Slashdot Mirror
MP3.com Removes "High-Bandwidth" Streams
mshiltonj writes "I noticed today that mp3.com no longer offers high-bandwidth streams for its genres or stations, although it looks like artists' playlists and individual songs are available in high bandwidth. mp3.com has lots and lots of free music that was free and legal to listen to online, and a good number of my "music bookmarks" were on mp3.com. I'll live (I've still got my favorite stream), but I don't think it's a good sign. Is streaming music doomed to die, not because of RIAA litigation, but because of expensive bandwidth costs?" I don't think bandwidth will be the determining cost - that's a price that has been falling and will continue to fall. But are things like iTunes store the future, or is it streaming?
MP3.com discovered that legal fees and bandwidth costs couldn't be covered by the very very small amount of cash coming in from ads.
Rather than go to a pay model they just decided to drop their higher streams... Maybe they should have had a system where you can pay some negligible fee (25 a year, perhaps) to hear the high bandwidth streams, and the low ones are free?
Some tunes are available only if you pay, but you can stream in hi-fi quality (128 kbit/s). Now, why would anyone pay for the tune, when they can just capture the hi-fi quality stream into an .mp3 file??
Before you call this stealing, think. It's just capitalism in action. Greedy agents acting on behalf of their own interests and agenda. If they can get something for free, they will. Morality has nothing to do with this.
It's business. It's the same thing the companies have been pulling, but now consumers can actually leverage their greediness directly.
Sucks to be the artist, though. But they would make peanuts with mp3.com in any case (been there, done that).
I've found Launchcast is much better than mp3.com for streaming stations, though if you listen to more than a certain number of songs per month (350? 400?) it goes into low quality mono for the remainder of the month.
Highly customizable though.
Rumor has it MP3.com recently laid off 40 people, roughly 15%.
When: May 08 2003
Maybe people just find Kazaa to be so much better.
-- the only thing we have to fear is really scary things
Linking that to the RIAA-hates-file-sharing, no one at RIAA bitched when people were trading tapes, but they get their panties in a wad over trading high-quality rips and copies. Maybe if everyone swapped 96k MP3s they wouldn't bitch as much... or maybe they would anyway.
Are we just greedy about quality? I think the mindset is something like "why shouldn't I have 128K stream?" I guess the spread of broadband is the answer. More multiple simultanious streams causes the server to split bandwidth down too low to actually stream. Funny that the spread of cable/DSL broadband is making cable/DSL broadband more obsolete.
I'd be interested to know why bandwidth is so expensive. Anyone here have any concrete answers? Surely with more and more of the world being connected, the cost should go down slowly, as there is just maintainance costs to deal with. Are we still paying for the initial road digging, satellite launches etc?
I used to the the engineer of a now defunct online radio service in the US (epointradio.com), way back in 2000-2001, and worked in various studio enviroments before that.
We streamed 24 and 56k MP3 and it sounded great. The secret to making it sound good was simplicity, just as with good home audio.
I used very little signal processing on the music end of things; arguably, none at all. I simply selected a sound card with very pleasant input clipping characteristics (read: free emergency limiting) and accurate digital loopback (for the controlroom monitors), followed by 1 bit of software gain reduction (to preclude the listeners' sound cards from reacting inconsistantly to the sometimes-peaked audio). The resultant bits were fed to LAME, with very carefully-selected parameters, and presented to the world.
Of course, we used REAL, COHERENT DJs instead of a mindless automatron playlist-spewer. It's non-trivial to select music sets which maintain consistancy, but by no means difficult. Even the "creative people" seemed to do a good job at it, and were able to handle manual gain management justfine with the live feedback they got from their 'phones or a pair of NS-10s.
Automatic gain control is useful for uncontrollable situations. I used some rather complicated compression, limiting, gating, and sidechain filtering for the mics during the talking parts of the program in order to keep things quiet and consistant. But even then, it was as little processing as I could get away with. I was faced with between 6 and 12 shifty people in a room full of live mics, any (or all) of whom might start talking at any time. I don't have that many hands.
On the other hand, modern music is generally already compressed and Fletcher-Munsenized to hell and back in the mastering process. It doesn't need any more "help." And that which still contains some element of dynamic content, such as Tool's Lateralus, uses it so artistically that it would be sinful to throw any of it away.
An engineer (or likely, several of them) spent hours, days, or weeks tweaking -that- -one- song to perfection. By homogenizing it with your ill-concieved one-size-fits-all processing, you've not only destroyed their work, but done a disservice to your listeners.
Compression was introduced to radio as a counter to road noise and static - admirable goals. Sometime later, it was used to keep levels somewhat consistant for lazy DJs, and kicked the processing up a notch. More recently, some marketing fucktard decided that it could be used to better compete with nearby stations, even to the point of modulating over them, and we ended up with the wall of multiband-limited pile of compressed shit that we get when we turn on the radio now.
These issues do not exist in the world of Internet broadcasting, where instead of static we get nearly limitless dynamic range. People listen to unprocessed music all the time at their PC, and are accustomed to hearing it that way, dynamic and timbral nuances intact.
And, unlike the stereo in your car, here the listener has control. If they want homogenized pre-processed shit, they can download (or merely enable) a plugin for it. Realplayer can do this, WMP can do this, along with Winamp and XMMS.
Why fuck it up in advance when they've got such a diverse array of tools to make things sound more to their liking right at their fingertips, if that's what they're after?
If you want consistancy, strive to make it sound the same from your stream as it does from the MP3 that they've got on their hard drive, put an ounce of effort into the mix, and keep your monitor setup appropriate to your audience.
This ensures that the puritans are happy because the chain is clean. The QSound users are happy because their settings don't need tweaked to accomidate the pre-process massage. And the "what's a soundcard?" people are happy, because it's still consistant with the noise they're used to hearing from their PC.
How do you improve on personal choice?
Kid-proof tablet..