ow does one identify music which has potential of being liked? Music, unlike textbooks, does not contain text or keywords. Its tags are not always valuable either. An interesting paper from Trinity College Dublin describes a method by which music adapts to the preferences of listeners (PDF). However, can this be done purely based on prior data? Data that is provided in advance unlike in real-time? A List of records maybe? Playlists perhaps? We seem to be coming closer to realisation of this idea.

Image similarity measures are one focus point of my research; also sparks to mind is Google’s notion of ‘Similar Pages’. Why not apply similar principles to music? I now collect big daily dumps of music that I listen to (output to files using the following technique ). Bound to each entry is the time when a track started. From this, one can infer which tracks are being skipped. Alternatively, full, raw playlists can be of use and might, in fact, be more manageable as well. By exploiting a large collection of playlists, the nature of the genres can be better understood.

Given all of this data, it can potentially be used for collabortive playlist sharing, somewhat like del.icio.us (see previous reference to del.icio.us with a gentle introduction). Users can then discover other songs they might like based on other people’s playlists. The more data, the more accurate statistics will be. Getting large lumps of input (playlists) is effortless too. Just imagine yourself the scenario:

You can automatically find playlists most similar to yours and recognise the most-played tracks on that playlist. Social software has seen great success recently, so exchange of music preferences and recommendations is probably the way to proceed.

There is a true advantage to having audio stored locally, as in the case of off-line podcasting. There is full control, in general, over the audio data in question. This means that podcasts can be sped up to fit the preferences of the listener. There is an XMMS plug-in called SndStretch (shown to the left) which I absolutely adore. It can speed up audio to be up to twice as fast, without change in pitch. No more squeaky voice due to change in pace. An half-hour show can be listened to it in only 15 minutes. This also works for music, which makes it more entertaining and variable.

Ever wondered how much you listened to a particular song?

XMMS possesses a great extent of flexibility. It provides facilities which, when customised appropriately, can dump complete song descriptions. XMMS can then progressively, accumulatively add song data to log files. Timestamps imply the duration of listening, which lead to possible future extensions, e.g.:

• Automatically parse (scan, hash, then analyse) the log file/s
• Compute an average listening time — that is before skipping to another track — for each individual song
• Erase music that is frequently skipped or just output a ‘prospective deletion list’ to another file

Input/output operations are made quite simple. These can be invoked as command-line strings whenever a new track begins. It is a built-in XMMS plug-in, which can be found under Preferences » Song Change » Configure (see illustration below)

Add the following line to the “Song change” Command field:

(date; echo %s) >> ~/My_playlists/playlist.log

I am providing an example log file, assembled naturally (not artificially. Honest!) over 20 minutes or so. I still need to get rid of the doubly appearances, apparently caused by xmms-kde in my dock (screenshot on the right). As well as giving a minimal user interface, xmms-kde enables control of the player with CTRL+SHIFT+keypad even if the player is neither visible nor in focus.

A music log file results in even more redundant data to store. According to a rough calculations, I will see it growing by 1 megabyte every month or so. Compression, however, should make it only 10-20% of its original size.

UPDATE 26/06/2005: The following command, which can be set to become a nightly cron job, will slice the large log file so that it is composed of daily ‘chunks’:

mv ~/[Path]/playlist.log ~/[Path]/playlist-`date +%Y-%m-%d`.log

This can be extended to form an annual and monthly hierarchy of directories.

See the later extension: producing history spreadsheets

Some months ago, an automatic and effective way of obtaining music was described: no need to run any heavy graphical program; no need to select music (only sources of music of the desired genre); no limits on bandwidth or volume.

Taking this further, Dr. Scott Tobkes explains about his method of bringing commericals to a halt.

Something I’ve had fun doing lately is recording live of archived streaming audio from radio stations around the country

There are innumerable subjects and programs. I record it to mp3 or wav, and listen in my car or PC, while fast-forwarding through the commercials. I know that there are aggregators that will make this a seamless process like Ipodder.

I don’t like to call them podcasts. I use this program…

eer-to-peer (P2P) and the World Wide Web (which we most commonly perceive as hypertext — HTTP) are separate. When one uses an Internet connection to directly communicate with a remote machine, then traffic is uncensored, often not monitored, and there are no restrictions imposed on the exchanged material. This justifies the potential illegality of P2P networking.

There are less controversial ways of obtaining media. When it resides on the World Wide Web and hosted by a trusted source, it is then subjected to copyright laws. For this particular reason, it is safer to download MP3 files from trusted Web sites, many of which distribute music for free in interests of self-promotion. Downloads are also far faster because data is delivered by a Web server. An older item explains how to automatically (READ: recursively) download music from the Web.

I have just discovered an inexhaustible source of music, which is also delivered as an RSS feed. That source is the MP3Blogs Aggregator. You must read this entry on the subject in order for everything to make sense. The gist: infinite amount of freely-distributed music reaching your hard-drive on a daily basis.

Let us consider 3 types (‘families’ or ‘classes’ if you like) of headphones:

Standard arc headphones are not very popular these days. They are relatively gigantic, at least when compared with today’s gadgets. Pros: Often come with a thick single wire that connects to one side only (this minimises wear). Cons: Large and unfashionable, susceptibility to outside noise.

In-ear: Often cheaply made. My last set lasted only 4 days, but it had a nice mechanism for isolating noise. Many headphones these days penetrate the ears, yet research has shown that this is damaging to the ear. When I searched for a headphones replacement 5 days ago I saw an odd in-ear set which would penetrate nearly 1 inch inside the ears. My knees could bend seeing the scary otoscope-like design.

Large headphones: my favourite type, but only suitable indoors for obvious reasons.

The bottom line: unless you foresee yourself an old person who is unable to listen to people at the table, care less about noise reduction. We are yet to discover the damaging effects on a generation of Walkmans, Discmans, Minidisks, MP3 players, etc.

The BBC takes a glimpse at Sony’s new portable MP3 players, which somehow imitate the iPod shuffle.

Unlike hard disk-based players, (Sony’s) flash devices hold fewer songs, using solid state memory rather than a disk.

The Sony line-up includes MP3 players which hold a similar number of songs to the iPod shuffle – about 250 – but the players have much longer battery life.

The home of the Walkman finally realised the importance of this entertainment niche.