Home > Archive > Compression > March 2007 > Audio Compression
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
|
|
|
| Sir,
I am udhay. I am a student and i am working on Audio compression for
my exam.
I want to know how does the compression take place in audio??
I have doubt in human earing capacity.The maximun hearing capacity of
a human is 20-20khz.
if a take a sample wave file,its coming around 40khz.
what it mean? can anyone help me to clear this doubt..
udhay
| |
| cr88192 2007-03-26, 3:55 am |
|
"Udhay" <udhyakumarm@gmail.com> wrote in message
news:1174883039.401672.90880@p15g2000hsd.googlegroups.com...
> Sir,
>
> I am udhay. I am a student and i am working on Audio compression for
> my exam.
>
> I want to know how does the compression take place in audio??
>
usually, it is converted to some form that compresses better, quantized, and
run through an entropy coder.
> I have doubt in human earing capacity.
it is much more sensitive than one might expect...
>The maximun hearing capacity of a human is 20-20khz.
>
depends, I have sometimes used 20Hz sounds to tweak the results in an
autonormalizer, in part because it can be theoretically loud but inaudible,
so I insert a few loud cycles and make things a little quieter...
> if a take a sample wave file,its coming around 40khz.
>
yes, humans hear about through this range, but this veraies from person to
person (one may have a range of, say, only about 40Hz to 12kHz, and another
may have a lot more).
the reason 44100 is typically used is because of some bizarre effect:
to generate a sound at a given frequency, you need about 2x the sample rate
as that frequency (more is better).
so, a 22kHz sound at 44kHz looks about like:
--| |--| |--
| | | |
| | | |
|__| |__|
and 11kHz looks like:
--\ /--
\ /
\ /
\__/
note that a lot of the detail in what is heard is spread over a wide
frequency range (endless overlapping waveforms), so one tries to preserve
all they can over at least the audible range.
> what it mean? can anyone help me to clear this doubt..
>
> udhay
>
| |
|
| On Mar 25, 11:45 pm, "cr88192" <cr88...@NOSPAM.hotmail.com> wrote:[color=darkred]
> "Udhay" <udhyakum...@gmail.com> wrote in message
>
> news:1174883039.401672.90880@p15g2000hsd.googlegroups.com...
>
>
>
>
> usually, it is converted to some form that compresses better, quantized, and
> run through an entropy coder.
>
>
> it is much more sensitive than one might expect...
>
>
> depends, I have sometimes used 20Hz sounds to tweak the results in an
> autonormalizer, in part because it can be theoretically loud but inaudible,
> so I insert a few loud cycles and make things a little quieter...
>
>
> yes, humans hear about through this range, but this veraies from person to
> person (one may have a range of, say, only about 40Hz to 12kHz, and another
> may have a lot more).
>
> the reason 44100 is typically used is because of some bizarre effect:
> to generate a sound at a given frequency, you need about 2x the sample rate
> as that frequency (more is better).
>
> so, a 22kHz sound at 44kHz looks about like:
>
> --| |--| |--
> | | | |
> | | | |
> |__| |__|
>
> and 11kHz looks like:
>
> --\ /--
> \ /
> \ /
> \__/
>
> note that a lot of the detail in what is heard is spread over a wide
> frequency range (endless overlapping waveforms), so one tries to preserve
> all they can over at least the audible range.
>
>
Sir,
Thanks for your mail. i want to how the audio data are stored in our
PC??
Is there any book for reading the compression and which algorithm is
good for lossless audio compression??
how we can find the low bit rate of an audio file?
udhay
| |
| cr88192 2007-03-26, 8:00 am |
|
"Udhay" <udhyakumarm@gmail.com> wrote in message
news:1174892368.902076.91330@o5g2000hsb.googlegroups.com...
> On Mar 25, 11:45 pm, "cr88192" <cr88...@NOSPAM.hotmail.com> wrote:
<snip>
[color=darkred]
>
>
>
> Sir,
>
> Thanks for your mail. i want to how the audio data are stored in our
> PC??
>
depends.
typically wav files are uncompressed. the wav format describes the basic
layout: bits, channels, sample-rate, ... the audio data is then a single
large chunk (basically, a large array of audio samples).
some formats employ some limited compression.
others, such as mp3, aac, or ogg vorbis, perform a long and complicated
process for transforming the waveform into something very different, and
then back again.
> Is there any book for reading the compression and which algorithm is
> good for lossless audio compression??
>
> how we can find the low bit rate of an audio file?
>
these are open-ended questions.
for lossless compression, about the only real option afaik is some form of
DPCM/LPC+entropy coding (RLE/LZ+huffman, ...).
an example of this approach is flac.
most everything else tends to be lossy...
note that it is difficult to predict how much a piece of audio will
compress, apart from compressing it.
> udhay
>
| |
|
| Sir
I want to know how a audio file is stores internally in our pC??
How to determine a length of an audio file?
Can any one help to show the structre of an audiofile.]
| |
|
|
| Oliver Wong 2007-03-28, 6:56 pm |
|
"Udhay" <udhyakumarm@gmail.com> wrote in message
news:1175058417.038601.295680@o5g2000hsb.googlegroups.com...
> Sir
>
> I want to know how a audio file is stores internally in our pC??
>
> How to determine a length of an audio file?
>
> Can any one help to show the structre of an audiofile.]
http://en.wikipedia.org/wiki/Audio_file_format
- Oliver
|
|
|
|
|