Can cropping allow more bitrate for picture ?

[thensa]

People,

I am using Xmpeg 4.5 -> Videoserver 0.93 -> CCE 2.64 to create SVCDS. I have heard that if you crop the picture in Xmpeg and then letterbox it you can allocate more bitrate for the actual picture and instead of wasting it on the black bands.

The problem is that the only examples of which numbers to use that I have seen so far are for NTSC and I need the numbers for PAL (going from 720 x 576 DVD to 480 x 576 SVCD). Other things which may be useful - I have a 4:3 tv and am using a standalone player to watch the SVCD's on. I also read that this is only useful when converting widescreen movies, is this true (because there is no black bands on a 4:3 movie) ?

Kind Regards.

TheNSA...

[poopity poop]

hi, I'm a little confused as to your question but if I may conjector an entirely new way for you to encode:
You might want to look into moving on to the next step and use what advanced encoders use.
We load the vob's into DVD2AVI and save the project
Then load the project into VFAPI converter and create a fake avi that frame serves the frames via a small fake .avi
Then you can load that into vdub -> resize & crop -> and make an uncompressed avi using the codec huffYUV.
Then load that uncompressed avi into CCE and encode

You will find this method outlined for a DivX movie on my website, but there is a nice flow chart that I think you will find helpful.
This method will also be a lot faster because when you do your passes, it is only reading the uncompressed avi and doens't have to resize and crop it every single time

[Selur]

Hmm,.. the main problem about the numbers I see is that they can differ from movie to movie (since most movie's are not exactly 4:3 or 16:9 on the DVD).
About the whole idea of cropping&letterboxing for (S)VCD encoding: I'm not too experienced in this field, but I don't think it can save much data, since u need to stick to a fixed resolution and as far as I understood it, it might work on some movies where u don't chance the aspect ratio with this,...

Cu Selur

Ps.:
@popity poop:
1. how much space does a normal 100min movie normally take in HUFFYUV?
2. why don't u use avisynth(and corpp&resize there) and input the avs into cee,should work if I'm not wrong.

[poopity poop]

Selur:
On my site I outline a bunch of differnt paths. One of them (which I advocate by a red line in the flow chart) takes the path of avisynth. I always use avisynth, I merely suggested an easier to understand route for a newbie.
huffyuv file size varies linearlly with frame rate and resolution is different depening on the surface area of the frame (resolution multipied out), it varies linearly. It also varies depending on motion and brightness and I'll give you the range in MB/sec of video. This is a rough estimation, I did on a clip that may not be the same brightness, motion, things liek that, but it will give you an idea of how large the file size could be. If you want an approximate file size of a different resolution or frame rate just multiply it linearly by a ratio. If you have no idea what that means or how to do that, then take some math classes:

480x480 (SVCD 23.974fps, before pulldown.exe obviously)): 8-9MB/sec of video
480x576 (SVCD 25fps): 9.5-11MB/sec of video
640x272 (2.35:1 AR, 23.974fps) 6-7MB/sec of video
640x368 (16:9 AR, 23.974fps) 8.5-9.5MB/sec of video

All used a bicubic resize filter A=1.00

So just multiply those numbers by the lenth and you'll get a good idea of how large your file will be. I suggest taking the higher number to get the highest file size it COULD be.
Overall its much better to make a huffYUV avi because your 2-passes will go many many times faster, during the encoding process.
Also you may say, gee I don't have 60GB's to spare. IF you don't there is an option in huffy that will enable more compression but be slightly lossy, but I can barely notice anything, it makes for about half the file size for what is listed above. I.E., the 480x480 clip is about 4-5MB/sec. Also the better way: go to www.pricewatch.com and buy a hard drive you can get one for about $1/GB ( 60 bucks). A good Hard drive for around $1.3/GB. Its REALLY REALLY cheap.

Just for comparrison, the same 480x480 clip using an uncompressed avi is exactly 16.1MB/sec of video.

[Selur]

*gig* I'm quite good at math and theoretical computer science stuff (studying cs) ;), but since I didn't use huffyuv for a long time and I couldn'z remember how large a normal movie would get with Huffyuv, I thought it wouldn't hurt to ask,... (I just wanted to know a vage number, which i got,.. so it's something between 40-70% of the uncompressed source,..)

About the space,.. space isn't my problem (500GB+) and speed also goes well on my dual machine,.. I just was courious because I never read/head about someone saving to huffyuv who wasn't capturing,...
(used an older version of huffyuv some years back to capture some vhs tapes ;))

Since I'm still a bit curious about the method, a little question:
Could you give some rough estimation how much (over all; for the whole ripping&encoding process) time this method saves over the 'normal' way?

Cu Selur

Ps.: Sorry, if I offended you in anny may with my statments,.. wasn't my intention. (kind of got this feeling when I read your reply,..)

[poopity poop]

of course not, in fact I see no way you could have offended me ha ha.
first things frist:
<drools over 500GB>
second thing:

Again it depends how much time it saves. It depends on three main things:
Which program you are using to encode (and therefore what format)
and
How many passes you are doing.
How many filters you are doing.

For example, for me I encoded Kenshin using smart smoother warpshap resize, two pass nandub on my 800mhz.
It was cluging along at 2fps for 2 passes which meant around 20 hours/pass.
If I made a huffYUV at 2fps, then do two nandub passes at 15fps. It would take about total of 4/5 the time it would have taken. Which isn't that much. But I knew myself that I was encoding from a single source and didn't process things twice, and made me feel better.

Anyway that was a stupid story.
If you are doing SVCD's at like 4 passes it will save about half the time.
If you are not using any noise reduction filters it may not save any time.
I don't know. I just like doing ti so I know that I'm not processing video twice. I'm just rambling now...

[Selur]

ok,.. thx for the info :)