Help - Search - Members - Calendar
Full Version: DivX 5 vs XviD 'stable' vs XviD 'latest' results
Hydrogenaudio Forums > Misc. > Off-Topic
Pages: 1, 2
Gabriel
5-pass?

Why would a codec need 5 pass?
smok3
QUOTE(Gabriel @ Feb 4 2003 - 09:53 AM)
5-pass?

Why would a codec need 5 pass?

i wonder that 2..., seems like it has something to do with a 'better' bitrate distribution (i guess lower the average bitrate is more passes should be usefull), it must be usefull, the best mpeg2 encoder 'cinema craft' is having the multipass option as well biggrin.gif (iam sure some1 will know the proper answer).
SometimesWarrior
QUOTE(Gabriel @ Feb 4 2003 - 09:53 AM)
5-pass?

Why would a codec need 5 pass?

The 3rd pass was noticeably different-looking from the 2nd pass in certain parts of the image. Was it better? Maybe. I'll compare the 5th and the 2nd to see if there's a noticeable improvement. Early reports say that increasing passes will lower the amount of edge noise or texture sparkling.

I have no idea why five passes would be needed. Ask the Divx guys. tongue.gif
tangent
QUOTE(SometimesWarrior @ Feb 4 2003 - 03:32 PM)
I just made a 5-pass Divx503 encode of "The Red Violin", with Garf's GT3b1 -q5 for the audio (turned out 144kbps) and it looks really good for a 1cd rip (GMC & B-frames, no qpel, "constant quality" biasing). I'm making an XviD comparison-encode, and I just want to make sure my settings are okay. I'm using the 02-02-2003 unstable build from Koepi's site, along with the Doom9 guide. My settings: luma masking, chroma motion estimation, GMC, qpel, B-frames (max 2 in a row), no curve compression, 10000-frame biased bit payback delay (this will probably cause Xvid to miss its target filesize?), no restriction on quantizers, H.263 motion estimation first-pass, New Modulated second-pass.

The movie is almost all low-motion, so I'm encoding it at 608x336 with sharp bicubic resize filter, although that gives it 0.148bits/pixel*frame. So far only the first-pass is being done, and it's going quite slow; about  0.4x realtime on my dual XP1600+, compared to 0.9x realtime for Divx503. It would be going faster, but I have to use Decomb's IVTC to fix the mastering engineer's screwy captioning job. But speed isn't the issue here, quality is. So are my settings not optimal for Xvid?

For Divx503, you would probably do better enabling qpel and disabling gmc.

I would use the 26/1/2003 unstable build instead. 2/2/2003 seems to have some oversizing problem. Bframe settings 2/150/100 or 2/100/200 should be ok. I would leave bit payback delay as default. Quantiser 2-6 for iframe, 2-16 for pframe. I haven't had much faith with modulated quantiser recently, so i tend to stick with the tried and test method of h263 for 1cd and mpeg for 2cds. Your choice of quantiser should be ok though.
SometimesWarrior
QUOTE(tangent @ Feb 4 2003 - 03:20 AM)
For Divx503, you would probably do better enabling qpel and disabling gmc.

I would use the 26/1/2003 unstable build instead. 2/2/2003 seems to have some oversizing problem. Bframe settings 2/150/100 or 2/100/200 should be ok. I would leave bit payback delay as default. Quantiser 2-6 for iframe, 2-16 for pframe. I haven't had much faith with modulated quantiser recently, so i tend to stick with the tried and test method of h263 for 1cd and mpeg for 2cds. Your choice of quantiser should be ok though.

I thought GMC was finally working in Divx503. I'll read though the threads at Doom9 again to see if I misunderstood. I didn't see anyone talk about qpel yet.

QUOTE(Neo Neko @ Feb 2 2003 - 12:06 AM)
K. I tend to leave me credits in. It just seems incomplete otherwise. But be sure to try setting the bitrate playback delay to 10000.

So should bit payback be the default 250, or maxed out? Also, Iago's guide recommended the quantizer limits 2-6 and 2-16 like you said, while Doom9's guide said leave them at 2-31 and 2-31. In hindsight, I would probably feel safer setting quantizer limits for unstable Xvid builds (akin to setting a minimum -b for Lame VBR MP3's). I don't feel like re-encoding The Red Violin for the Nth time (pardon the pun) but I will try your recommendations for the next movie I make. By the way, where can I find the Jan.26 build?

Quick summary of quality comparison: I was really surprised and impressed by how good Xvid looks. My last test with Xvid was half a year ago, when it didn't come close to Divx502 quality, and it's really progressed nicely since then. The few high-bitrate "nature" scenes in The Red Violin (ocean, fire, rain) turned out better in Divx, but Xvid didn't massacre them at all. There were scenes when I liked Xvid more than Divx, when Xvid's edge noise wasn't worse but its sharpness was better. Also, sometimes Xvid's textures "stuck" to moving objects better. On the whole, Divx's consistent quality still won me over. By the next stable Xvid build, however, I might change my tune.
Althalus
using divx5.03 I did a 3 pass of First Knight (1cd). The third pass looks noticably better than the 2nd pass

Especially they sky, scenes with fire, etc.

like someone said on doom9

2nd pass = 90%
3d pass = 97%
4th pass = 98%
etc smile.gif

that rough estimate fits with my findings so far. I'm sticking with 1pass + 2nth passes.
mithrandir
QUOTE(SometimesWarrior @ Feb 3 2003 - 10:41 PM)
But even compared to film, I don't think DVD has the necesary resolution to provide film-like transparency.

No, DVD is nowhere close to providing the resolution of film. DVD is 720x480 whereas quality 35mm filmstock is equivalent to about 3000x2000. I shouldn't say "film-like transparency" because that can be misleading.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2009 Invision Power Services, Inc.