My way of using CCE (with DV codec. Is this OK?)

[jmartina]

I am reading this thread and other in this forum, and I can see
that all you talking about avisynth, batch files etc.

The method that I am using is very simple:

1. DVDdecrypter
2. VOBEdit to demux all audio and subpic streams that I need
3. IFOEdit to save celltimes to file
4. DVD2AVI or xMpeg to covert movie to Microsoft or MainConcept DV
5. CCE with DV from step 4. using instructions found in this forum
6. IFOEdit authoring

The result is FANTASTIC!

For example LOTR (wich is 171 min. long) played on my 72cm Sony
is so good that i wasn't shure is that original or copy inside.

Same result I get for Gladiator and StarWars Episode II.

No problem at all, so why all this mess with .avs, scripts, etc???

[69Mws]

The reason for AviSynth is, that you have a lot of possibilities, options, filters, etc.

For example: i had an avi file with 23,976 FPS and i encoded it in CCE with an avs-script, which changed the framerate to PAL (25 FPS) and resized it properly for anamorphic (PAL)DVD and this all at once....okay, this was a quite strange example, but it shows you the possibilities of AviSynth :)

And it's said to be very fast at frameserving.

What speed does CCE indicate with your method? What means "...convert movie to Microsoft or MainConcept DV"?

Greetz
69Mws

[jmartina]

DVD2AVI or xMpeg asks you to provide codec for conversion, right?

I give them to use Microsoft DV codec or MainConcept DV codec (Digital Video).
So my movie is converted from MPEG (VOB) to DV AVI file (approx. 12GB per hour
so my LOTR was some 36 GB). I am not resizing or croping anything.

Then I feed CCE with that AVI using bitrate calculated according to needed size,
and using parameters needed for DVD (mostly using instructions from this site).

My configuration is:

Gigabyte 8IHXP
P4 2.4GHz
512MB RAMBUS PC1066
IBM GXP180 180BG 8MB cache
Gainward GeForce4 Ti4600
Windows XP Pro SP1

Conversion speed:

DVD2AVI using MainConcept DV codec 40-45 FPS
CCE average 1.85

By the way, I am using only PAL DVDs.

[atreides93]

The reason people use AVIsynth etc, is to avoid that huge step you're doing where you convert the mpeg-2 to DV avi!! That seems like an awful waste of time/effort to convert it to AVI just so you can feed it to CCE.

[69Mws]

Yes, right, that's exactly the thing what's called "frameserving" to CEE. You avoid creating a very huge avi File and spare a lot of time. That's the trick with AviSynth.

For DVD re-encoding you just need a simple script, because no resizing is needed. Here's an example (AviSynth must be installed and you have to place the mpec2dec.dll manually where you like it):

LoadPlugin("K:\WINNT\system32\MPEG2DEC.dll")
mpeg2source("F:\cube\2_Streams\cube.D2V")
ResampleAudio(44100)

As you can see, only the project file is needed from DVD2AVI and creating a Project File takes only a few minutes!

I have mostly a speed of 0.6 with CCE 2.50 running on:
P III 1 GHZ
640 MB RAM
NTFS Partition on Seagate ATA100 Drive with 7200 rpm

Greetz
69Mws

[jmartina]

Let's say that huge DV file is not the problem, does CCE runs slower
with frameserving than directly with DV AVI file?

[69Mws]

I can't answer this, but 'til now i have never heard, that frameserving with avs is considered to be slow. I would suggest, that you just give it a try. Install Avisynth and get the mpeg2dec.ll or mpeg2dec2.dll.

Creating a project file takes only very short time in DVD2AVI and the avs script is also simple, as mentioned above.

Even it would be a bit slower with avs, you have to be aware, that you spare a lot of time by not creating this huge avi.

Greetz
69Mws

[jmartina]

Thanks, I will try and post results.

[auenf]

Originally posted by jmartina
Let's say that huge DV file is not the problem, does CCE runs slower
with frameserving than directly with DV AVI file?

i think the frameserving does work out slightly slower than encode from DV AVI, but the convert to DV + CCE encode total time will be slower than the frameserve directly to CCE encode.

having said that, the ->DV->CCE would be more reliable than the frameserving method, which has been known to cause a few problem (slices, CRC error, to name two ;) )

personally i encode captures of TV, and use DV avi with CCE all the time ;)

Enf...

[jmartina]

Here are first results from my testing:

120 min movie 2.35:1 so it is larger than 4.7GB even after
removing unneeded audio and subs.

Configuration (different than mentioned few posts before):

Gigabyte 845PE Ultra
P4 2.4
1GB DDR333
OnBoard Promise RAID 2x120GB IBM


I will not count jobs that are common (decrypting, etc.)


Method 1:

DVD2AVI - aprox 45FPS time: 67min (using Canopus DV codec)
CCE - aprox 1.85 time: 65min

Total: 132min



Method 2:

Frameserving using VFAPI - very fast
CCE - aprox 1.2 time: 100 min



So in my testing (at least on this machine) CCE speed is significantly
reduced. Total time is better, but:

I have tried three movies using both methods,
and on Star Wars Ep2 I had "Too many frames dropped" error in IFOEdit authoring
using method 2. I tried one more time, and it was same. With method 1 there were
no problems at all.

[69Mws]

Frameserving via vfapi is slower as far as i know.

Try just loading the DVD2AVI Project File (.d2v) with an avs script into CCE --> see my example above!

Greetz
69Mws

[Arky]

Don't forget that if you are doing 3 passes or more, and are using filters (and only if you are using filters!!), then it might be quicker to render a filtered video stream to DV codec (or HuffyUV would be quicker to render to, assuming you have enough HDD space) and then feed this to CCE. This is because you will only need to filter once, and this will potentially save time.

This is the issue I discussed at length, with LB, in the following thread:

http://forum.doom9.org/showthread.php?s=&threadid=36500&highlight=huffy

And some other related threads:

http://forum.doom9.org/showthread.php?s=&threadid=39263&highlight=huffyuv+speed

http://forum.doom9.org/showthread.php?s=&threadid=26056&highlight=huffyuv+speed


Arky ;o)

[bb]

Originally posted by Arky
Don't forget that if you are doing 3 passes or more, and are using filters (and only if you are using filters!!), then it might be quicker to render a filtered video stream to DV codec (or HuffyUV would be quicker to render to, assuming you have enough HDD space) and then feed this to CCE
I wouldn't recommend to save the filtered file through a DV codec, especially if you are using denoise filters, because with DV you'll introduce noise again. If you can save time this way (i.e. slow filtering) better use Huffyuv.

bb