Hello, some off-topic notes: The German "Linux-Magazin" (http://www.linux-magazin.de) features in this months edition a title-story "MP3 Most Wanted!". The Article describes (too shortly) the technique, encoders, (much too shortly) software decoders and (too long) MP3-walkmen. I like the part about encoders best. It first says that there are no free encoders around and that the FhG-Encoder is much too expensive (459,- deutschmarks for linux instead of 110,- dm for the windows-version). And then: "As usual in free software, where there is a will there is a way. Because layer 3 is part of the ISO standards, there is a reference-implementation. it has many bugs and misses some features of modern L3-encoders (Joint Stereo, VBR) and the encoded songs show clearly audible errors at 128 kbit/s. [...] The developers use a gap in judiction [excuse my translation] and distribute patches against the iso-reference, not complete encoders. in the meantime an active developer-scene has formed and there is a project wich has made a promising second start: LAME!" [...] LAME bases on the iso-implementation which it extends with many important features such as VBR and Joint-Stereo. And above, the quality is clearly better than that of the iso-code, because many bugs were fixed and improvements included. Lame is the first free project wich produces acceptable quality at 128 kbit/s! The development proceeds rapidly and there is a new version pactically every weekend. There is a mailinglist [ok, we might expect quite a few mor people in here soon] so everybody might join developing. Main developer Mark Taylor is very helpful and seems to invest a lot of work into this project. There are many co-developers discussing on the mailinglist witch have a very good knowlegde of digital signal-processing. This all gives cause for hope. I cannot at this time recommend LAME for production use without restrictions, but this is only a question of time and maybe it is so at the release of this article [i have to excuse for my translation here!]. The latest version tested by me is v3.28, which gives impressing results already. Only in some special songs there are some artefacts by compression, but these are really tolerable. A definite advantage of LAME: It�s fast! On my Alpha 533 SX164 it�s speed factor is 1.35. On an overclocked Celeron 450 it is even faster. If you want to encode something fast, you should cosider LAME. A further nice feature is the graphical analsysis tool to compare input and output and to analyse already encoded songs. LAME is therefore the only serious free implementation of a MP3-encoder. All others lack in quality, because they are only speed-optimized variants of the ISO-reference (Blade, 8Hz, etc.) [then some sentences about "encode" and commercial products by FhG and Xing] To some to an end, a hitlist of the linux-encoders by quality[...]: 1. FhG-IIS 2. LAME 3. Xing 4. encode 5. BladeEnc, 8Hz [...]" Article by Nils Faerber, Linux-Magazin 11/99, pp 34-46. Translation by me. [everything in sqare-brackets are notes by me] Best regards, Olaf mailto:[EMAIL PROTECTED] -- MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
