Intersting article on a new way of storing music files:
http://news.yahoo.com/s/cmp/20080402/tc_cmp/207001130
Better than MP3?
Moderators: Sharp, X-Trade, Pepperpotty, karmathanever
- georgeinar
- Platinum Member
- Posts: 3425
- Joined: Sat Jun 15, 2002 5:15 pm
- Location: Chicago
- Contact:
Better than MP3?
George Nelson is 2loose
http://www.cdbaby.com/artist/2loose for my cds
http://www.soundclick.com/2loose or follow my tweets = 2loose_buzzgoth (twitter name)
<a href="http://cdbaby.com/cd/2loose2"><img src="http://cdbaby.name/2/l/2loose2_tiny.jpg" width="50" height="50" alt="album cover" border="0"></a>
http://www.cdbaby.com/artist/2loose for my cds
http://www.soundclick.com/2loose or follow my tweets = 2loose_buzzgoth (twitter name)
<a href="http://cdbaby.com/cd/2loose2"><img src="http://cdbaby.name/2/l/2loose2_tiny.jpg" width="50" height="50" alt="album cover" border="0"></a>
- Timo
- Platinum Member
- Posts: 3106
- Joined: Thu Jan 24, 2002 8:53 am
- Location: Kaoss central, England
- Contact:
Sounds like MIDI, and realtime physical modelling?
Has anyone physically modelled individual pianos (Steinway, Boesendorfer, Bechstein, or even cheap and cheerful uprights, etc.)?
The other benchmark is the human voice. Currently I guess Vocaloid is the closest, although this still uses extensive sampling, and still sounds like an over-autotuned model.
I think we're rather a good distance from synthesising convincing vocals and piano!
You'd need a damn fast computer to model every nuance of vocals, the backing, the instruments, the whole mix, the effects/processing, mastering, everything in realtime. The expressive data stream would probably end up requiring more bandwidth than an MP3 download!
Otherwise the concept reminds me a bit of old skool "tracker" sequencers, from the Amiga, etc., which expressively played optimised samples and could easily fit on a 860k floppy disk along with game data, etc. The sampling/optimisation was an art in itself, though. But it was still just that, manipulating samples, not physical modelling.
Has anyone physically modelled individual pianos (Steinway, Boesendorfer, Bechstein, or even cheap and cheerful uprights, etc.)?
The other benchmark is the human voice. Currently I guess Vocaloid is the closest, although this still uses extensive sampling, and still sounds like an over-autotuned model.
I think we're rather a good distance from synthesising convincing vocals and piano!
You'd need a damn fast computer to model every nuance of vocals, the backing, the instruments, the whole mix, the effects/processing, mastering, everything in realtime. The expressive data stream would probably end up requiring more bandwidth than an MP3 download!
Otherwise the concept reminds me a bit of old skool "tracker" sequencers, from the Amiga, etc., which expressively played optimised samples and could easily fit on a 860k floppy disk along with game data, etc. The sampling/optimisation was an art in itself, though. But it was still just that, manipulating samples, not physical modelling.
<img src="http://www.infekted.org/timo/userbar-atmossphere.png" border="0" align="bottom" alt="Korg Moss Soundset"> [Free Moss Set For All Workstations With Moss Expansion]
<img src="http://www.infekted.org/timo/userbar-virus.png" align="bottom" border="0" alt="www.Infekted.org - Access Virus Community"> [Infekted.org - Original Access Virus Forum & Community]
Trinity V3 PBS | Radias KB | Virus TI Snow | Virus Indigo 1 | 505 Groovebox
<img src="http://www.infekted.org/timo/userbar-virus.png" align="bottom" border="0" alt="www.Infekted.org - Access Virus Community"> [Infekted.org - Original Access Virus Forum & Community]
Trinity V3 PBS | Radias KB | Virus TI Snow | Virus Indigo 1 | 505 Groovebox