First things first. Media formats are just like condoms. They come in every colour, price, texture and flavour there is out there. Formats are very varied, but they can be easily grouped in two main categories: Lossless and lossy formats.
Lossless. The packing monsters.
Lossless formats, are just as its name says: What you hear is what you get. These formats are powerful, crunch formats that provide 1:1 to sound ratio. Basically, they work grouping together media information, so the size is considerably compressed. The best definition I’ve heard for them, is that they are “packers”.
The algorithm works, groping similar sounds in a digital formats, so the sound given by the source format (i.e. a CD) is the same in both the digital file and the sound recorded into the CD. Examples of this are WMA and ACC in their lossless versions. Of course, there are also open formats, like FLAC and Monkey’s Audio (APE). Almost all files you’ll find in the net, encoded with lossless, will be either of this formats.

FLAC Logo
The first impression may seem all-glorious because of the fact of the 1:1 sound ratio (actually, it sounds glorious), but as everything else it has its downside. First of all, it’s compression ratio is somewhere around 40% and 50%, depending on the album. Although it’s not a halfway bad ratio, the space is an important factor to check.
The average album is written in an average 700 MB CD. The actual effective size may be around 450 MB and 650 MB, depending again, on each album.Given this, we would have an average of 225 MB and 325 MB, for each album. If you have a small collection, lets say, something around 40 albums, you would have about 11 GB, of full lenght, full HD albums.
The sound would be great, but unless you have a big disc, you’ll be out of space in no time. That’s one of the main problems, the other factor, is that unless you replace the firmware of the all-so-known iPod to Rockbox, you won’t be able to carry the FLAC files with you, as Apple is run by ‘meatbags’ and they don’t support FLAC. And believe me, don’t even try to use AAC. You’ll get dissapointed, and it won’t actually cut it.
This may be the case for most of us, but if you run a business based in music, it will be worth the try getting an extreme amount of available space only for music, with FLAC coded files. It’s worth the try.
Lossy codecs. Your entire music collection in your pocket.
You have seen them everywhere. The so-called MP3 is everywhere you look. Media players built in propietary systems, cell phones, audio systems, cars, iPods… Common. There are even a couple of electrodomestics that support MP3 playback by now. Although it’s heavily massified, this doesn’t necessarily means that it’s the best.
Although it may not be the best format, as it’s the most massified, it’s also the most compatible. You can further read abou audio containers at Radified, Hydrogen Audio forums and JTHZ. I won’t write wether it should or shouldn’t be used. I’ll use it due to compatibility, as it will fit anything ranging from standalone or car stereos, iPods and computers.
Moving on, lossy formats rely usually on psycoacoustic models that eliminates segments of audio it thinks that won’t be perceptible to a common user. The information translated into bytes, is then refered as a rate of speed, named Kbps, or Kilobytes per second.
Of course, lossy formats suffer a quality decrease, caused by the compression of the codec. This compression may or may not be perceptible by a set of variables. The quality of the audio system used, the hearing capacity of each person, volume, and other variable such as ambient noice and attention.
Talking a little about MP3
MP3 is the short way of MPEG audio layer 3. It was the first lossy codec developed, which gave birth to a new era of music and the industry of mobile audio. A true technology breakthrough for some, but a degeneration of formats for some others.
The MP3 marked and era. The file sharing era. The time when Napster was the ultimate sharing program that could be seen anywhere in the world is long gone, but other file sharing programs have quickly arrised.

Napster logo
The fact is, that a whole industry was born. Digital media containers, not only MP3, but also other propietary formats such as WMA or AAC, were rapidly introduced to the market, and the mobile media players were born. Lets face it. The so-called iPod couldn’t have even been possible without the sole existence of the MP3.
On the technic part, MP3 codecs are used by all propietary media players. Each players commonly has a “Copy from CD feature”. Most players commonly encode audio in CBR (Constant Bit Rate) with speeds of 128, 168 or 192 Kbps. Of course, these setting can be pointlessly tweaked, to use higher or lower bit rates. Anyway, they mostly produce low quality files.
If you want a 5 minutes personal test, go and grab a copy of Encspot and install it. You will need later if you follow this series of posts to create top-notch files. Basically, Encspot is an audio codec analyser. Open with it any MP3 file you have in your disc, for accurate results, use one that you encoded and one downloaded from a P2P network.
Encspot’s output will be the codec used, Kbps rate, method used, size and quality. 128 Kbps and lower rates will commonly be marked as low quality. I have found files encoded with Xing and FastEnc at 192 Kbps that are marked as low quality.
One of the main objectives of this posts, will be to encode high quality MP3 files, using LAME and the RazorLAME interface.

Lame encoder logo
The way of the MP3
MP3 can be coded in three ways. They’re CBR: Constant Bit Rate, ABR: Average Bit Rate and VBR: Variable Bit Rate. Again for further reading of this methods, check the sources stated before. For this tutorial, I’ll be using VBR with quality level 3.
The VBR works in a very particular way. The best way to exemplify it, is with a scattered graph.

Graph CBR vs VBR
When you use a CBR method, lets say, 160 Kbps, each part of the track will be encoded at 160 Kbps. As seen in the scattered graph, the bigging of the track increases from absolute silence to midst complexity. None of the sounds in this area required much quality, so disk space is wasted.
Following, at the midst of the track, a peak is created. This may be caused by a raise of the track’s volume or complexity, between a number of reasons. This peak needs way more quality than the previous area and than the supplied, resulting in quality loss. It can be heard in the form of volume lack, weak low sound and small frequency high sounds.
Basically, if the graph is outside of the selected speed rate, the area has quality loss, if the graph is inside, it wastes space.
On the other side, VBR uses a different approach. It’s psychoacoustic model will try to adjust the speed rate to the graph. A silence, will be encoded with the lowest rate possible, whereas a complex sound will be encoded with the highest rate possible. Just like CBR and ABR, VBR can be tweaded to attain a good quality and good compress ratio. Personally, I encode my files with the VBR method, with a quality index of 3. The complete use of this will be mentioned in a few posts.
Useful MP3 tools
The main characteristic of man that makes it stand from other species, is that he uses tools. The main characteristic of a media freak and all other species, is that he uses tools to make his audio files the best around.
There are a nice amount of tools out there to make top notch files. It’s not only about quality, but also about the use of ID3 tags to identify them. There are some important tools that stand out for their simplicity and power. Basically, this tutorial covers the use of 5 important tools: The LAME codec, RazorLAME interface, MusicBrainz Picard, VSO Image Resizer and MP3Tag. A short explanation of them:
- LAME codec: The LAME codec is the best MP3 codec in the market. It’s open source, and it’s so good, that there are some companies that had used it for their products. Some of them (Sony..) had used it illegally, thinking that as it’s open it can be used the way they want…
- RazorLAME interface: RazorLAME is an interface coded to work with the LAME codec. It’s specially useful to batch code the files. It’s customization level is very varied, but I’ll only add a line to it in this tutorial.
- MusicBrainz Picard: Ever popped a CD in with a media player? It often shows it’s information like artist, track titles and album name. Wow, it must be black magic! Certaintly, it’s not black magic, but the process is almost magical. Media players do this by submitting track durations to CDDB, and the server sends back the media information that it believes to match. MusicBrainz is a community driven, free database with the same information, working with PUID’s. Picard, is the tagger of choice when it comes to semi-automatic tagging.
- VSO Image Resizer: A simple, yet effective program to resize images. I’ll be using to resize artwork’s for albums.
- MP3Tag: The best tag editor out there. Works at a blazing speed, making everything very neatly. My weapon of choice to add the ‘Genre’ tag and artwork to full albums.
- EncSpot: A piece of software by GuerrillaSoft. Although it’s not a free tool, there is one version (which can still be obtained from internet) that is given with a free license. EncSpot is your analizer of choice. Basically, it will yield you information about the quality of a given file. It’s main feature is tellingĀ you the bitrate, as well as the codec used to encode a file.
All this tools will be used way too much in the next posts. They help to create good quality MP3 files, which at the same time are neatly ordered and compatible with every other player in the market. Stay tuned for the next post “Organization Structure”, on how to make your own folder hierarchy to keep organized your music library.

