Advice to an audiophobe ??
dtimms at iinet.net.au
Thu Dec 25 14:04:30 UTC 2008
William Case wrote:
> On Wed, 2008-12-24 at 08:30 +1030, Tim wrote:
>> On Tue, 2008-12-23 at 10:55 -0500, William Case wrote:
> 1. Gives me a whole range of adjustments for different channels. (I
> assume channels means different sources e.g. Master, Headphone, PCM etc.).
Just to clarify, a source device generates an audio signal( line in, PCM
(pulse code modulation = wav uncompressed audio), that might then get
processed (eg volume control, master, headphone), and then sent to a
destination (often an output device like a speaker connector or
Because a typical soundcard has an internal hardware mixer, it can
usually mix together various inputs (sources) like CD input, mic input
and recorded audio signals, and produce a single output signal (mostly
in stereo=2 channels). When mixing together externally received signals,
no main CPU processing power is used, unless you are trying to record to
hard disk etc.
> 2. gives me two choices and
> 3., 4., 5. gives me only Master.
> Which should I choose and why?
capture means recording - usually from a line in (eg from a vcr or mp3
player etc), or from a microphone. So exclude those from your choice.
> If I should be using HDA NVida (Alsa-mixer), why do I have PulseAudio
Consider pulse audio to be a real-time digital mixer and volume control,
where the audio calculations are performed inside your main CPU. In the
default setup, once pulseaudio has done it's processing, it passes the
result to the alsa driver which outputs the audio data to the soundcard.
The soundcard turns the digital audio data into analog audio signals for
use with amplifier, speakers, or headphones.
Pulseaudio also has enhanced capabilities like remembering that when you
playback with xmms that you like to output via your amplifiers and
speakers, but when you are viewing a flash video, to playback into your
headphones instead, at a different level. Another capability let's you
choose the destination playback device while the material is actually
being played. A third capability let's the output go to an audio device
on another machine. Obviously, this is a bit trickier to set up.
>> These individual mixer input controls should normally be left off if you
>> never use them, as they can each introduce noise (hiss, beeps and
>> burbles, etc.) to the system.
> I will turn them off except for Master and Front. I will experiment
> with PC Speaker. Of course these are only available to me if I use the
> default alsa mixer setting.
It's not a one or other setting, both parts will still be involved;
pulseaudio will process, mix, and attenuate sound signals, whereas alsa
will drive the physical hardware. The setting you are seeing lets you
decide whether to control the physical driver volume levels or the
software generated pulseaudio volume controls. If you mute or turn the
alsa master way down, it wont matter how high you turn the pulseaudio
mixer, since the alsamixer comes after the pulseaudio one in the audio
chain. (also true for the reverse).
If you play back a loud audio file, and turn both the pulseaudio source
and master up full. Then change to the alsa setting. You can then use
the also setting to set up an absolute maximum level that you would want
to hear, by adjusting the master. Then you could go back to the
pulseaudio setting to adjust the playback to a comfortable setting, and
from then on only use the pulseaudio setting.
>>> * How is sound related to video ?
>> Sound is the sound, video is the picture... The question is too vague
>> to be answerable.
In digital format, sound and vision are both represented with digital
1's and 0's. With all video and audio file types, there is a packing
together of the audio and video information into the one file. The
multiplexed file provides information about when to playback each frame
of video in relation to the audio in the file. For example, an mpeg2
(dvd) file might have two frames of video, then 2 of audio, then 1 of
video, two audio in an order to achieve a consistent throughput of audio
and video data.
>>> * Why are there so many files associated with producing sound?
In digital audio, the most basic file type is waveform (.wav), where
each momentary value of audio is stored, on a 1 for 1 basis. Experiments
and calculations can show us that for something we store as quality
musical recording we need to sample that momentary value at 44kHz (times
per second) or higher so as not to disrupt our digital recording with
audio aliases. Since we also seem to enjoy the spatial enhancement
produced by stereo or more channels, the file needs to store both left
and right information. Finally, we found that if we only store the
digital value using a small no of bits per sample, when played back we
hear a harsh, chunky sound, rather than the CD like quality of using 16
(or more) bits per sample. The catch with all that is it takes up a lot
To solve space issues (less a problem now that storage space costs a lot
less), compression schemes were developed. Most take advantage of
reducing the number of channels eg to mono, reducing the sampling rate,
or the number quantizing levels (bits/sample); but this is done in
context of the type of audio being compressed - eg human voices are
typically of lower frequency, and can sampled at a slower rate, and with
The biggest jump in compression was with psychoacoustic modelling, where
it was found that in a complex sound, a listener does not notice that
certain frequency (pitch) sounds become inaudible (or masked) by other
The reason there is so many formats, is because developers were
essentially competing to produce more highly compressed audio files,
without noticeable change in quality, when using a certain type of
audio, over a certain communication medium. Eg: when the fastest home
internet connections were slow modems, compression made it possible to
transmit voice signals over your internet connection. If you tried to
transmit music of higher quality that voice, you would have large audio
distortions that made it difficult to hear the original material.
You might like to play with the audio editor program audacity (perhaps
from rpmfusion if you want to be able to import and save in certain
compressed formats (mp3)). It shows you a graphical representation of
the audio file, and eg lets you choose a zoom, start and stop position,
and just play back small parts of a file, so that you can work out what
the sound "looks" like to a computer.
Hope that helps a bit more ;-)
More information about the users