From: cr88192@gmail.com
On 1/22/2026 5:31 AM, Tim Rentsch wrote:
> Michael S writes:
>
>> On Sat, 6 Sep 2025 05:28:16 -0500
>> BGB wrote:
>>
>>> Just randomly thinking again about some things I noticed with audio
>>> at low sample rates.
>>>
>>> For baseline, can note, basic sample rates:
>>> 44100: Standard, sounds good, but bulky
>>> 32000: Sounds good
>>> 22050: Moderate
>>> 16000: OK, Modest size, acceptable quality.
>>> Seems like best tradeoff if not going for high quality.
>>> 11025: Poor, muffled.
>>> 8000: Very poor, speech almost unintelligible (normally).
>>> But, it is seeming like a "weird hack" may exist here.
>>
>> 8000 x 8bit (mu-law in USA, A-law in majority of the world) was a
>> standard sampling rate for digital back ends of analog wired telephony
>> for more than 50 years. I didn't check, but would assume that it still
>> is.
>> Most people founded it quite intelligible.
>
> Yes but bit rate isn't the whole story. First the measure is not
> "good sound" but only "understandable sound". Second telephony
> does frequency filtering in a very different way than digital
> audio does. Voices on phones are recognizable but still easily
> differentiable from the original. Music played via phone-quality
> audio sounds terrible.
>
In general, A-Law generates better sounding audio.
But, at low sample rates (eg, 8 kHz), I had noted that ADPCM gives more
intelligible speech than A-Law, even if the A-Law "sounds nicer".
Like, if you want background music, 8kHz A-Law is OK.
But, to understand what anyone is saying, less ideal IME, as it tends to
be more muffled than what I was getting with ADPCM. Contrast, ADPCM is
different, in that while intelligibility is higher, it also has a more
"gritty" sound (particularly in variants at 2-bits / sample).
Had also noted:
Low pass filtering reduces "grittiness" by also negatively effects
intelligibility;
High pass filtering preserves intelligibility but can cause a more
"tinny" sound (say, if one uses a 2kHz high-pass filter, so that the
ADPCM only really encodes the 2kHz to 4kHz band).
In my past testing, seems like 2kHz to 4kHz is the most important band
for speech intelligibility (at least for me). It is improved with the
4kHz to 8kHz band, but it seems like this is less important.
So, ranking bands for relative importance:
< 500 Hz : Mostly N/A (*)
0.5- 1 kHz: 5th (barely noticable)
1- 2 kHz: 3rd (makes sound "fuller")
2- 4 kHz: 1st
4- 8 kHz: 2nd
8-16 kHz; 4th (small effect)
16-32 kHz: Mostly unnoticable
*: Mostly inaudible as sine waves, but still very audible if some form
of square wave. But, it would seem like I may be "mostly deaf" in this
range (I can detect sounds here more through tactile senses than by
hearing them via ears). This effect though more depends on touching a
surface, and the relative properties of the object in question.
Subjective responses from house-cats do seem to imply that to them there
is a more obvious difference between the original audio and high-pass or
band-pass versions.
Like, play a normal version of a song, and they remain calm, but play a
2kHz high-pass version and they dig in their claws and seem displeased
(implying a more obvious difference in the sound of the audio than my
own subjective experience).
>> Certainly more intelligible
>> than cellular telephony, until less then 20 years ago cellular improved
>> a little.
>
> Cell phone audio... even today, ick, double ick, and triple ick.
Cell phone audio:
Pretty much unintelligible.
Mostly warbling noises and hiss...
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
