Subject: Re: [linux-audio-user] why is my mp3 file bigger than my wav file?
From: Ross Vandegrift (ross_AT_willow.seitz.com)
Date: Sat Feb 02 2002 - 18:25:49 EET
On Sat, Feb 02, 2002 at 04:02:57PM +0200, Kai Vehmanen wrote:
> On 1 Feb 2002, Laura Conrad wrote:
>
> > ecasound -i /dev/dsp -o lingham.wav
> [...]
> > ecasound -i lingham.wav -o lingham.mp3
> [...]
> > -rw-rw-r-- 1 lconrad lconrad 2322432 Feb 1 17:11 lingham.mp3
> > -rw-rw-r-- 1 lconrad lconrad 1163619 Feb 1 17:06 lingham.wav
>
> My guess is that the lingham.wav is in 8bit, mono, 8kHz format. In this
> case the above figures make sense. A 2.3MB mp3 file encoded at 128kbit/s
> contains around 2.5 minutes of audio. Ok, let's calculate how much a
> 2.5min, 8bit/1ch/8000Hz wav file would take:
Nah, lame knows about different wave file formats and will automatically:
1) Resample into a better format for compression, if it's really wacked out
2) Apply low or high pass filters to fit the sample rate into a better subset of
the psychoacoustic model.
3) Examine the wav after 1 and 2 and try to decide if it should use MPEG1 Layer
3 (ie, mp3), MPEG1 Layer 2.5, or MPEG2 Layer 3.
> Mp3 and similar compression mechanisms are good for preserving quality of
> cd-quality (or similar) audio content, while reducing file sizes. In your
> case, the source file already was of low quality.
But shouldn't the MPEG2 or Layer 2.5 encoder have kicked in and realized it was
a low-quality sample?
Ross Vandegrift
ross_AT_willow.seitz.com
This archive was generated by hypermail 2b28 : Sat Feb 02 2002 - 18:15:57 EET