even tone which is about 20 seconds long. What am I doing wrong?

 

[Sokel]

What am I doing wrong?

Nothing, that's the test.

 

[DavidEdwinAston]

Thanks Sokel. Okay, it has defeated me. Easily done, I know.

 

[DavidEdwinAston]

Thanks Sokel. Okay, it has defeated me. Easily done, I know.

My wife and I have just listened to the tone, full volume on this s24 ultra. It appears to be the same, from beginning to end.

 

[DavidEdwinAston]

My wife and I have just listened to the tone, full volume on this s24 ultra. It appears to be the same, from beginning to end.

I should add, she wasn't in the kitchen.

 

[Sokel]

I should add, she wasn't in the kitchen.

What you should do is load all the test files at your player and then compare the "pure" one against each of the others starting from the -20 one until you reach the one you no longer hear any noise in the background.

 

[DavidEdwinAston]

What you should do is load all the test files at your player and then compare the "pure" one against each of the others starting from the -20 one until you reach the one you no longer hear any noise in the background.

I am saying that the zip file only produces one playable tone, on this phone, and the Android Tablet which i have tried

 

[Sokel]

I am saying that the zip file only produces one playable tone, on this phone, and the Android Tablet which i have tried

The .zip includes 7 files, all with the same tone but with different level of noise.
The goal is to determine the one that you can not hear the noise behind the tone.

So, you start with the pure one (no noise) and compare it with each one of the rest six beginning with the one labeled -20 , then -30, etc.
When you no longer hear noise you can vote that one.

 

[pma]

I am saying that the zip file only produces one playable tone, on this phone, and the Android Tablet which i have tried

The zip file (with 7 wav files) is linked in the post #1. The post #1 is always relevant.

Once more I repeat the link

https://pmacura.cz/testsignaltonoise.zip

If you are able to unzip 1 file only, then there is a problem with your HW/SW. If you are listening from cell phone speaker or PC notebook built-in speaker, then your vote would be invalid and confusing the issue. Smallish speakers are unable to reproduce the 200Hz base tone properly. The test is intended to be used with hifi speakers or headphones, not with toys.

 

[DavidEdwinAston]

The zip file (with 7 wav files) is linked in the post #1. The post #1 is always relevant.

Once more I repeat the link

https://pmacura.cz/testsignaltonoise.zip

If you are able to unzip 1 file only, then there is a problem with your HW/SW. If you are listening from cell phone speaker or PC notebook built-in speaker, then your vote would be invalid and confusing the issue. Smallish speakers are unable to reproduce the 200Hz base tone properly. The test is intended to be used with hifi speakers or headphones, not with toys.

Finally pma, on my desktop, in Ubuntu, all seven files are playable.(Creative T15 loudspeakers). Brilliant. I will listen again and vote!

 

[telecesne]

...The test is intended to be used with hifi speakers or headphones, not with toys.

I was reading the thread, and then seeing responses about playing on cellphone/iPad, I thought - hmm, I don't think that is going to work too good.

The 'not with toys' got a good chuckle out of me.

 

[telecesne]

"... establishing the channel noise we can tolerate in all situations for all content."

This resonates with me. As engineers and scientists, we ALWAYS seek perfection. 0 distortion, 0 noise, the pursuit of better is always a goal. BUT - sometimes, we have to step back and because of limited resources (time, money), we have to find the 'good enough' spot.

I think that statement sums it up - finding the channel noise that can be tolerated in all situations for all content is the 'good enough' point for those who get caught up in perfection.

There comes a point where there is diminishing return on going beyond the 'good enough'.

Many on this thread have pointed out - once it gets below a threshold, you may measure the change, but you certainly can't hear it - well, unless you are from planet Krypton (Superman reference).

 

[tmtomh]

A lot of misunderstanding here - as always.

Your test is an interesting one, as many of us have said, and I appreciate you creating the files and posting this.

And of course, you cannot explore S/N ratio without having a signal - no one disagrees with that, and contrary to your typically condescending and dismissive attitude, everyone understands that.

And at the same time, such a test is limited, for very good and well-understood reasons that have been noted repeatedly in this thread.

More precisely, since you hold yourself up as an experimental scientist with superior understanding to most other members here, there is nothing about your test that demonstrates with any scientific precision that equipment noise levels of -110dB (or whatever) are excessive, silly, or unnecessary.

To be clear, I am not saying you are incorrect about this - I actually agree that comparing gear based on one having -110dB and the other having -140dB is indeed pointless for music listening purposes - both are functionally transparent when it comes to noise.

However, this is an opinion that is supported only in an informal and very limited way by your test - because -110dB is so far below -60 or -70dB that by common sense -110dB must be far more than we need. But your test does not allow us to draw any scientifically rigorous conclusions about this because we don't actually know what the level is when noise is truly, reliably inaudible. And we don't know that from your test because even if we use the 200Hz tone in your test files as our amplitude-reference signal for a SNR test, the fact remains that when we listen to music at the same volume as we will listen to this 200Hz tone, there are rests in music and there are silences between tracks. And so the experience of listening to the equipment during those moments of silence or near-silence in the source will be part of the same music listening experience as listening to that 200Hz tone.

This is just one reason (or aspect, or example, as one prefers) why your test is interesting and instructive, but does not actually provide the kind of scientific proof of the sweeping claim you are making. This is exceedingly common with your posts: you provide a great practical test or set of measurements - and then you use them to grind your pre-existing axe and make dismissive comments about the entire orientation of this site and @amirm's equipment tests, and in the process you make sweeping claims that exceed what your test or measurements can actually support.

And then when anyone points this out, you throw a little tantrum and say everyone else is too dumb to understand.

It's quite predictable, and rather tedious.

 

[pma]

This is (again) your personal interpretation and I do not feel like arguing in a personal, non technical debate. If you sticked with technical arguing, then I probably would answer.

 

[pma]

Simply put - the signal to which the white noise is added is always the same. The only difference is the level of the added noise, thus only S/N is changed.

 

[pma]

Test results as of now, August 2, 2025, 9:26 am CET

Please note that the test question is How much S/N can you hear from the test files. There is no generalizing, no suggestions to other kinds of test base stimulus. The results apply to the 200Hz tone used and the level of the added noise.

 

[amirm]

This is (again) your personal interpretation and I do not feel like arguing in a personal, non technical debate. If you sticked with technical arguing, then I probably would answer.

The discussion is technical. You made lofty claims that such a test has a bearing on equipment performance. It does not. It is a random data point with no justification.