Quantcast Cqtester - digitalFAQ.com Forums [Archives]
  #1  
11-19-2003, 04:27 PM
rendalunit rendalunit is offline
Free Member
 
Join Date: Apr 2002
Location: san jose, Ca
Posts: 1,148
Thanks: 0
Thanked 0 Times in 0 Posts
http://www.boraxsoft.de/CQTester.shtml


I just discovered this cqmatic-like tool by borax

I'm trying it out now. It looks cool.


I pasted the link above here for english translation
Reply With Quote
Someday, 12:01 PM
admin's Avatar
Site Staff / Ad Manager
 
Join Date: Dec 2002
Posts: 42
Thanks: ∞
Thanked 42 Times in 42 Posts
  #2  
11-19-2003, 04:58 PM
jorel jorel is offline
Invalid Email / Banned / Spammer
 
Join Date: Aug 2002
Location: Brasil - MG - third stone from the sun
Posts: 5,570
Thanks: 0
Thanked 0 Times in 0 Posts
thanks friendal,
seems cool...i'm trying to follow the pictures, can't understand any word!


oh...now i see the link to "translate"!

double thanks!
Reply With Quote
  #3  
11-20-2003, 04:08 AM
kwag kwag is offline
Free Member
 
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
Another "CQMatic, Tok, acp" (me too) tool, with an "MVCD" accent

-kwag
Reply With Quote
  #4  
11-20-2003, 04:54 AM
incredible incredible is offline
Free Member
 
Join Date: May 2003
Location: Germany
Posts: 3,189
Thanks: 0
Thanked 0 Times in 0 Posts
Send a message via ICQ to incredible
Quote:
Originally Posted by kwag
Another "CQMatic, Tok, acp" (me too) tool, with an "MVCD" accent

-kwag
Borax and his CQtester are independend from MVCD or KVCD and some time ago he showed there this CQ testing appl.
He programmed this also a time ago, and the first Question I made there was - when I saw the offer in his Thread - , if this appl. got its own core and if nothing of the core is taken frome somewhere else. I do beleive him as he assured it.

It works refered to TmpgEnc's "Source Range" Settings different in comparison to CQmatic.
YOU HAVE TO SET a range! So the whole movie range has to by choosing "Cut everything except selected area" (as I remember the text in TmpgEnc).
Also afterwards the "source range" checkbox still has to be checked!

And be shure to set the right settings refered to prediction sampling.

Im also in contact with Borax for optimization and he's is still developing. So the Whole Sampler can be movie percentual based and so on.
A nice bonbon is the "curve" where you can watch the CQ approach

The Sample length in case of KVCD has to be set to:

- 0,96 sec! PAL
- 1 sec! NTSC 23.976 encoding
- 0,8 sec! NTSC 29.976 encoding

The formula is the following: ((1/fps)*GopSeq-length) = sec. of Sample Length.
So for example (1/25)*24 = 0,96 sec. in case of PAL 25fps (as KVCD standard GopSeq Length is 24)
It does not work with stream percentual based offset. So you have to set the Sample frequency.

One week ago Borax and I we talked about the GOP-Lenght based Sampler Length setting, and if this appl. could read out of the Projects GOP Settings so the right Sampler Length will be calculated.

As he's still developing and if you in here find something which could enhance the precision, just send him your thoughts via email.

And as "babelfish.altavista" doesn't outputs a full translated-link to paste somewhere else, I give you a Link translated by google:

EDIT: Borax informed me via email that he translated the HowTo in english:
http://www.boraxsoft.de/CQTester_eng.shtml
Reply With Quote
  #5  
11-20-2003, 11:13 AM
kwag kwag is offline
Free Member
 
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
Well, I wish Borax a lot of luck with "Adventures in CQ prediction"

-kwag
Reply With Quote
  #6  
11-20-2003, 02:00 PM
rendalunit rendalunit is offline
Free Member
 
Join Date: Apr 2002
Location: san jose, Ca
Posts: 1,148
Thanks: 0
Thanked 0 Times in 0 Posts
I like it that cqtester goes all the way to 100 Sorry kwag I know I sound like a broken record I can't stand 2 cd encodes that are cq=90 and audio 256kbps because I like to think that the video is as good as it can be. The numbers below show that cq 100 is very slightly better than cq 90.

I did some comparisons:

Code:
                      bitrate               Q. level
groundhog day
1:41 sample           -------               --------
                      peak      avg         peak        avg
cq 90   24.281 mb     2265      1824        4.00        1.61 
cq 100  24.574 mb     2267      1847        4.00        1.57

groundhog day (motion adaptive script)

cq 90   23.853 mb     2272      1792        3.55        1.53
cq 100  24.201 mb     2280      1818        3.55        1.48

Divx source (v4 optimal script)
1 minute sample

cq 90   18.542 mb     2531      2108        5.04        2.41
cq 100  18.560 mb     2531      2110        5.04        2.40
I promise I won't bring this up again
Reply With Quote
  #7  
11-20-2003, 02:16 PM
kwag kwag is offline
Free Member
 
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by rendalunit
The numbers below show that cq 100 is very slightly better than cq 90.
How about CQ=95
Could you make a quick test, and compare a clip at CQ=95 and another at CQ=100
Perhaps, if there is a slight quality gained, then I'll use 95 as MAX in CQMatic
But if there's no visible quality from 95 to 100, then I'll set 95 as ceiling
Quote:
I promise I won't bring this up again
No, please, bring it all up

-kwag
Reply With Quote
  #8  
11-20-2003, 06:25 PM
rendalunit rendalunit is offline
Free Member
 
Join Date: Apr 2002
Location: san jose, Ca
Posts: 1,148
Thanks: 0
Thanked 0 Times in 0 Posts
hey kwag,

I'm doing two complete encodes at cq 95 and cq 100. Is it possible to export the data from bitrate viewer to another app like Excell? I'd like to overlap the two graphs so I can find a spot where there's a difference in bitrate and view the clips at that exact spot side by side to see if there's any visual difference (or post the two clips for someone else to judge them). I'm using a trial version of BV btw and it doesn't have all the features as the registered versions (i think).

thnx,
ren
Reply With Quote
  #9  
11-20-2003, 08:17 PM
kwag kwag is offline
Free Member
 
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
Hi ren,

I think the best way is to open your encoded mpeg files in Vdub, and select a high action frame. Then copy that to clipboard, load it in Windows paint, and save as .PNG. Do the same with both encoded files, selecting the same frame. Then we can blow it up, and visually compare

-kwag
Reply With Quote
  #10  
11-21-2003, 02:03 AM
rendalunit rendalunit is offline
Free Member
 
Join Date: Apr 2002
Location: san jose, Ca
Posts: 1,148
Thanks: 0
Thanked 0 Times in 0 Posts
I encoded "Groundhog Day" at 352x240 with cq=95 and cq=100.
Code:
Mpeg2Source("C:\DVD_VIDEO\VIDEO_TS\groundhog.d2v")
LanczosResize(352, 176, 0, 0, 720, 480)
AddBorders(0, 32, 0, 32)
cq-95 final size=1.048 gb sample 9.28 mb ~47 seconds
cq-100 final size=1.058 gb sample 9.37 mb ~47 seconds

See if there's any visible difference between the two. I thought I saw a little more noise around the legs and feet of the people standing near the cliff in the cq95 sample.
Reply With Quote
  #11  
11-21-2003, 03:14 AM
kwag kwag is offline
Free Member
 
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
Hi ren,

I can't see any difference at all in Vdub
They both look identical, even after zooming frames.

-kwag
Reply With Quote
  #12  
11-21-2003, 04:33 AM
incredible incredible is offline
Free Member
 
Join Date: May 2003
Location: Germany
Posts: 3,189
Thanks: 0
Thanked 0 Times in 0 Posts
Send a message via ICQ to incredible
Also a nice way to do a 1:1 comparison on pics is as you said ... copy both windows of each sample to clipboard ... pasting them into Photoshop .... and afterwards copy both into a pixel-matching layer! based .psd file.
By doing this you just can switch on/off the first layer to see a very clear comparison ... thats the way I do it when testing AVS routines/filters.
Reply With Quote
  #13  
11-21-2003, 01:48 PM
rendalunit rendalunit is offline
Free Member
 
Join Date: Apr 2002
Location: san jose, Ca
Posts: 1,148
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by incredible
and afterwards copy both into a pixel-matching layer! based .psd file.
good idea Here's the .psd file which is two layers; layer1- 704x480 cq95 and layer2 cq100. They look exactly the same!

Now I'm convinced that there's no visible difference from cq95 and cq100 In fact there probably isn't any difference from cq 90 and cq 100.

ren
Reply With Quote
  #14  
11-21-2003, 05:16 PM
kwag kwag is offline
Free Member
 
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by rendalunit

Now I'm convinced that there's no visible difference from cq95 and cq100 In fact there probably isn't any difference from cq 90 and cq 100.

ren
Hi Ren,

Please do one more test, if possible
Compare the CQ=90 and CQ=100.
This is important, because the whole idea of using a MAX CQ is to speed up prediction. Maybe there is a difference between 90 and 100, but maybe between 92 and 100 there's not, so then I can change the ceiling to the maximum level.

-kwag
Reply With Quote
  #15  
11-21-2003, 08:32 PM
rendalunit rendalunit is offline
Free Member
 
Join Date: Apr 2002
Location: san jose, Ca
Posts: 1,148
Thanks: 0
Thanked 0 Times in 0 Posts
Sure, I'll test a sample with cq90 and 100 this evening.

edit: after I fix the broken window that I threw the dogs rubber ball through because I thought it was open!!!!
Reply With Quote
  #16  
11-21-2003, 08:46 PM
kwag kwag is offline
Free Member
 
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by rendalunit
edit: after I fix the broken window that I threw the dogs rubber ball through because I thought it was open!!!!
You do that too
I have a datshund, and all she wants to do all day, is for me to throw a tennis ball so she can chase it. Luckily, I don't have any glass windows around

-kwag
Reply With Quote
  #17  
11-22-2003, 10:04 AM
WaSaBi WaSaBi is offline
Free Member
 
Join Date: Oct 2003
Posts: 58
Thanks: 0
Thanked 0 Times in 0 Posts
seem to be good =)

where to find forum support for this tool ?
Reply With Quote
  #18  
11-22-2003, 10:07 AM
Encoder Master Encoder Master is offline
Free Member
 
Join Date: Mar 2003
Location: (^_^)
Posts: 504
Thanks: 0
Thanked 0 Times in 0 Posts
Reply With Quote
  #19  
11-22-2003, 06:47 PM
dazedconfused dazedconfused is offline
Free Member
 
Join Date: Nov 2002
Posts: 316
Thanks: 0
Thanked 0 Times in 0 Posts
Gosh Encoder Master, whatever could you possibly mean??? (http://www.kvcd.net/forum/viewtopic.php?t=5666) Just kidding!. Is it merely a coincidence that the link for that thread ends in "666"

Yes, let's all try to follow kvcd.net's unwritten rules of etiquette by not mentioning the "M" word around here....peace is already such a rare commodity these days.

-d&c

EDIT: Since EncoderMaster's preceding post has now been edited twice since he originally posted it (once since I replied), I'd simply like to prevent the possibility of my words being taken out of context by anyone by clarifying that: 1) EM originally made a small faux pas by posting a direct link here to a particular website, and his post was then edited some hours later so it said something similar to 'oops, I can't post a link to that site here...you can think why!', and 2) then it was re-edited again later by EM to remove those words, leaving only the 'homer' emoticon behind. EM, please understand that I mean no offense to you here by posting this!...I'm only including this footnote because your post has been changed twice now, and I do not want my own words or intent possibly being misunderstood or taken out of context by anyone because of this. I inferred that you were having a laugh about past happenings here, and that was my basic intention as well. Regards.
Reply With Quote
  #20  
11-23-2003, 03:37 AM
Encoder Master Encoder Master is offline
Free Member
 
Join Date: Mar 2003
Location: (^_^)
Posts: 504
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Yes, let's all try to follow kvcd.net's unwritten rules of etiquette by not mentioning the "M" word around here....peace is already such a rare commodity these days.
Your're right but I don't mean the M OK?
Reply With Quote
Reply




Similar Threads
Thread Thread Starter Forum Replies Last Post
Quelqu'un a tester CQtester? WaSaBi Conversion et d'Encodage de Vidéo (Français) 0 11-21-2003 07:17 PM

Thread Tools



 
All times are GMT -5. The time now is 04:04 AM  —  vBulletin © Jelsoft Enterprises Ltd