Re: [Lame-dev] A question about lame technic

[Alexander L. <Alexander@Leidinger.net>]

Quoting Phạm Quang Dương <sol...@gm...> (from Fri, 23 Oct  
2009 14:04:23 +0700):

> Hello Lame developer team,
> First, I present myself, I'm a IT student from Hanoi University of
> Technology, Viet Nam.
> Now I and my group take part in a project of building a website for the
> blind (URL is tamhonvietnam.net but current it's not ready available). At
> this website we delivery the voice books which are recorded by the
> volunteers for the blind without any benefit.
> 1 of my group missions is bulding a web service that compresses/convertes
> the recorded Wave file to Mp3 file. I have searched by Google and then I
> found your project. I find that Lame is a good mp3 encoder and useful for
> our project.
> But we have to prove the Lame power to the project leader. I have searched

I assume the cost matters in this case, and LAME does not cost anything.

If it has to be mp3, the only way open source encoder which matters is  
LAME. A lot of other software (even commercial ones) use LAME, or are  
able to use LAME (some kind of de-dacto standard). Maybe this counts  
as some kind of proof.

There's no _real_ way to prove, you can only take different encoders  
and make listening tests.

Another good format is ogg vorbis, and there are some not so  
widespread formats available as open source, just in case it does not  
need to be MP3.

> but could not find enough information about Lame technic.
> I have a question that where I can find the Lame technic description, encode
> algorithm, what make Lame different with other mp3 encoder...
> E.g when resampling wave file, what base Lame use to decide which sample
> will be removed. How to Lame change the bit rate (bit per sample). In VBR
> what base Lame use to choose the bit rate?...

The psychoacoustic code (it removes frequencies) is the major part of  
each encoder. For the commercial ones you will not find information at  
all, as this is the major part which is responsible for the money  
flow. For the open source encoders (LAME or ogg vorbis, or other  
ones), most of the time it comes down to "read the source", as the  
people are focused on writting the software and do not spend time to  
write documentation.

Bye,
Alexander.

-- 
Innovation is hard to schedule.
		-- Dan Fylstra

http://www.Leidinger.net    Alexander @ Leidinger.net: PGP ID = B0063FE7
http://www.FreeBSD.org       netchild @ FreeBSD.org  : PGP ID = 72077137