Filtering help needed

[x264]

Heck, you could even try it out and see if your interlacing is gone. What a novel idea!

Combed frames are still present.

[manono]

Should i change any of these settings before saving the file as .avi?

I don't know what your final output is to be, but it's most likely the same as your source (YV12). So, I (most of the time, unless the next step is using CCE for MPEG-2 encoding, in which case I might set it for YUY2) set Mode for YV12. I also check 'Use Multithreading' as my CPU is multithreaded.

You still have the interlacing? What kind? The little bits you see from flapping mouths in animations? Then you have several choices. You can tighten up the CThresh, use override files, put on VinVerse afterwards (which you said you don't want to use), ignore it because it probably can't be seen when watching it at full speed, substitute a good frame from nearby using FreezeFrame, interpolate a new frame to replace it, or other things I haven't thought of.

Do you have an untouched sample showing the problem?

[x264]

For heavy filtering and 2pass encoding, it might be a good idea to first encode the output to a lossless format and then use this output as source instead of running the filtering path 2 times (during 1st and 2nd pass).

This is what i want to do. What should i do after creating the .avi file?

[manono]

This is what i want to do. What should i do after creating the .avi file?

Then you create a new script using your lossless AVI, something like:

AviSource("Lagarith.avi")

and feed that to your encoder.

[x264]

Ok. thank you for your help

[x264]

The lossless .avi file is around 16 gb. Why is the filesize so big compared to the DVD vob files?

[Selur]

What happens is:
DVD content get's decoded to RAW uncompressed video data, this raw video data get's reencoded with a lossless codec, since the compression ratio of lossless content isn't really high you end up with larger files.
-> you might want to start reading up some of the basic of video compression,...

[x264]

What should i do in order to improve the encoding speed in x264?. I always use the preset : Slowest.

[Selur]

use a faster preset
slower means:

- slower:
                                    --b-adapt 2 --direct auto --me umh
                                    --partitions all --rc-lookahead 60
                                    --ref 8 --subme 9 --trellis 2

source: x264 --fullhelp
so lowering in example the lookahead value or subme or setting trellis to 0 might be the first things I would try if I want to basically stay with slower.

[jagabo]

Use a faster preset? Use Multithreaded AviSynth (and specify the number of threads)? Use simpler filters in AviSynth?

[x264]

Will there be any benefits in using multithreaded avisynth for core2duo processor?

[Selur]

In general it totally depends on the filters you use, but most filters do gain from multithreading.
To check out if it might be worth the effort run your script through AVSMeter, if the speed it is decoded is faster than you encoding, than probably not.

[x264]

Ok. Could you tell me what changes i need to make in this script for multithreading? The guide on the doom9 forum is confusing.

MPEG2Source("C:\Users\Abhijith Nair\Desktop\VTS_01_1.d2v", cpu=0)
trim(2157,31667)++trim(33226,33587)
crop(8,0,-6,-0)
Dehalo_Alpha_mt(darkstr=0.0, brightstr=1.3)
Dither_convert_8_to_16()
Dither_Quantize(10, reducerange=true, mode=0)
Dither_Out()

[Selur]

it's not just changing the script, you also need another avisynth.dll and sometimes other filter dlls.

Your script will later probably look like:

setMtMode(5,2)
MPEG2Source("C:\Users\Abhijith Nair\Desktop\VTS_01_1.d2v", cpu=0)
setMtMode(2)
trim(2157,31667)++trim(33226,33587)
crop(8,0,-6,-0)
Dehalo_Alpha_mt(darkstr=0.0, brightstr=1.3)
Dither_convert_8_to_16()
Dither_Quantize(10, reducerange=true, mode=0)
Dither_Out()

[x264]

I changed the original avisynth.dll with the multithreaded avisynth.dll. Can you explain the first and third line in the above code?

[x264]

Avsmeter crashed when i ran the above script.

[Selur]

first line sets mode 5 and number of threads to 2
third line changes mode to 2.

as a rule of thumb:
a. source filters should always be used with mode 5
b. filters that use the gpu should always used with mode 5
c. nearly all other filters run fine with mode 2

Avsmeter crashed when i ran the above script.

no surprisingly, since you are probalby using the lates Avisynth 2.6.04 alpha without the appropriate mt_masktools-26.dll version (http://forum.doom9.org/showthread.php?p=1619023#post1619023) like I wrote "and sometimes other filter dlls."

[x264]

without MT


MT

[Selur]

so using avisynth mt will speed up your avisynth script, now the question is, will it hurt or help you encoding speed, since more cpu power an RAM will be used for decoding.

[x264]

so is this going to slow down my encoding speed?

[Selur]

it might, depends if the avisynth script was more of a bottle neck than the encoding through x264 -> instead of asking you could simply start an encoding and see for yourself

[x264]

Is there anything that can be improved in the above scripts besides the filtering?

[Selur]

If you use a system with a CUDA capable NVIDIA GPU (preferable with a VP5 chip) you could buy&use DGDecNV instead of MPEG2Source.
Other than that, there is nothing more in you script other than filtering,...

[x264]

I have DGDecNV, i don't use it much because of the CUDA error. During the encoding process only 50% ram i being used, how do i make the encoder use more ram for the encoding process?

[Selur]

No clue what you mean when you write 'the CUDA error', but it guess you either do not have a license or you do not have a CUDA capable graphic card.

[x264]

No clue what you mean when you write 'the CUDA error', but it guess you either do not have a license or you do not have a CUDA capable graphic card.

A green screen pops up when i preview the script in Avspmod. I'm having the license.

[Selur]

then it might be a good idea to contact the author (neuron2, which is also a member here and on doom9)

[jagabo]

I have DGDecNV, i don't use it much because of the CUDA error. During the encoding process only 50% ram i being used, how do i make the encoder use more ram for the encoding process?

You don't need to use more RAM. What you want is to make sure CPU usage is near 100 percent while encoding (use Task Manager). If AviSynth is still a bottleneck increase the number of threads, SetMtMode(mode ,threads). If you plan to use the computer for other things while encoding set x264's priority to low.

[x264]

How do i know which mode i have to use?

[Selur]

How do i know which mode i have to use?

are you even reading what we write here or are you just randomly posting questions?

as a rule of thumb:
a. source filters should always be used with mode 5
b. filters that use the gpu should always used with mode 5
c. nearly all other filters run fine with mode 2