vp5 released today...

[buba king]

I got a e-mail today. On2 could offer me 50% better video quality over all current video codecs..

So i installed the real one player, the suckiest plyer on earth. and the vp5 decoder.

and i was not impressed by the samples they provided.. just another hype if you ask me..

email:

Dear On2 Friend:

Today, On2 Technologies is releasing the most powerful independent video codec in the world. More efficient than H.26L and delivering higher quality at lower bit rates than MPEG-2, MPEG-4, Real 8, Windows Media 8, Apple Quicktime 5 or DIVX, this codec will significantly alter the landscape of video streaming and downloading. It will also increase the number of businesses and homes that can get full-screen, full-frame rate, TV-quality video.

As you may know, On2 already counts a number of large streaming and set-top box integrators among its customers. Some of the attractive features of VP5 are listed below:

- Competitive codecs typically require data rates 50% higher than VP5 to achieve comparable quality
- Ability to handle interlaced content
- Decodes from 120-210 times faster than standards-based codecs (like H.26L)

To see VP5 compared to other codecs, go to www.on2.com and visit the VP5 preview section, or for our detailed press release, http://www.on2.com/ pressreleases.php3?qs1=vp5_release_feb_21_02

If you have questions about VP5, please call us at (917) 237-0500 or join us on our conference call today at 5 PM Eastern Standard Time, when we will discuss our new codec and the company's 2001 financial results (to be released at 4pm EST). The number is 1-888-881-4892.

Thanks for your continued interest in On2 Technologies.

Best,

Douglas A. McIntyre
President and Chief Executive Officer
On2 Technologies, Inc., The Duck Corporation


Any comments?

[Acaila]

Sounds like a normal advertising line just to sell their product. But I'm gonna check it out.

The only sucky part is I have to install RealPlayer :(

[Acaila]

Just a little quote from their page:

Known Issues:

Lack of overlay support with certain video card Because of the RealPlayer video
architecture, VP5 may look blocky or fuzzy when stretched or blown-up to
full-screen. This happens when RealPlayer cannot get overlay support from the video
card; it may also randomly lose overlay support when stretched to certain video
sizes, or if another application (such as MediaPlayer) is using the overlay.

Setting RealPlayer maximum bandwidth too low can cause excessive buffering
If the network settings in RealPlayer are set below the actual bandwidth of the file,
then RealPlayer may buffer for extended amounts of time before playing, or not play
the file at all. The best choice is to set to the highest speed possible.

Black or green screen after pausing or seeking after pausing or seeking within
RealPlayer and then playing again, there may be a black or green screen for a few
seconds. This is an issue with the way RealPlayer deals with keyframes, it can be
fixed by setting keyframes closer together.

Now aren't these fun? :(
Maybe if we could put it into an OGG container the seeking issue will be gone :D

[-h]

Well let's see..

50% better than H.26L. Seeing as H.26L isn't anywhere near finalised yet (and even so, is far outperforming MPEG-4 in terms of compression), VP5 must be at least 5 years ahead of anyone else in the video compression field. That alone is enough to make me skeptical (given their past..... average performance).

And decoding 120 times faster... well what can I say. Even with the added compexity of H.26L, it could still decode at full speed on a K7 1 GHz. VP5 can decode video at 2500 fps? I don't think so.

And even then.. blaming the "overlay" for blockiness during playback is verging on farcical. This mob deserves to be shunned by anyone with experience in video comperession. It's just some rabid marketing employee, foaming at the mouth with the opportunity to spout buzzwords, with little to back it up.

If they open the source, perhaps it will be of some worth. To someone. Somewhere.

-h

Edit: Maybe I was a bit harsh first time around.

[Acaila]

You were...but just a little bit.

I agree the features they talk about could be nowhere near reality, but I'm not going to judge them on past experiences.

This mob deserves to be shunned by anyone with experience in video comperession.

I need first-hand proof before I can decide if the encoder is worth it. Anyone with experience in video compression should not make up their mind before having seen the facts. :)

I'll start running some tests when I get back home, and then I have the weekend to toy around with it...

[-h]

I need first-hand proof before I can decide if the encoder is worth it. Anyone with experience in video compression should not make up their mind before having seen the facts. :)

Well.. let's just say I have much much much much much much much much much much more faith in the H.26L team (who have proven themselves to be staggeringly capable) than the VP5 team... whomever they may be :)

This is not disimilar to the endless "I can compress any file down to 10%!!!" claims repeatedly posted on comp.compression and other pure data compression boards - if it were possible to compress data as well as they claim, I can guarantee that the frighteningly smart people we all love would've already found and exploited it.

-h

[bullitB]

Has anyone here actually looked at these samples?

I don't know about you guys, but the VP5 samples at 500 kbit/sec looked really good to me. Much, much better than anything I've seen at similar resolutions from DivX, XviD, RealVideo...anything. Now, I know looking at pre-compressed material isn't the best method for comparing video codecs, but even so, On2 may be onto (no pun intended) something here.

As for H.26L...isn't it like a year away from completion, let alone implementation? If On2 can get this codec out when they say they can (this summer), we may have a legitimate alternative to a certain codec with $0.02/hour patent licensing fees...

[-h]

I don't know about you guys, but the VP5 samples at 500 kbit/sec looked really good to me. Much, much better than anything I've seen at similar resolutions from DivX, XviD, RealVideo...anything. Now, I know looking at pre-compressed material isn't the best method for comparing video codecs, but even so, On2 may be onto (no pun intended) something here.

Without an independent test, such samples don't mean much. The samples that Microsoft chose for WMV8 / WMA8 made it look pretty good as well...

As for H.26L...isn't it like a year away from completion, let alone implementation? If On2 can get this codec out when they say they can (this summer), we may have a legitimate alternative to a certain codec with $0.02/hour patent licensing fees...

And what are On2's licensing terms going to be? They have patented this stuff, remember, and I haven't seen VP4 getting much use ;)

The performance of the current H.26L code is far far better than any MPEG4 codec in use today, by at least 50%. For On2 to claim it can best H.26L by a further 50%, while decompressing 120 times faster (which means they can't be using entropy coding, motion compensation, fourier/dc/wavelet transforms, or really anything in use by any other codec made in the last 7 years), just strikes me as being a bit too extraordinary.

When someone I trust posts screenshots, from a source I can verify, demonstrating VP5's superiority, I might change my mind :)

-h

[Doom9]

I've already asked their PR department if I could get a review copy.. let's keep our fingers crossed

[24hourloop]

Originally posted by bullitB
Has anyone here actually looked at these samples?

I don't know about you guys, but the VP5 samples at 500 kbit/sec looked really good to me. Much, much better than anything I've seen at similar resolutions from DivX, XviD, RealVideo...anything. Now, I know looking at pre-compressed material isn't the best method for comparing video codecs, but even so, On2 may be onto (no pun intended) something here.

As for H.26L...isn't it like a year away from completion, let alone implementation? If On2 can get this codec out when they say they can (this summer), we may have a legitimate alternative to a certain codec with $0.02/hour patent licensing fees...

I must agree. I thought their stuff looked better than DivX at the same data rate. Of course their claims are exaggerated. People expect that. Seen a car commercial lately?

So I believe we ought to wait for the encoder and give it a whirl ourselves.

[Acaila]

I checked the vids out and all I can say is VP5 blows DivX4 away at equal bitrates.

Too many other things speak for DivX for atm so VP5 is far from being competition. Quality alone doesn't cut it. There's compatibility/playability which is also very important. So far it can only be viewed with RealPlayer. The low bitrate (200kbps) video played without a problem on my system, but the high bitrate (500kbps) one ran into serious resource problems. It would not run smooth no matter what I tried.
But from what I could see VP5 at 500kbps equals DivX4 at about 900kbps.

[Vanos_b]

From what I could see in the trailers, VP5 is only better in smoothing the image and I think that with the same amount of detail loss XviD/DivX could produce a better image (same bitrate, of course). Speed is a problem too as on my Duron 900 the 500kbps movies don't run smoothly if a have 2-3 other (IE5) windows opened (XviD works perfectly). In fact the quality is quiet similar to VP3, only that (I think) VP5 uses VBR. We'll live and see.

[diji1]

So far it can only be viewed with RealPlayer.

Sheesh, it was all sounding like it might be worth checking out up until i read that :) ... lol, real player.

[movmasty]

hey -h, what is H.26L ?????????