KVCD Predictor or Manual Prediction - digitalFAQ.com Forums [Archives]
 digitalFAQ.com Forums [Archives] KVCD Predictor or Manual Prediction

#1
08-03-2003, 09:03 PM
 vhelp Free Member Join Date: Jan 2003 Posts: 1,009 Thanks: 0 Thanked 0 Times in 0 Posts
Hi all..

I know.. we're all busy doing lots of fun stuff

I've ben reading up on this "prediction" stuff, and my eyes are about to
pop out of my head if I don't take it easy, ...but it challenging and fun

I'm not sure which is the best method, but I'm going to go out on a limb,
and give a few of the suggests in "kVCD Predictor" thread a shot, manually
of course.

But, what really confuses me, is all the formulas out their. My eyes just
keep rolling around like crazy!! But, I'm going to try a few things
out tonight (now)
.
.
I just wanted to let you all know, and perhaps, keep a gernal of my efferts
and trials and tribulations and whatnots here. Maybe even code a little
something along the way.. either as an extension to vcalc or just something
tiny and compact. Anyways..

One last thing.. I have an Excel spreadsheet that I started some time ago,
with prediction, but had issues w/ it and as such, you know how it goes..
It's a bit messy, but I'm going to start it fresh, w/ some clean calculaitons
and "build" on it from their. I may post my results of this spreadsheet for
those curious maybe.

I'm off to my first attempt at "prediction" hehe.. wish me luck !!

Attempt #1

-vhelp
Someday, 12:01 PM
 Site Staff / Ad Manager Join Date: Dec 2002 Posts: 42 Thanks: ∞ Thanked 42 Times in 42 Posts
#2
08-30-2003, 12:59 AM
 leo Free Member Join Date: Aug 2003 Location: Mexico Posts: 25 Thanks: 0 Thanked 0 Times in 0 Posts
Hi guys.
I think prediction is the weaknes kvcd (maybe the only?), I've been trying with ACP, TOK, CQMATIC with various results.
My method?
Trial and error, most of the times I have to encode two or three times; but I dont mind cuz the results are greeaaat
If you know of any other method or soft please let me know.
#3
01-27-2004, 09:43 PM
 knowbody Free Member Join Date: Jan 2004 Posts: 3 Thanks: 0 Thanked 0 Times in 0 Posts
This is my first post here and I too have been using trial and error. Not having read many posts on the subject, my own experience has led me to believe that filesize prediction is likely to be problematic. I will give an example;
Using the 2-hour, 352x240 template, I have encoded the daylights out of a certain one-hour and forty-minute movie. What I discovered while encoding this movie, and by only making adjustments to the Rate Control Mode (using CQ in TMPGEnc), there is a huge difference between a setting of 78 and 79. My goal had been to adjust the setting for the highest possible value, producing the largest filesize that can still be burned successfully to a single CD, and I encoded the movie at every value between 75 and 82. At 79 the filesize was 811 megabytes, while a setting of 78 dropped it down to 737. That's a very large difference for a single-digit change to make and a much larger jump existed between these two values than between any two of the other adjacent settings.
So, with that said it is my newby conclusion that an individual is likely to either overshoot a movie and be forced to tweak-down and try again, or encode a filesize that seems like it could be bigger but in fact cannot be made so without exceeding the capacity of a CD.

I'd also like to say that KVCD is truly awesome, and I'm so glad to be able to get away from 2-disc VCD movies, burning an entire movie on a single CD with remarkable quality, that I could jump for joy !
#4
01-27-2004, 11:29 PM
 kwag Free Member Join Date: Apr 2002 Location: Puerto Rico, USA Posts: 13,537 Thanks: 0 Thanked 0 Times in 0 Posts
Welcome knowbody,

Check the CQMatic forum area.

-kwag
#5
01-28-2004, 01:30 AM
 knowbody Free Member Join Date: Jan 2004 Posts: 3 Thanks: 0 Thanked 0 Times in 0 Posts
Hello kwag,
Thank you for the greeting!
I've been doing a little reading and made another post in the File Prediction forum. I see that you are the site admin here, and so I'd like you to know what beautiful and unexpectedly fine results I've been getting with KVCD. I had heard of it at least a year ago but at that time had no experience with TMPGEnc and so I wouldn't have known what to do with Template if it was handed to me on a silver platter, LOL.
I've come a long way in a short time and just for the heck of it thought I'd try my hand with KVCD. I'm very glad I did and actually plan to use DivX and KVCD exclusively from here out. Like leo and vhelp I too have been attempting to work out filesize predictions. This codec is certainly worth the work involved, that's for certain !!
Thanks again for the welcome and tip. I'm looking forward to using and learning all I can about KVCD.
#6
01-28-2004, 08:48 AM
 kwag Free Member Join Date: Apr 2002 Location: Puerto Rico, USA Posts: 13,537 Thanks: 0 Thanked 0 Times in 0 Posts
Thanks knowbody
Just keep on reading the forum. There's a ton of information and a lot of guides, available on the main page at www.kvcd.net on the right side (under Articles)

Have fun
-kwag
#7
04-23-2004, 11:41 PM
 kev23m Free Member Join Date: Feb 2004 Location: India Posts: 185 Thanks: 0 Thanked 0 Times in 0 Posts
I have always used Manual File prediction and never ever has it gone wrong.

Have the Sampler.dll in plugins.

Add "Sampler(100,length=24)" at the end of the script.

Formula = Total Frames(taken from Tmpgenc)/100/23.976(ntsc) or 25(Pal)*0.98 = X

Video Size Required = 800 less Audio Size less 10 mb.

Video Size / X

This has never failed me as yet.
#8
04-25-2004, 01:50 PM
 rendalunit Free Member Join Date: Apr 2002 Location: san jose, Ca Posts: 1,148 Thanks: 0 Thanked 0 Times in 0 Posts
that's the "old school" method for the times when there was no CQMATIC
#9
04-27-2004, 07:49 AM
 kev23m Free Member Join Date: Feb 2004 Location: India Posts: 185 Thanks: 0 Thanked 0 Times in 0 Posts
Quote:
 Originally Posted by rendalunit that's the "old school" method for the times when there was no CQMATIC
Yep i guess this is what all prediction tools are based on so if you are not to lazy , run some samples and you get it right.
#10
08-06-2004, 08:10 PM
 Shibblet Free Member Join Date: Jul 2004 Location: Wasilla, Alaska Posts: 108 Thanks: 0 Thanked 0 Times in 0 Posts
Xvid's codec has an option of 2 pass mode where you actually put the video file size in for your second pass. It scans through the movie for the first pass, and on the second pass, it makes a video file that is very very close to your input file size.

Isn't there any way of adding a 2 pass mode that does one prediction pass and a CQ or CQ_VBR 2nd pass that adjusts itself to fit your specific file size for KVCD's? The quantizer matrix is very similar to Xvid.

I don't know if this has anything to do with MPEG-1 and MPEG-2 being that much different from MPEG-4, but there has got to be a fool-proof way of predicting file size.
__________________
Well if it's not a wolf, then it's a damn big dog.
- Rabbit, from "The Fable"
#11
08-07-2004, 04:56 AM
 Dialhot Free Member Join Date: May 2003 Posts: 10,463 Thanks: 0 Thanked 0 Times in 0 Posts
Quote:
 Originally Posted by Shibblet Isn't there any way of adding a 2 pass mode that does one prediction pass and a CQ or CQ_VBR 2nd pass that adjusts itself to fit your specific file size for KVCD's?
CQ (or CQ_VBR) are QUALITY based. If there wheer any mean to have a QUALITY based encoding where you can control the bitrage allocation all encoder would have implemented it.

Quote:
 I don't know if this has anything to do with MPEG-1 and MPEG-2 being that much different from MPEG-4, but there has got to be a fool-proof way of predicting file size.
All mpeg1 or mpeg2 encoders also have 2pass encoding mode ! Tmpgenc has one of course. But is it NOT quality based anymore and the result is really worst (in words on quality).
#12
08-07-2004, 09:46 AM
 black prince Free Member Join Date: Jul 2002 Posts: 1,224 Thanks: 0 Thanked 0 Times in 0 Posts
@knowbody,

Have you tried CQTester with Calcumatic I get 97% or closer filesize
predictions with this method:

Quote:
 Here's a quick step by step for 352x240 for 1 CDR using CQTester: 1) DVD Decrypter (movie only) 2) DVD2AVI v 1.77.3 (check for force film) create .d2v and .ac3 files 3) Headac3he - create .mp2 with you desired audio bitarte (keep track of file size) 4) Moviestacker - select resolution and overscan for avs script 5) CalcuMatic - select .d2v and keep track of min bitrate and avg bitrate make note of predicted video and audio file sizes to compare with final files. 6) Tmpgenc - select avs script and kvcd template a) set max bitrate 2000, min bitrate (from calcumatic) b) open source in adv tab and using cut/edit select the whole movie c) go to file save to "tpr text file" for .tpr d) close Tmpgenc 7) CQTester - a) select .tpr file from Tmpgenc b) set no. cd's to 1 c) set MB pro CD to 800 d) set Audio MB to .mp2 file size from headac3he e) set mulitplex offset to 0 f) click on "Opt + Enc" button 8 ) wait until CQTester has finished and check file size created for .m1v 9) Use "Mplex" to mutliplex audio + video 10) Burn mpg file from mutliplex Note: My results for actual video file size were 2% to 3% less than predicted for five movies I tested with various lengths of 90min up to 180min.
-BP
#13
08-08-2004, 02:59 AM
 incredible Free Member Join Date: May 2003 Location: Germany Posts: 3,189 Thanks: 0 Thanked 0 Times in 0 Posts
Borax also got an english parte of his cq tester page:
http://www.boraxsoft.de/CQTester_eng.shtml

 Similar Threads Thread Thread Starter Forum Replies Last Post kwag Avisynth Scripting 84 07-30-2004 09:12 PM DKruskie Avisynth Scripting 4 05-14-2003 06:44 PM SansGrip Avisynth Scripting 165 12-28-2002 07:29 PM black prince Avisynth Scripting 0 11-22-2002 09:58 AM kwag Video Encoding and Conversion 0 11-18-2002 06:25 PM

All times are GMT -5. The time now is 01:15 PM  —  vBulletin © Jelsoft Enterprises Ltd
 Top of Page  -  Site Home  -  Forum Home  -  Archive  -  Forum Policies