View Full Version : Java Xvid Stats Viewer 0.41


bugsan
28th December 2003, 23:09
new version 0.41 build 19022004

- fixed minor bugs with realvideo stats file.

---------
hi :) first, sorry for my worse english :D

xvid 1.0 comes with a new .pass file.
So i decided to code a tool which prints statistics and graphs.

Written in Java, it works on Linux and Windows.
(you need Java Sun JVM http://www.java.com/)

http://membres.lycos.fr/xym0x/jxvid1.gif
http://membres.lycos.fr/xym0x/jxvid2.gif
http://membres.lycos.fr/xym0x/jxvid3.gif
http://membres.lycos.fr/xym0x/jxvid4.gif
http://membres.lycos.fr/xym0x/jxvid5.gif

you can download it here : JXvidStatsViewer.jar (http://membres.lycos.fr/xym0x/JXvidStatsViewer.jar)
if you have installed Sun JRE, just double click, or tip this :
C:\foobar\> javaw -jar JXvidStatsViewer.jar

or the windows binary
JXvidStatsViewerW.ex (http://membres.lycos.fr/xym0x/JXvidStatsViewerW.ex) (rename it .exe)

if you have an idea for a graphic, send an email bugsan666@hotmail.com

CruNcher
28th December 2003, 23:41
very nice work bugsan :)

a quantizer graph would be nice :)

Koepi
29th December 2003, 01:07
A quantizer graph would be quite useless, as the statsfile usually has ~2 different quant values ;)

But indeed, very nice work bugsun! Thank you for the nice tool!

Regards
Koepi

tiki4
29th December 2003, 12:46
Huh, just came over this. I will try it immediately I'm back home. The screen shots really look beautiful. Nice work.

tiki4

Seiby
29th December 2003, 13:00
Very nice!

I'll give it a try :)

CruNcher
29th December 2003, 15:49
@Koepi hehe i meant the same for the Debug output :)

Bogalvator
2nd January 2004, 17:28
Surprised this hasn't generated more interest, I think this is a pretty nifty tool.

Are you going to develop it? I was thinking maybe an "average P frame distance" display would be quite useful (give a rough idea of how many B-frames in a row you're getting for a given B-VOP sensitivity)

Anyway, cheers for a great tool Bugsan!

sh0dan
4th January 2004, 00:00
Looks really great!

I guess resizeable/zoomable graphs would already be on the todo-list.

Looking forward to whatever the "AviSynth" section will include! :)

It could be great if it could be used for selecting frame ranges and applying specific filters for the ranges. Actually it just made me think it could be very interesting to be able to load stat-data into an AviSynth filter... Hmmm... :)

bugsan
5th January 2004, 03:38
new version 0.3

- fixed minor bugs
- added average consecutive b/pframes (ibpbbp: b=1.5 p=1)
- resizable window (=resizable graphs)
- enabled avisynth tab, just see what it does. I am too much tired to explain this in english.

bugsan
5th January 2004, 15:36
update v0.35

- new motion graph
- copy to clipboard button

GolovachLena
5th January 2004, 15:54
Originally posted by sh0dan
I guess resizeable/zoomable graphs would already be on the todo-list.

They're already resizeable and zoomable, i guess. Try dragging the border of the main window. Well, may be it's not that zoom you possibly mean, but looks nice anyway. Cool tool indeed and platform independent that is very, very nice for me personally, because i'm constantly trying XviDing under Linux.

Human_USB
5th January 2004, 16:18
How do I get the code to work??


movie = movie+clip.Trim(0,19834).Deen("w3d",1,4,2,4,3)
movie = movie+clip.Trim(19835,20640).Deen("w3d",1,10,5,4,3)
movie = movie+clip.Trim(20641,20777).Deen("w3d",1,4,2,4,3)
movie = movie+clip.Trim(20778,21903).Deen("w3d",1,10,5,4,3)
movie = movie+clip.Trim(21904,22046).Deen("w3d",1,4,2,4,3)
movie = movie+clip.Trim(22047,23938).Deen("w3d",1,10,5,4,3)
movie = movie+clip.Trim(23939,36000).Deen("w3d",1,4,2,4,3)
movie = movie+clip.Trim(36001,36566).Deen("w3d",1,10,5,4,3)
movie = movie+clip.Trim(36567,39297).Deen("w3d",1,4,2,4,3)
movie = movie+clip.Trim(39298,41221).Deen("w3d",1,10,5,4,3)
movie = movie+clip.Trim(41222,41974).Deen("w3d",1,4,2,4,3)
movie = movie+clip.Trim(41975,43245).Deen("w3d",1,10,5,4,3)
movie = movie+clip.Trim(43246,44321).Deen("w3d",1,4,2,4,3)


Thanks
-Jason

sh0dan
5th January 2004, 16:44
Add source at top like this:
clip = MPEG2Source("file.d2v")


Remove the first "movie" reference in the first line, like this:
movie = clip.Trim(0,19834).Deen("w3d",1,4,2,4,3)


Add this at the end:
return movie

Human_USB
5th January 2004, 16:57
Thanks for your help..... Would this work?


SetWorkingDir("C:\Filters\") #Filter Folder

##Mpeg2 Loader
LoadPlugin("mpeg2dec3.dll") #mpeg2source

LoadPlugin("Deen.dll") #Deen
LoadPlugin("UnDot.dll") #Undot ^^Always Use^^
LoadPlugin("kerneldeint140.dll")

clip = mpeg2source("C:\DVD\Rips\FLCL\flcl-non.d2v")

trim(8000,8720) #30 seconds

Import("C:\Filters\kerneldeint.avs")

##Crop
crop(4,2,712,476)

LanczosResize(640,480)

movie = clip.Trim(0,19834).Deen("w3d",1,4,2,4,3)
movie = movie+clip.Trim(19835,20640).Deen("w3d",1,10,5,4,3)
movie = movie+clip.Trim(20641,20777).Deen("w3d",1,4,2,4,3)
movie = movie+clip.Trim(20778,21903).Deen("w3d",1,10,5,4,3)
movie = movie+clip.Trim(21904,22046).Deen("w3d",1,4,2,4,3)
movie = movie+clip.Trim(22047,23938).Deen("w3d",1,10,5,4,3)
movie = movie+clip.Trim(23939,36000).Deen("w3d",1,4,2,4,3)
movie = movie+clip.Trim(36001,36566).Deen("w3d",1,10,5,4,3)
movie = movie+clip.Trim(36567,39297).Deen("w3d",1,4,2,4,3)
movie = movie+clip.Trim(39298,41221).Deen("w3d",1,10,5,4,3)
movie = movie+clip.Trim(41222,41974).Deen("w3d",1,4,2,4,3)
movie = movie+clip.Trim(41975,43245).Deen("w3d",1,10,5,4,3)
movie = movie+clip.Trim(43246,44321).Deen("w3d",1,4,2,4,3)

return movie


@bugsan

Thank you for a great program. Under the AviSynth tab can you add a low area too.

Thanks
-Jason

bugsan
7th January 2004, 13:08
new version 0.40

+ graph for PSNR files (produced by compare() and compareyv12())
+ an option to adjust details
+ win32 binary available (just a jar file in an exe file)

Human_USB
7th January 2004, 15:48
Thank you again for a great program. Maybe one day you will make it so we can edit the stat file in your program and then see a before and after graphs.

Bogalvator
12th January 2004, 01:24
Forgot to say nice one on the update, especially the average P & B frames distance displays, and also on the executable - now no-one has an excuse not to use your program!

One tiny thing for next version - make it so it loads up *.stats files by default as well as *.pass files (or is there a smug difference between the 2 that I don't know about?)

Looking forward to the next version :)

bugsan
12th January 2004, 20:10
*.pass is the new stats file name (current CVS and koepi beta3)

The exe file is a simple way to execute it, but the JVM is needed.

Dreassica
12th January 2004, 21:52
w3d in deen?? I either am outdated or u made a typo there, since i was under teh assumption such a function wasnt in there.:confused:

Manao
12th January 2004, 22:10
It's a mode from the latest version of the filter made by MarcFD. It was made available on a french forum : look here (http://atlas2.tgv.net/~media-video/forum2/viewtopic.php?t=2808&highlight=deen+w3d).

Edit : the translation of the readme is here (http://forum.doom9.org/showthread.php?s=&postid=407922&highlight=w3d#post407922)

Dreassica
12th January 2004, 23:05
Seems liek i wasnt up to date on my filtyers then ^_^"

Thnx for tyhe links, ill try it out.

Prettz
14th January 2004, 19:33
very impressive for a java program.
heheh just kidding, looks amazingly professional

sxd
14th January 2004, 23:56
The AviSynth feature, even though I have no clue how it works, is quite nice. I did a test 1 CD encode of a chapter of a DVD with both low and high motion scenes. The bits per pixel went down from 0.646 to 0.598 with Undot() as the normal filter, and Unfilter(-5, -5) as the high filter. I didn't test it with the defaults (I don't have or normally use Deen), but I suspect I would get a very nice increase in compressablility.

bugsan
19th February 2004, 15:04
version 0.41

- fixed minor bugs with commented lines in stats file.
This tool should work with realvideo stats file.

crusty
19th February 2004, 16:48
It still says version 0.40 though..just downloaded it.

But it looks like a nice tool. Could come in handy.

Aktan
20th February 2004, 09:38
I forgot to mention this Bug I found a long while back. It seems that the Viewer considers all S-Frames as B-Frames which messes up the stats.