DVDREasy: OPV in HC

dvdreasy · #1 05-20-2005, 12:11 PM

Hi,

I have almost ready OPV support for HC

The result is excellent although the final size (after authoring) is a little undersized (4.15 GB). I think I have to teak it a little to produce a more closer result (4.30 GB)

I will do a couple of more tests and post a new release soon . . .

Cheers,

DVDREasy

Prodater64 · #2 05-20-2005, 12:25 PM

Quote:

Originally Posted by dvdreasy

Hi,

I have almost ready OPV support for HC

The result is excellent although the final size (after authoring) is a little undersized (4.15 GB). I think I have to teak it a little to produce a more closer result (4.30 GB)

I will do a couple of more tests and post a new release soon . . .

Cheers,

DVDREasy

Can you explain how do you do it?

dvdreasy · #3 05-20-2005, 03:29 PM

Quote:

Originally Posted by Prodater64

Can you explain how do you do it?

Hi Prodater64,

here is how I'm doing it :
1) put all the right settings for 2 passes VBR
2) encode 1% of the movie with avisynth command “SelectRangeEvery(1200,12)”
3) grab from HC log file the average quant value
4) encode the full movie in OPV with that average quant value

Note: For using this method you have to force the setting "*GOP 12 1"

In fact, after take the average quant value from HC log I multiply that value by a corrective factor (I'm working on it - for now is 0,9980) but, from the tests I did the corrective value can be lower

Cheers,

DVDREasy

kwag · #4 05-20-2005, 05:35 PM

Quote:

Originally Posted by dvdreasy

Hi,

I have almost ready OPV support for HC

Please don't drop the regular 2-pass support

-kwag

dvdreasy · #5 05-20-2005, 05:37 PM

Quote:

Originally Posted by kwag

Please don't drop the regular 2-pass support

Of course not . . . thats for sure

incredible · #6 05-20-2005, 05:45 PM

Are you aware that selectrangeeverys first integer is not the number of the samples but the period (distance in frames to next slice).
So that wont be constant when doing diff. movie source lengths.

Sampler() from SansGrip used as first integer the number of samples in a lenght of the second integer.

If you want to use a percentual movie prediction using selectrangeevery then you could try this ...

Code:

function slicer(clip input, int percent, int Goplenght, int Gopmulti, string "ping")
{
PercCount = (Framecount(input)/100)*percent
period    = int(Framecount(input)/PercCount)*(Goplenght*Gopmulti)
input=(ping == "pong") ? input.trim((period/2),Framecount(input)):input
selectrangeevery(input, period, (Goplenght*Gopmulti))
}

Usage:
slicer(2,15,1,"ping")

2= 2% to be predictioned of the whole movie
15 = Length of the GOP in frames of the later Encoding
1 = GOP Lenght multiplier (for longer slices)
"Ping" = A Ping will be done

"Pong" would mean a Pong would be done WHERE the offset is done automatically by the function internally. The offset is calculated so the pong prediction samples exactly between the ping slices.

Thats based on teh PingPong method but how ever you can see the math and adapt it on a single prediction without offset where a user could also tweak the prediction %age in your appl.

dvdreasy · #7 05-20-2005, 06:09 PM

@incredible

THANKS !

Encoder Master · #8 05-21-2005, 02:09 AM

Quote:

here is how I'm doing it :
1) put all the right settings for 2 passes VBR
2) encode 1% of the movie with avisynth command “SelectRangeEvery(1200,12)”
3) grab from HC log file the average quant value
4) encode the full movie in OPV with that average quant value

Like I do but it's not the same Q value then the 2pass encode like you said. It's under that and also prediction take time and so I think a support for CQ isn't good. Because 2pass gives me the best quality and the perfect filesize.

And also the q isn't constant if you check the max bitrate.

incredible · #9 05-21-2005, 02:27 AM

Its a nice "theory" using the Q float value out of a sliced 2-pass pre-encode, but as we know all encoders do at least sometimes change in their Q in a minimal way even if encoding the same! source twice.

So why not doing or at least trying the way using a prediction as we have been used to. No matter if a "simple" prediction ... my "ping Pong" one or a "long/short" one.

Encoder Master · #10 05-21-2005, 02:36 AM

Quote:

Its a nice "theory" using the Q float value out of a sliced 2-pass pre-encode, but as we know all encoders do at least sometimes change in their Q in a minimal way even if encoding the same! source twice.

Yes, your absolutly right. But it's a nice fast way to get a very close value. And then I think your Method would be the best choise to predic the value. But as I said 2Pass is the best and why so complicated?!

dvdreasy · #11 05-21-2005, 09:36 AM

Hi,

Although I know that this isn´t the PERFECT way to achieve the optimal Q value, its VERY fast and gives a very close value . . .

Now, I'm using a corrective factor of 0,97 and all the movies I did are ON TARGET :

Did 3 movies (01h:56m, 02h:10m and 02h:43m) and the final size (after authoring) is between 4,29GB and 4,32 GB

Like I said the 2 pass option continues to be available.

Cheers,

DVDREasy

Prodater64 · #12 05-21-2005, 11:30 AM

But gop 12 1 is for ntsc encodes, isn't?
What for PAL ones?

I think if he reach a more or less accurate method, it is worth to use it, as 2 Pass VBR takes a lot of time, and CQ quality is also good enough.
(and I already know that time doesn't matter)

incredible · #13 05-21-2005, 12:17 PM

A 12 Gop is a "cream" Gop, means for best Quality IF enough Bitrate is available.

Usual: PAL = 15 and NTSC = 18

DVDreasy meant that IF the length of the slices in the 2pass Prediction is 12 then also the later encoding in the GOP has to be 12.

But slices using 15 is ok (its the very limit as the prediction ALSO has to estimate the encoders Motionvectors handling and therefore each slice shouldnt be too short. Often I use a Gop multiplier of 2 in my Predictions, thats why I integrated it in slicer()

dvdreasy · #14 05-21-2005, 12:57 PM

Quote:

Originally Posted by incredible

A 12 Gop is a "cream" Gop, means for best Quality IF enough Bitrate is available.

DVDreasy meant that IF the length of the slices in the 2pass Prediction is 12 then also the later encoding in the GOP has to be 12.

Yep, you'r right !

Look, I'm not a expert in encoding stuff . . . Just a curious.

I didn't expect so much posts about this thread !

KVCD.NET forum really ROCKS !!!

Cheers,

DVDREasy

kwag · #15 05-21-2005, 01:27 PM

Quote:

Originally Posted by dvdreasy

Look, I'm not a expert in encoding stuff . . . Just a curious.

Hey Pedro,

If you can code an application like DVDREasy, than makes you fall under the "Expert" category immediately

Quote:

I didn't expect so much posts about this thread !

KVCD.NET forum really ROCKS !!!

-kwag

rds_correia · #16 05-21-2005, 06:06 PM

Hi Pedro,
OT on-
Let me congratulate you for such a wonderful tool you're working on.
It is nice to see so many coders working in the forum and (for me) especially the TUGAS (a.k.a Portuguese ones

).
OT off-
You are using the exact same tools that I was using before I stoped visiting the forum for a while.
I just wonder if HC has gain some speed and if Muxman has already accomplished full compliancy?
Anyway I do trust these tools will work wonders under your GUI.
Do keep the options open as for OPV or 2-Pass.
Just remember: here at kvcd.net we give more credit to quality than to encoding speed

TIA

Cheers

dvdreasy · #17 05-21-2005, 06:39 PM

Quote:

Originally Posted by rds_correia

I just wonder if HC has gain some speed and if Muxman has already accomplished full compliancy?
Anyway I do trust these tools will work wonders under your GUI.
Do keep the options open as for OPV or 2-Pass.
Just remember: here at kvcd.net we give more credit to quality than to encoding speed

Hi Rui,

Hc is a little faster in 0.14 beta release. DVDREasy use MUXMAN since 0.13 release and I never had a complyance problem.

The open options in DVDREasy for HC are the ones you can change in Encoder settings (see the picture in this thread). other settings are passed to HC automaticly (interlaced, TFF, Aspect rate, scan method, etc)

Remember, this is an EASY one click tool to be used for all the people (with or without encoding knowledge).

Cheers,

DVDREasy

audioslave · #18 05-23-2005, 04:25 PM

Wow! Thanks a lot for this excellent tool!

I don't know if this already have been answered, but is this OPV version out yet?

dvdreasy · #19 05-23-2005, 04:33 PM

DVDREasy 0.6.3 will be available SOON

incredible · #20 05-24-2005, 04:19 AM

Hey DVDreasy ....

OT/
Maybe the following wishes/requests already have been mentioned but here's what I would like to have:

a) Option to also include a second Audio Track as maybe many users beside their native Language Track would like to enjoy having the orig english Track. I think there its no Problem to mux/author a DVD containing 2 Audios (Look at the .exe's/Binaries provided wth DVDauthorGUI).

b) Being able to encode the Videotrack just as single ES, as I often like to author a DVD by myself so no DVDauthoring via DVDreasy needed (but maybe I missed something as I hadnt the time to get really deep into your GUI

.... So DVDreasy could be also just used for prediction purposes on HC.

BTW: If I remember there was mentioned somewhere that the avs script import in HC causes Problems when using Win2000?? As Im not able to import avs scripts

HC freezes. No matter which script (AVS 2.56 used and latest HC Release) ... I havent been that much in the HC and also DVDreasy Threads so maybe that also has ben diskussed/solved before