Simplified RobShot? 1 movie per DVD....

[Roveer]

I've decided that I want to put a single movie per DVD and in certain cases, the movie will fit without modification. In other cases it will be bigger than the 4.38gb even after stripping and will require re-encoding.

Can I use a simplified RobShot method to do this? I understand that the 1st pass is to get the video information so you can decide how 'low' you can set the bitrate and still get a good picture. If I'm only trying to get a movie onto a single disk I'm usually looking way above that point. Here's an example:

single movie when stripped of everything but the movie and the english AC3 soundtrack comes out to 5.6gb Unfortunatly too big for a single DVDr, so I'll re-encode.

I've been using a bitrate calculation to give me approximate percentages to reduce the bitrate by and so far it's been fairly close.

Can I just do a one time multipass in CCE (3 passes), set my bitrates (usually 3000-6000 in AVG, 0 in Min and 7000-8000 in Max) and let it encode?

What Im asking is, there isn't any special magic in doing the 1pass encode, then switching to those values (above), and doing the additional 2 passes versus doing all 3 passes right away with the final settings?

Roveer

[SiC]

The first pass in Robshot's method is strictly for letting you see and make fine adjustments to your video. Basically to fit the most (minutes) onto a single DVD-R. But if your bitrate is above 3000kbps you more than likely will not have to do this first pass since you more than likely will not have any "grey bitches".

But just to let you know, the time to make a single pass VBR + 3 multipass VBR is no different than making just a 3 multipass VBR. The reason for this is because when you set to 3 multipass VBR, it will first create the *.vaf file which takes just about the same time as making a single pass VBR. If you do single pass VBR + 3 multipass VBR, in the first pass it will create the *.vaf file while it is encoding the video. When you select just 3 multipass it will have to create a *.vaf file first, then encode your video after.

[Roveer]

That's exactly how I thought it would work. Just wanted to make sure I wasn't missing any of the magic. Now, to play with the other end of the sprectrum and see how few bytes I can use to make a movie that still looks good, (see my other posts).

Thanks,

Roveer

[mikeathome]

Hi,

what I started to do now is:
- 1pass VBR Q:30 (!), or 40, Q = Quantization Factor (the lower the better; all for CCESP).
- let ehm run till 50% and see if the fianl movie will fit on the DVD-R
- in most cases (with 2.35:1 anamorph movies, ~2h length) the result will below 3.8GIG which will leave ~400MB for one 5.1 DD Audio track + ~100MB for menus.

You can't beat quality, even if you do a three pass VBR, you hardly will see any difference. This way you decrease handling time drastically.
With 4:3 aspect ratio titles you may need to go to 40-50 quantization factor. Still near original quality.

Only with really LONG (special edition) titles like T2 you'd better follow Robshots approach.

Most of my final editions are between 3-4 GIG in total and they all hold ~2h of movie and the original menus.

mike

[tyee]

Hi mikeathome

I tried your technique last night and it looks good. I was wondering what min and max bitrate your using, and what image quality priority and anti noise filter settings your using here?

I'll make a 3 pass also tonight and compare and see if I can see any difference.

Tyee

[tyee]

Hi Sic

[QUOTE]Originally posted by SiC
[B]The first pass in Robshot's method is strictly for letting you see and make fine adjustments to your video. Basically to fit the most (minutes) onto a single DVD-R. But if your bitrate is above 3000kbps you more than likely will not have to do this first pass since you more than likely will not have any "grey bitches".

Sic, are you saying that if I set my bitrate minimum to 3000kbps and max to about 8000kbps (for DVD encoding), then just select 1pass VBR, or 3pass VBR, that's it?

I should now get an excellent encoding and not have to worry about using Robshots 2nd step of searching for the grey parts?

Tyee

[Roveer]

See this post for a really great way to do what your asking:

http://forum.doom9.org/showthread.php?s=&threadid=18296

Look at SiC's suggestion in that thread. I've done 2 movies that way and it gives me just the right file size!!! I'm using Min=0 and Max=9000.

Regards,

Roveer

[Chumboy]

mikeathome....

followed robshots and your method and when i hit "encode" in ccesp, i save the ecl file, and then ccesp errors because it can't find the .vaf file.

how do you get the .vaf file?

thanks,
chum


----------------------

nevermind....
i didn't run the vifpset.bat file in VFAPI to install the codec.
i did and ccesp is working fine now...

thanks.

[tyee]

Roveer
If your specifying min,max, then using Sic's calculation for average, then you must be using 3pass VBR?

For 1pass VBR you can't specify average bitrate, only Q.

Tyee

[Roveer]

Correct, I use 3 pass VBR. I've been pretty satisfied with the results. I usuall set it up for overnight processing and finish with DVDMaestro in the morning. Like I said, I'm happy with the results.

Roveer

[SiC]

When I encode 2 movies to fit 1 DVD-R, I use 9 pass LOL!! When I'm doing a single DVD-9 to DVD-R, then I set mine to about 3 passes.

@Chumboy, do you get an error that says the Index file is too big??? Then it says it can't find the *.vaf file???? If so, that is a bug in CCE 2.62 that I've run into...... my solution was to use 2.64. Now if you get that error when you attempt to click on the Advanced button, that is normal because you have no *.vaf file. The *.vaf file will be created when you make a Single or Multi-pass VBR. If you are going to just do the multi-pass without doing the first Single pass, then you do not need to press the Advanced button.

[Chumboy]

SiC,

i am using 2.64.

no...didn't get the "file too big" error.
just "can't find *.vaf"

i installed the vfapi codec and it worked fine.

it's now on it's 5 hour 2nd pass.

[mikeathome]

Hi,

@Chumboy:
To get a .vaf you need to run an individual 1pass VBR (with Q factor setting) encoding. Ensure you check the 'Create VAF' just above the Audio (uncheck Audio btw.).

That's why I recommend, do it at right quality (Q: 30/40) straight from the beginning. You may save a hell lot of encoding time, if after that first run your video file is about size you wanted.
I set bitrates as follows:
min: 0
max: 8500
You hardly will ever reach 8500, so it's absolutely sufficient, but has one main adavantages (may be never of use for you, but you never know) if you later decide to add a more sophisticated audio/subpic stream you have some freedom. I've heard of an authoring app' which does not accept individual streams above 8500kbps for that reason.

mike

[tyee]

Hi Mikeathome

I tried the 3passVBR last night and compared it to the 1passVBR with Q between 30 and 40 and I agree with you. I only did a 2 minute clip to test but I could hardly tell any difference between them. I set the 3passVBR to have an average bitrate of about 5000kbps, and to get the same average bitrate on the 1passVBR (using bitrate viewer), a Q of 40 was required. There were a couple of bitrate peaks higher in the 1passVBR.

So my question is, if we use a Q of 40 with min. set to 0, and max. set to 9000, does this mean that the average bitrate of any movie we encode will be about 5000kbps? If yes, what's the point of using 3passVBR anymore?

Tyee

[mikeathome]

Originally posted by tyee
Hi Mikeathome

I tried the 3passVBR last night and compared it to the 1passVBR with Q between 30 and 40 and I agree with you. I only did a 2 minute clip to test but I could hardly tell any difference between them. I set the 3passVBR to have an average bitrate of about 5000kbps, and to get the same average bitrate on the 1passVBR (using bitrate viewer), a Q of 40 was required. There were a couple of bitrate peaks higher in the 1passVBR.

So my question is, if we use a Q of 40 with min. set to 0, and max. set to 9000, does this mean that the average bitrate of any movie we encode will be about 5000kbps? If yes, what's the point of using 3passVBR anymore?

Tyee

Hi Tyee,

thank for confirming my findings !
BUT (always a but), don't rely on that 5000kbps AVG with Q:40. What this 1pass VBR does, is really adapating.
Try yourself the same with an anamorph 2.35:1 and a 1.85:1 movie. You'll realize quite some differences. The 1.85:1 will be much higher.

About your question:
Try one of the old James Bond Movies. I had quit some difficulties to get it below 3.9GIG. Dr.No for instance. It would have required a Q: of 70 which is inacceptable. So, in this case I had to do a 3pass VBR to drive overall file size down.
The same you might be faced with superlong Movies (like Pearl Harbor or T2 Special edition). Here I still recommend a 1pass VBR at 40 (may be it fits ;-) if not -> go RobShot with 3pass and AngelMatrix. This way you don't waste your time.

mike

[tyee]

Hi mikeathome
Very interesting. I guess it would vary depending on aspect ratio and quality of movie.

Looks like the 1passVBR with Q = 40 is the way to start the encode process and then go Robshots if required just like you say.

BTW, are you bothering with all this matrix stuff. mb1 says you must use bbMPEG to mux the matrix into the header of the mpeg2 file for proper decoding (if using 2.50) but I can't seem to get it to work.

I've tried using the Andreas matrix for encoding without patching the header, and it did look better than the 2.5 standard, but I haven't tried Angel matrix yet. Is it better than Andreas?

Maybe your using 2.62 and don't have this problem? It writes matrix into header automatically doesn't it?

Tyee

[mikeathome]

Originally posted by tyee
Hi mikeathome
(...)
BTW, are you bothering with all this matrix stuff. mb1 says you must use bbMPEG to mux the matrix into the header of the mpeg2 file for proper decoding (if using 2.50) but I can't seem to get it to work.

I've tried using the Andreas matrix for encoding without patching the header, and it did look better than the 2.5 standard, but I haven't tried Angel matrix yet. Is it better than Andreas?

Maybe your using 2.62 and don't have this problem? It writes matrix into header automatically doesn't it?

Tyee

Hi,

yes I use CCESP 2.62, so no need to patch anything.
Can you post the Andreas Matrix, then we can compare. Do you have the AngelMatrix? You can compare on you own if you want.
http://www.dvd-svcd-forum.de/phorum/ikonboard.cgi?s=3c7cae19504dffff;act=ST;f=20;t=7

mike

[tyee]

Hi mikeathome
Andreas matrix is right out of Tsunami's patcher. Do you have it?

Tyee