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

[pma]

New files

Frankly, I have not been prepared to so many nit-picking comments on inaudible distortion components in the 200Hz base file, especially if the file used with added noise was always the same. As I hate pointless debates and I do prefer doing the job instead of plain debating about nothing, I have prepared a new set of files and measurements, now with digitally generated 200 Hz file without any visible distortion components. The files are located at:

https://pmacura.cz/testgen_signaltonoise.zip

The new base file spectrum is this:

So, now you may "enjoy" the test files without that disturbing, 0.0019% THD distortion. I am curious how it would change the results. I have no power to change post #1, so cannot explain it in the first post of the thread. I could update the original zip file, but it would confuse the issue.

 

[solderdude]

A 70dB limit is always what I found myself too. Everything above that is 'nice to have' and shows good engineering.

One could say that when a DAC's volume control is used as main volume control and the amplifier following it is at max. volume 20dB better S/N ratio for DACs is desirable.
So > 90dB S/N ratio (not muted) is more than enough for home audio.

 

[pma]

A 70dB limit is always what I found myself too. Everything above that is 'nice to have' and shows good engineering.

One could say that when a DAC's volume control is used as main volume control and the amplifier following it is at max. volume 20dB better S/N ratio for DACs is desirable.
So > 90dB S/N ratio (not muted) is more than enough for home audio.

Exactly.

And I prefer operating the DAC at full volume being followed by analog volume control or low noise analog preamp.

 

[solderdude]

And I prefer operating the DAC at full volume being followed by analog volume control or low noise analog preamp.

so do I... I generally set the player to -3dB simply because not all DACs perform optimally at 0dBFS but rather 1 to 3dB below that.

 

[pma]

Same here, also for the level optimisation reason.

 

[danadam]

To me the test just confirms that the instantaneous dynamic range of human hearing is about 60 dB, but that was already commonly known, wasn't it?

I'm not sure what's the relevance of this to DACs though. OK, you need at most 60 dB SNR when "S" is at 0 dBFS. But you also need 60 dB SNR when "S" is at -10 dBFS and maybe -20 dBFS. Then you probably don't need that much anymore when "S" is at -40 dBFS.

So, unless you have some strange DAC that heavily modulates the noise floor based on the signal level, it seems like what you should test is the absolute dynamic range, not the instantaneous one.

 

[tmtomh]

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.

"So much misunderstanding - as always" is also a personal interpretation, which was precisely my point. I am perfectly content for others to read our exchange and decide for themselves what they think.

 

[sceptical1]

so do I... I generally set the player to -3dB simply because not all DACs perform optimally at 0dBFS but rather 1 to 3dB below that.

I thought the reason for analog volume was to protect against the full volume software glitch from your $100 dac damaging your $5000 speakers.

 

[solderdude]

I thought the reason for analog volume was to protect against the full volume software glitch from your $100 dac damaging your $5000 speakers.

Yep, works for that too.