CQMatic and TMPGEnc EXpress 3.0
 

From what I can understand, both TMPGEnc 2.5 and TMPGEnc EXpress 3.0 can be installed and run independantly. Yes? If so, why not run CQMatic with 2.5 and use the resulting CQ value in 3.0? This is assuming you create a new kvcd template for 3.0 with settings from your favorite existing template and import the matrix(I saw it posted somewhere?). At least it can be considered a starting point.
From what I am reading, 3.0 is impressive. Did I read correctly?

Thanks

Re: CQMatic and TMPGEnc EXpress 3.0
 

I'm almost positive you'll get a different target file size, if you use the same CQ in both encoders :roll:

-kwag

Hi,

I confirme you that with the same CQ, you get different filesize! I did the test, and there is a big difference! Something like 3 points to get the same filesize.
When you know that the CQ/filesize curve is not linear, you understand that it's imposible to get the CQ in tmpgenc 2.521 and use it in tmpgenc 3.0...

Salu2
Fabrice

What are the chances of seeing a prediction tool like CQMatic for 3.0? I understand that there are no text projects which makes the task more difficult.
Has anyone ever taken the time to map the CQ curve and make a graph? Yeah, I know, another big task.

Thanks

Yes, I did it a time ago...

Have a look there: http://www.kvcd.net/forum/viewtopic.php?t=5609

The chanes to get a tool like cqmatic for tmpgenc 3.0? 0!

Salu2
Fabrice

All depend... how much do you offer ? :-)

I can only offer my gratitude :wink:

Thanks for making the graph Fabrice. The repetative nature is obvious. I will spend some time reading the entire thread.
I would love to ask the maker of tmpgenc how he/she ended up with such an oddly unlinear scale? I know some of the mods have tried to get him/her to rewrite the cq part of the prog. I'm guessing it is not easy or it would have been done. I'm not asking for linearity, but even if the scale was logarithmic, we would all have it much easier.

Thanks

Manual prediction works well with TMPGenc3XP.

Just tried it and its real great for mpeg-2 , not tried Mpeg-1 as yet.

Did a 179min movie with Min Bitrate : 300 Max : 2500 CQ : 73.11 Aspect Ratio : 16:9

Output size for video 666mb

Was nowhere close to the same bitrate and Cq using TMPGenc Plus.

I did select the options under " Others " and just copied the kvcd matrix.

tested this encode on a dvd player it works fine till i forward once i do that then blocks appear and have to restart the movie.
This is in mpeg-1 or Mpeg-2 done with TMPGenc3Xp.

Can you point me to the manual prediction you are refering to?

Thanks

Old school

no of frames / 100 / 24 for ntsc or 25 for pal *0.98 = x
File size required / X = Sample Size

Add line at the end of the script
Sampler(100,length=24) AND encode the sample to match sample size.

Just ran a few tests , increasing max bitrate has no effect on quality n size just the min bitrate in Tmpgenc3XP.

That means that tmpgenc never need to reach the max you have set for encoding your sample.

1/ either your sample is a "slow" one and that is not the better part to encode if you want to test the encoder strongness
2/ or tmpgenc 3.0 does not use the whole bitrate we ask it to use and that is REALLY bad thing.

That would also be the reason for the smaller size. Just doing some tests with similar setting in Tmpgenc plus and checking.



Just did the same sample with Tmpgenc Plus and increased max bitrate from 1800 to 7000 the sample size increased by 0.26 , difference in sample sizes btween Tmpgenc Plus * 3XP is 1 mb.

Okay, so it's surely the sample that is not the "best choice" you could done. Try with an other, more "tricky" to encode.

I increased the sample size to 200 and the size was 14.2 when it should have been 15mb.

So i cut out a song & dance sequence from the movie and tested with that , here with max bitrate at 7000 the size was 7.78 it was supposed to be 5.95 and with 2500 it was 7.78. So no change at all.

Gonna try and encode the whole movie and see the outcome , will let ya know.

Did the whole movie .

Min Bitrate - 0 Max - 7000 Cq 74.56 Runtime 2h34m Output size - 652mb.

I had burned a sample and checked on the player and the picture was real good.

I never knew what cq matic was still dont realy :P but i think i got an idea baiscly tells u an extremly good questament with the frame count a and cq value?Am I correct? Im still reading links in this post and all, But I'd like to see more on this topic myself im learning alot im pretty much self taught so ar ever since i joined here ive advanced alot faster and suprizing myself!!!
Thanx for the hard work and for sharing it!!
-Trigger911 :wink:

CQ Matic wont work with Tmp3XP as we cant save a text based project.

When your testing try and increase max Bitrate and see if there is any difference in size at all.

ok ill try that at a l8r time im trying to finsh some stuff so i can watch my movies again y do companys put crappy readers in dvd players???

ps how would i go about cutting a peice off a vob and how long should it be? i dont mind taking the time to see id like to know my self or i can upload the diff setting on a buddies host so other my look at them.

To cut the peice off a vod i use Dvd2avi it saves a d2v project .