The officially official Devuan Forum!

igorzwx

Member

Registered: 2024-05-06

Posts: 105  

Statistics is a very exact science

Over 5% of the world’s population – or 430 million people – require rehabilitation to address their disabling hearing loss (including 34 million children). It is estimated that by 2050 over 700 million people – or 1 in every 10 people – will have disabling hearing loss.
_https://www.who.int/news-room/fact-sheets/detail/deafness-and-hearing-loss

Linux reached 4.03 percent of global market share in February, according to data from research firm Statcounter. That takes Linux past the 3 percent milestone it reached in June 2023. While we’re still far from the Year of the Linux Desktop, interest in Linux has somewhat grown lately.
_https://arstechnica.com/gadgets/2024/03/linux-continues-growing-market-share-reaches-4-of-desktops/

15 years ago, the data were different: deaf people - 2%, and, correspondingly, Linux Desktop users - 2%.

It seems to be a result of "natural selection" caused by PulseAudio. Therefore, it should not be difficult to verify these data with mathematical modeling.

Last edited by igorzwx (2024-07-03 14:58:19)

aluma

Member

Registered: 2022-10-26

Posts: 646  

Re: Statistics is a very exact science

Ok, this is a fragment of a famous song recorded in 2 bits from 16 standard CD quality.
Can you tell who sings what?
http://ant-audio.co.uk/test/2bit_fragment.zip

igorzwx

Member

Registered: 2024-05-06

Posts: 105  

Re: Statistics is a very exact science

If you open this wave with Audacity, you see nothing there, and Audacity's spectrogram shows nothing.
This may mean, perhaps, that human beings are not expected to hear anything.
However, if you amplify it with Audacity tool (50dB), you can hear the song and a lot of noise.
But it is still below -43dB. You can amplify it more, but then you may have a horrible noise as well.
It is, indeed, a sort of popular song. I do not know "who sings what".
I would not pass a test on popular songs.

aluma

Member

Registered: 2022-10-26

Posts: 646  

Re: Statistics is a very exact science

This old example from one of the developers of digital audio devices was given to an audiophile forum so that the “experts” could feel what they are “broadcasting” about with aplomb. This signal is just above the noise of power supplies and can be heard on a truly high-quality system with a large dynamic range. My case is a Hi-End tube amplifier, speaker with a sensitivity of 93 dB.

igorzwx

Member

Registered: 2024-05-06

Posts: 105  

Re: Statistics is a very exact science

Yes, you may need an amplifier, especially for high frequencies, if you have problems with hearing.

aluma

Member

Registered: 2022-10-26

Posts: 646  

Re: Statistics is a very exact science

This may mean, perhaps, that human beings are not expected to hear anything.

This simply characterizes the quality of the sound card.

You don’t hear 2 bits out of 16 due to the high noise level, and no amount of resampling or software tricks will help. This is common on most computers. In addition to just noise, impulse noise increases jitter.

But, firstly, on the forum we express a subjective opinion, which by definition is undeniable, and secondly, hearing and the emotions it evokes are even more subjective. The brain does its job and, for example, it is impossible to determine the naturalness of reproduction without having previously heard acoustic instruments; the brain has nothing to compare with. And electronic instruments do not have their own “voice”.

Therefore, I do not dispute the opinion of my colleagues about the advantages of different methods of audio reproduction, but my audiophile experience suggests a skeptical attitude towards attempts to judge the quality of the remaining 14 bits.

And I don't consider Bill Reynolds, the developer of PCLinuxOS, to be a "slave", "defective", etc. only because it uses PulseAudio.

igorzwx

Member

Registered: 2024-05-06

Posts: 105  

Re: Statistics is a very exact science

Audacity FFT spectrogram has nothing to do with the quality of my sound card (Intel HDA codec).
The default minimum is -90dB. This is below the audible level.
_https://support.audacityteam.org/audio-analysis/spectral-analysis#plot-spectrum

After amplification, you can hear the song, but the level of rather unpleasant noise is very high.
But this because, perhaps, the song was recorded with the help of PulseAudio.

Last edited by igorzwx (2024-06-29 21:45:31)

aluma

Member

Registered: 2022-10-26

Posts: 646  

Re: Statistics is a very exact science

This is because the noise level of your card is at the level of a 2-bit signal.
In my experiments, I looked at these levels with an oscilloscope at the output of the power amplifier when playing a DVD player CD-RW disc (with a pulse power supply) and a CD player (with an analog one). The noise level was 8...10 times less.
The noise of the switching power supply was a little more and was different to the ear.

Last edited by aluma (2024-06-30 04:39:45)

igorzwx

Member

Registered: 2024-05-06

Posts: 105  

Re: Statistics is a very exact science

It has nothing to do with my sound card. The audio file was amplified with Audacity.
The strong and unpleasant noise can be easily detected on spectrograms.
You can install Audacity (sudo apt install audacity) and verify this.

Perhaps, you do not know how Audacity works. You open an audio file with Audacity. Select the whole file and amplify it. Save the result. Then you can select a time interval and build a spectrogram.

If you think that Audacity's amplifier (with default settings) is producing strong audio distortions, you can file a bug.

> mediainfo 2bit64_fragment.wav | grep Sampling -A1
Sampling rate                            : 44.1 kHz
Bit depth                                : 16 bits

Instead of fooling yourself with an oscilloscope, try to record a low level background noise (or sound) with a portable Olympus Hi-Res Audio Recorder. Then, you can amplify it with Audacity, and see (and hear) the difference.

But why do you need such a strong amplifier? Do you have problems with hearing?

Last edited by igorzwx (2024-06-30 07:42:20)

aluma

Member

Registered: 2022-10-26

Posts: 646  

Re: Statistics is a very exact science

Instead of fooling yourself with an oscilloscope...

?

But why do you need such a strong amplifier? Do you have problems with hearing?

You don't have any idea what you're talking about. Your sound card doesn't have enough gain for you to hear a 5-bit signal, and it's masked by noise.

P.S. I’ll add, maybe it will become clearer.
There is a real sound signal, you do not hear it, the card cannot reproduce it.
This concludes the question.
You can enhance it in an audio editor to make sure that there is something there, but this has nothing to do with real playback.

Last edited by aluma (2024-06-30 09:04:49)

igorzwx

Member

Registered: 2024-05-06

Posts: 105  

Re: Statistics is a very exact science

Your sound card can reproduce the sound, because it has a very powerful amplifier (as you proudly declared before).
I may agree that such sound cards might be very helpful for semi-deaf PulseAudio users.

Last edited by igorzwx (2024-06-30 12:24:54)

aluma

Member

Registered: 2022-10-26

Posts: 646  

Re: Statistics is a very exact science

I already said that you have no idea what you are trying to judge. There are no "very powerful" Hi-En amplifiers. Mine is 4.5 W, with normal listening measured at less than 0.1...0.2 W.
But I had just one question, the answer I received was, you can’t hear. The rest is insignificant.

igorzwx

Member

Registered: 2024-05-06

Posts: 105  

Re: Statistics is a very exact science

Can you hear ordinary music recordings without amplification?
If you feel that you really need an amplifier, you may better visit a doctor. Otherwise, very soon, you may need a much more powerful amplifier.