I modified this script again to show a recommended cq to use using the simple formula new cq= (wanted sample size*cq used)/sample size). After you've encoded a sample put the size (mb) and cq at the top of the script and open it in WMP. I haven't decided whether this is totally useless yet because it's probably easier to use a calculator :lol: :lol:
Code:
Prediction("Band of Brothers", 1, 0.99, SampleSize, CQ, MinAudioBitrate=320). Sampler(length=24) edit: I cleaned up the code a little bit |
Hi Kwag
Just wondered if the formula for prediction: Predicted MPEG size = (( Total frames/MovieTimeInMinutes) / 24 ) * MPEG sample file size is the same for NTSC and PAL. Cheers Jeff |
Quote:
-kwag |
Quote:
|
evening guys.. rendalunit,
can I ask a really shtupid question..?? ..?? can this now be incorporated in FitCD's calculations ?? ..or is this not necessary ?? Told ya it would be shtupid, he he.. Have a good evening all. -vhelp |
Quote:
ren |
here spot, i mean rendalunit, he he..
I was having a looksees at FitCD's code and doing a few what-if's while I was reading some of these posts here (yestday I think) anyways.. then I came accross this post, and thought I'd just ask a really shtupid question..?? ..?? w/out actually thinking it out through.. hence the shtupid'ness of the question :roll: (soon or later, I have ta ask a shtupid question) after looking around where's at the source code, (I can't remember) I stopped and gave up on it for something more appealing (my project) but I remembered roming around in it, and this thread jarred my memory. Thats all. -vhelp |
Quote:
"...is much easier than using a calculator ... ....avs into Tmpg and hit "preview"- ..." "Thanks Graal_CPM, Mauddib, Racer99, Kwag-- you guys rock!!!!" you too rendalunit! :wink: |
Sorry for the long post :? I copied and pasted the code by rendalunit and ran it in WMP and got an error stating:
http://www.digitalfaq.com/archives/error.gif This refers to the line I highted in red. When I comment it out the script works fine. Am I missing something?? Below the quote is my script. Also, where do you put in the CQ and sample size because it states "After you've encoded a sample put the size (mb) and cq at the top of the script and open it in WMP." Thank you for your time and patience :) Quote:
Code:
LoadPlugin("C:\WINDOWS\Desktop\VCD Filters\MPEG2DEC.dll") |
SampleSize and CQ need defining with =x added to them
i put the Prediction function variables at the top of the script so it's easier to edit. Code:
LoadPlugin("C:\encoding\MPEG2DEC2.dll") CQ=0 with the sample size in mb and CQ value used: i.e. SampleSize=7.143 CQ=76.5 |
ok I copied the above script and substituted my filters in instead but the sample size of my first run with a CQ of 60 is giving me a 1.3MB size :?: And what usually takes about 15min to do a sample is taking like 2 mins.
Also, where do you comment out the prediction function when you're ready to encode the whole movie :?: Thanks Here it is Code:
LoadPlugin("C:\WINDOWS\Desktop\VCD Filters\MPEG2DEC.dll") |
Quote:
Quote:
Boy, do I feel like I'm back in school :lol: |
Quote:
Quote:
Quote:
-kwag |
FINALLY :!: :!: :yippie: It's working :)
Thanks rendalunit and kwag. http://www.digitalfaq.com/archives/error.gif Now I can go to sleep while it encodes :lol: |
Here's my 2 cents: I modified the prediction function to take 2 sample sizes as input. Provides for a bit better accuracy. I also called it predict2 so that it will not be confused with the original.
Code:
#------------------------------# |
Hyia Kwag.. Been a LONG while since I last posted here. I have been watching this thread for little while and am very interested in what I have been reading.. I understand most of it since the scripts seem to be doing most of the work. I do have a question tho. What if you are not ripping but recording straight to MPEG-2 (DVD specs) and re-compress from there? I have a Dazzle DVC-II and recording Indy Jones and the Temple of Doom. I maxed out all my settings for maxium quality when recording. I am trying to adjust your scripts for the KDVD D1 template and I seem to be overshooting my target filesize by at least 20 MBs!! I was using CQ_VBR. I can assume that this is incorrect now. I am re-encoding right now at just CQ. Since this is for KDVD, I set the number of CDs at 2 (this is pretty close to my final target filesize). The formula states that I should use CQ 29.3(ish). That seems awefully low but will try it next. Is there any special modifications to make to allow for KDVD D1 (or even half D1)?
My apologies for the long winded post :) |
Hi SodGawd,
Long time no see :lol: Are you sure you're using the latest KDVD templates :?: It seems to me that the value of CQ you calculated is extremely low for CQ templates. That would be about the right CQ for the old CQ_VBR templates. So please check your template again agains the ones currently posted :) -kwag |
Quote:
Quote:
Also, since I am compressing VHS (29.97fps/interlaced) I am trying to see if I get a noticable diff between IVTC and Interlaced. |
Newton's Method to predict CQ
Hi, Kwag. Long time no see ;)
I've been doing mostly CCE lately and playing with Q prediction. I implemented a variant of Newton's method to find roots of an equation to find a suitable Q for RoBa in 3 to 4 tries, outsmarting binary search. I recently expanded my simulator spreadsheet to support TMPGEnc's CQ mode. I though you might be interested in taking a look at this thread in Doom9's forum: http://forum.doom9.org/showthread.php?s=&threadid=60191 This stuff might turn out to be useful for CQMatic. Best regards, |
Hey thanks r6d2 :D
I'll take a look now! -kwag |
Site design, images and content © 2002-2024 The Digital FAQ, www.digitalFAQ.com
Forum Software by vBulletin · Copyright © 2024 Jelsoft Enterprises Ltd.