The long road

Had just bought another 9 CDs to add to my not-so-small Tango collection recently. As a result, decided once again to seriously explore various options for digitising my collection. There are two main reasons for doing this: weight, and ease of access and search. After all, having to cart around about 50~60 CDs is no mean feat! 🙂

Actually, I had previously converted my somewhat smaller Tango collection into MP3s by using LAME, the free and open-source MP3 encoder that is reputed to have the best quality. However, guess it was more just a try-out and there was no urgent need, so was not entirely satisfied with the result (in terms of compression vs quality). This was despite my using the “alt-present extreme” switch. Each 3-min song took up about 4-5MB.

About a year ago, experimented briefly with Ogg Vorbis, which is supposed to use a lower bit-rate for the same quality as MP3. This meant the 3-min songs took up more like 2-3MB. However, got more and more confused with different versions available for download. In the end, could not bring myself to convert everything to Vorbis simply because of the minimal hardware support – not that I have a hand-held MP3 player now anyway…

Right now, the main consideration is quality since hard disks are so much cheaper.

From the discussions on HydrogenAudio, I have concluded that Musepack (MPC), at relatively high bit-rates, is probably the best lossy audio encoder right now. In fact I had set my mind on using this format to archive my Tango collection – despite zero hardware support for this format – until I discovered lossless compression…

A note about lossless audio compression algorithms. The current state-of-art compression ratio for lossless encoding is about 50% of the original wave file. This means that for a collection of 70 CDs, something like 20GB is needed! However, decided to go down this path anyway because 1) right now, harddisks are relatively cheap; 2) having a lossless collection means I can always transcode to MP3, MPC, etc., when I feel the need.

So, as of now, the softwares I intend to use to digitise my entire Tango collection is:

• WavPack: a hybrid lossless audio compression that has the option of producing a lossy file plus a correction file.
• Foobar2000: an advanced audio player that has a low memory footprint and native support for several popular audio formats.
• MP3Tag: despite the MP3 in the name, this is actually a very powerful universal tagger. WavPack files use APE tags, which are more versatile than the ID3 tags in MP3 files.
• ReplayGain: this is another useful utility that can be found at Rarewaves, which normalises the volume of compressed files without changing the original files.

Furthermore, they are all free!

For the technically inclined, on my P4 2.6GHz with 1GB RAM, compressing a full CD (just under 600MB for my music) into MPC takes close to 15-min, but slightly over an hour into WavPack.