Use the open source community to test & improve your stuff for you, and then close things down and make money with it.
Hey, it seems to be a succesfull formula! Any more examples of stuff like this happening?
I know of one: http://www.freechess.org
Take GPL code, improve it a little, and then sell it off (to the USCL in this case). Doesn't matter it's a GPL violation...the author can't prove anything anyway, unless he pays $$$ lawyers.
I'm not going to ruin the quality of my songs just because the ones that make the players are too short-sighted to include Vorbis support, forcing me to transcode everything (which introduces horrible artifacts)
If the player wont support Vorbis, I'm not going to buy it.
It's not because _you_ don't care that _nobody_ _else_ cares.
Monkey's Audio is a fine program. It's by far the best lossless compressor today. The only problem is that it is Windows only and nonfree.
Now the really good news:
Monkey's Audio's author and the Ogg guys have already been talking and he is willing to opensource his format and get it working in the ogg framework, as well as to cooperate to make it even better.
Unfortunately the poster didn't mention this, so I will.
This is a tuning release. Although all infrastructure like channel coupling is in place, the encoder itself is not ideally tuned yet. One of the goals of this release is to get people to test the new modes and report possible problems (samples were it goofs up). If you do this, be sure to try a blind test. Your mind _will_ play tricks on you otherwise.
Two known problems currently are pre/postecho on some really hard samples, and occasional 'hissing' in the low bitrate modes (< 160).
Both are known and will be fixed in the very near future. RC3 is already expected next week.
>Although it is possible that the casette input amp is less accurate near the low end than the CD input amp, I doubt it.
You're basically sending the singal through two totally different paths. Additionally, the CD input is likely to be of higher quality than the casette input (why make that good?).
Vorbis is very interesting for hardware manufacturers because of the lack of licensing fees and patents. Although support it not that good at the moment, there have been companies that expressily said Vorbis support would be preferred over mp3pro. Some are waiting for 1.0, some already have unofficial support.
WinAmp will support Vorbis by default in the next release. Why _not_ make the move? Unless you have a non-upgradable portable, there's only good points to it.
>The slashdot editors are the ones who decide which stories get posted (decisions I disagree
>with as often as not BTW), not those readers who happen to have moderator priveleges at a given moment.
Slashdot editors _do_ moderate. Either that, or by some magic all moderators had the exact same ideas at the somethingawful.com messup.
I agree that the parent could have been modded down by readers. The end result stays, namely that any post on slashdot which gives the remote hint that slashdot could have done something wrong is usually quickly at -1, no matter how factual it is.
Slashdot posted a link to a site which already had trouble paying its bandwidth. Of course it was slashdotted into oblivion. The owner responded by redirecting all slashdotters to goatse.cx.
Every post which said something about this was modded down instantly, and the front page claimed
'link removed because people were being redirected randomly'.
>I guess its okay for the/. editors to mod a
>post pointing out their flaws down into the nine
>realm of hell - but its bad that this LT guy
>astroturfed?
The really sad thing is of course that this is completely true, as has already been demonstrated. (you're at 0, Troll at the time of this posting).
The Somethingawful debacle is another nice example. Everything that remotely indicated slashdot could have done something wrong was -1 within seconds.
>Does anyone think a human could beat a computer
>in a non-timed game?
The comp would win. It just waits a human lifetime
between moves:)
There has been a match between a top human
correspondance player and the top programs, and
he did remakably bad. Then again, he had never
played a computer before.
Slashdot Mirror
They knew what codecs they were listening to, did not have to ABX and as far as we could verify different bitrates where mixed.
It's harder to get a test _worse_ than the Washington Post did.
--
GCP
OpenDivX vs. DivX4
Use the open source community to test & improve your stuff for you, and then close things down and make money with it.
Hey, it seems to be a succesfull formula! Any more examples of stuff like this happening?
I know of one: http://www.freechess.org
Take GPL code, improve it a little, and then sell it off (to the USCL in this case). Doesn't matter it's a GPL violation...the author can't prove anything anyway, unless he pays $$$ lawyers.
--
GCP
Fixed point ogg decoders have already been written. This is how the HipZip supports Vorbis.
--
GCP
>Look, nobody really cares about Ogg Vorbis.
Heh well I do.
I'm not going to ruin the quality of my songs just because the ones that make the players are too short-sighted to include Vorbis support, forcing me to transcode everything (which introduces horrible artifacts)
If the player wont support Vorbis, I'm not going to buy it.
It's not because _you_ don't care that _nobody_ _else_ cares.
--
GCP
When this was first introduced the joke was that Ogg would be safe as long as we don't meet any alien races with more than 255 ears :)
--
GCP
>I'll wait patiently.
Hmm perhaps you misunderstood.
Channel coupling is in. You just cant specify for yourself what mode to use. The encoder does it for you.
--
GCP
Reality check: bzip2 gets nowhere near the compression of a special lossless audio compressor.
--
GCP
With RC2 just being out, this was rather hard to do. Also, the encoder is not fully tuned, so waiting for 1.0 final would be advisable.
Another factor is the bitrate. The ordering of codecs @ 64 kbps can be totally different than that @ 160kbps.
--
GCP
>There are no patent royalties for software MP3
>decoders.
This is wrong. Go take a look at http://www.mp3licensing.com/royalty/index.html
It's not because you don't pay that there are no licenses.
BTW. That page should be _the_ reason why Vorbis will succeed.
--
GCP
This is nonsense.
Monkey's Audio is a fine program. It's by far the best lossless compressor today. The only problem is that it is Windows only and nonfree.
Now the really good news:
Monkey's Audio's author and the Ogg guys have already been talking and he is willing to opensource his format and get it working in the ogg framework, as well as to cooperate to make it even better.
--
GCP
Unfortunately the poster didn't mention this, so I will.
This is a tuning release. Although all infrastructure like channel coupling is in place, the encoder itself is not ideally tuned yet. One of the goals of this release is to get people to test the new modes and report possible problems (samples were it goofs up). If you do this, be sure to try a blind test. Your mind _will_ play tricks on you otherwise.
Two known problems currently are pre/postecho on some really hard samples, and occasional 'hissing' in the low bitrate modes (< 160).
Both are known and will be fixed in the very near future. RC3 is already expected next week.
--
GCP
>Although it is possible that the casette input amp is less accurate near the low end than the CD input amp, I doubt it.
You're basically sending the singal through two totally different paths. Additionally, the CD input is likely to be of higher quality than the casette input (why make that good?).
I think your test was flawed.
--
GCP
I don't know what you've done, but Oggenc should run at 4x _at least_ on your system.
The 128kbps mode is not ideally tuned (IMHO), but the problems are known. 160kbps is already a lot better.
--
GCP
MP3 is just MP3, you can't put Vorbis in that
;)
Ogg is sortof like AVI/ASF, with Vorbis being the MP3 and Tarkin being the DivX
--
GCP
Better quality at a smaller filesize.
Vorbis is very interesting for hardware manufacturers because of the lack of licensing fees and patents. Although support it not that good at the moment, there have been companies that expressily said Vorbis support would be preferred over mp3pro. Some are waiting for 1.0, some already have unofficial support.
WinAmp will support Vorbis by default in the next release. Why _not_ make the move? Unless you have a non-upgradable portable, there's only good points to it.
--
GCP
>Why would you want lower bitrates?
Streaming
--
GCP
Vorbis supports up to 255 fully coupled channels
So basically this is already done.
--
GCP
Hmm, isn't replying to spam bad because it will
mark your address as 'in use' with spammers?
This looks like a fun thingy, but I wouldn't
use it to acutally reply to spammers for sure...
--
GCP
Been there, done that:
e ct %20FAQ.htm
ftp://cap.connx.com/pub/Chess%20Analysis%20Proj
Building something that actually plays chess
over such a network is not going to work. Latency
is critical for all current parallel chess algorithms.
--
GCP
>The slashdot editors are the ones who decide which stories get posted (decisions I disagree
>with as often as not BTW), not those readers who happen to have moderator priveleges at a given moment.
Slashdot editors _do_ moderate. Either that, or by some magic all moderators had the exact same ideas at the somethingawful.com messup.
I agree that the parent could have been modded down by readers. The end result stays, namely that any post on slashdot which gives the remote hint that slashdot could have done something wrong is usually quickly at -1, no matter how factual it is.
--
GCP
Slashdot posted a link to a site which already had trouble paying its bandwidth. Of course it was slashdotted into oblivion. The owner responded by redirecting all slashdotters to goatse.cx.
Every post which said something about this was modded down instantly, and the front page claimed
'link removed because people were being redirected randomly'.
Randomly eh?
--
GCP
>I guess its okay for the /. editors to mod a
>post pointing out their flaws down into the nine
>realm of hell - but its bad that this LT guy
>astroturfed?
The really sad thing is of course that this is completely true, as has already been demonstrated. (you're at 0, Troll at the time of this posting).
The Somethingawful debacle is another nice example. Everything that remotely indicated slashdot could have done something wrong was -1 within seconds.
'We are slashdot. We are hypocrites'
--
GCP (come on, mod me down, see if I care)
>If it's cleanroom, tell 'em to fuck off
>directly.
How are you going to do this cleanroom?
You'd need to know at least how the format
works. Even then a cleanroom implementation
can have trouble with patents.
If the only way to decode it is via patented
technology, you're screwed. Just don't use
the format. Just don't buy DVD's.
MP3 is in _exactly_ the same situation today.
--
GCP
>Anyway, we are talking Blitz Chess here
:)
No, the games are standard (long) timecontrols.
>Does anyone think a human could beat a computer
>in a non-timed game?
The comp would win. It just waits a human lifetime
between moves
There has been a match between a top human
correspondance player and the top programs, and
he did remakably bad. Then again, he had never
played a computer before.
--
GCP
>Not really, the "Power chess 98" engine drew
>DeepBlue.
>Ever heard of World Chess Computer Championships?
As has already been stated in another post, they
drew a PROTOTYPE.
--
GCP