thanks,i was waiting for this!
:D

Let's get ready to rumBLEEEEEEEEEE :lol:
It's playtime :mrgreen:

Thanks SansGrip :wink:
-kwag

heheh enjoy ;).

Works like a charm SansGrip :D
Just a small detail. When you run the "Sample Encoding Helper", and you go to step 2, could you add extensions .m1v and .m2v on the dialog? Because it defaults to .mpg, and the samples are usually video only .m1v of .m2v ( if MPEG-2 ) files.

-kwag

I gathered from the file size prediction thread that one's supposed to encode the sample with a system stream, and hence with a .mpg extension. Wouldn't this more accurate than a .m1v or .m2v without a system stream?

Actually, it should be more accurate using only the video stream, because the audio is already being calculated separately. So the .m1v or .m2v is a video only stream without any other dependencies. And that's the way it's encoded too. At least I encode video stream only ( ES Video Only ) in TMPEG.

-kwag

By "with a system stream" I meant the "System (Video only)" option. Since the final encode will have three streams -- video, audio and system -- and we're factoring out the audio, that leaves the other two streams to account for in the calculations. If you encode the samples as an elementary video stream with no system stream, wouldn't that cause inaccuracies since the final encode will have all three streams in it?

Or am I confused? ;)

Well, at least all the samples that I've done manually with the formula, I always used ( ES Video Only ) stream. The final file size has always been ~2% to ~3% less than predicted.

-kwag

The lack of a system stream might explain it. If your final size with video and audio streams only is, say, 780mb, with the system stream it'll be 780 * (2048 / 2018) = 791mb.

Or maybe I'm barking up the wrong tree. If you think it would be best to add m1v and m2v to the "open file" dialog it'll take me about a minute. My only concern is that if it really is best to encode samples with both a video and system stream, allowing ES files might throw people off...?

Ok, I just ran a test on two prediction samples. One as ES Video only and another as System ( Video Only ).
File sizes:
ES (Video Only): 9,752MB
System (Video Only): 25,705 8O

So obviously, it has to be run as ES stream.

-kwag

Wow, that's a huge difference. I'm going to cancel my current encode and run the same test on my reference source material. If I get that big a difference I'll release 0.2a with .m1v and .m2v in about 20 minutes.

Edit: Just goes to show there's no substitute for testing! Thanks kwag :).

Hey, anytime :D
Maybe TMPEG is screwing up somewhere :roll: . I'm on 2.59 PLUS

-kwag

I guess it's not TMPEG. Just tried it with 2.58 PLUS, and got the same large file size using System(Video Only).

-kwag

2.59 must be acting up, because I get completely expected results from the two encodes.

System (video only): 15.3mb
ES (video only): 15.1mb

In this case, the video and system stream is...

15.1 * (2048 / 2018) = 15.3

...exactly the right size :?: :?:.

This is probably a stupid question, but... are you SURE you're hitting "System (Video only)" and not "System (Video+Audio)"?? 8O

Edit: I'm using 2.58 not-plus...

Yes. 100% sure. Just running it right now to be sure. System(Video only), not the last one(Check button). Same result 8O
So maybe the PLUS is screwed up 8O , but then 2.58 does the same too :!:

-kwag

2.58 "regular" works as expected for me. I'm wondering if this is a PLUS issue? I can't think of anything other than a bug that would account for such a massive difference in file size :?:... Video + system streams should be just over 1.5% larger than video alone 8O.

Downloading TMPEG standard now...

Edit: Same result with the 2.59 free version 8O 8O

-kwag

heheh, and here was me thinking I was just about to go to bed... :mrgreen:

I know I am going to bed :)
It's 4:00AM here 8O .
Good night ( or morning 8O ) everyone :lol:

-kwag

What?? You mean you can SLEEP... at a time like THIS?? :lol:

Just kidding. Have a good one :D.

KVCD Predictor on Windows 98se
 

deleted

Re: KVCD Predictor on Windows 98se
 

You'll need the Microsoft .NET runtime, which can be downloaded from Windows Update.

You'll need the Microsoft .NET runtime, which can be downloa
 

Is this the 20meg plus file they list?

Can I run Predictor on 98se, and if so, what extra files do
 

deleted

Re: Can I run Predictor on 98se, and if so, what extra files
 

That's strange. Did you get it from Windows Update?

.NET
 

deleted

Hi nicksteel,

You can't use TMPEG's wizards with CQ or CQ_VBR encoding modes. Download FitCD, which has the file prediction script to encode your sample, and then use SansGrip's KVCDPredictor program to calculate your sample size.
Get it here: http://www.kvcd.net/forum/viewtopic.php?t=1369

-kwag

If you're encoding with CBR mode then you don't need to use the KVCD Predictor because it's for predicting final file size with VBR mode. You'll have much better results if you don't use the TMPGEnc wizard and leave the rate control set to CQ_VBR mode and use the excellent KVCD Predictor program :wink:

I would put that movie on two cds with the KVCDx3 template and use 128kbps audio or higher. Download a sample from the main page to see if your player supports this resolution (528x480 or 544x480)

Kwag- we simultaneously posted :lol: :lol:

deleted

Hi SansGrip,

SansGrip wrote:
I've tried everything except re-install Win ME OS. :( .NET will not uninstall
using add/remove programs.

KVCD Predictor would be nice, but .NET too troublesome. I've written
a similar file size predictor using "Autoit" with a similar GUI screen as
KVCD Predictor. I input the movie length and the desired audio bitrate
and the program calculates the audio file size and the differance for
1 or 2 CD selected and the differance becomes the video target file size.
I've saved Tmpgenc .tpr files for CQ 20 to 60 test of the avs test script.
My program starts with CQ=40, since my range is 20 to 60 CQ. depending
on the est. file size result from CQ 40. It will use either 30 which is mid-
point between 20 to 40 or 50 which mid-point between 40 to 60. My
second pass refines the CQ range further until the results approximately
gives me the closet target file size for video. Crude, but effective and
after 5 tests it's very accurate. :D Thanks to the discussions on this
forum the ideas really helped to cut down on the numbers of test
passes to reach the approximate file size. :D I'm sure KVCD Predictor
far surpasses my crude program, but It works for me. 8)

-black prince


-black prince

That really is odd. I forgot to try installing .NET on my laptop, and I won't have time to do it tonight, but I shall try to do it tomorrow. Hopefully it'll give me an insight into your problem.

Glad to hear you found a solution regardless :).

Re: .NET
 

When I installed the runtime I did it from Windows Update and there were no problems at all. I've never tried installing it from the Microsoft download page.

Let me know how it goes.

I'm downloading again from Microsoft and will try again.
 

Worked great on second download. Don't know what happened.

Would really appreciate advice on my last post. I assume that I change the settings from FILM to NTSC. And I would really like to know if the 704x480 would be better than the 528x480 for widescreen HDTV playback.

Nick
Houston

@SansGrip,

I can't mirror your ~sansgrip/ directory. Only your ~sansgrip/avisynth directory. Don't know if this is on purpose, or your provider has incorrect attributes for your home directory. Or maybe you have to create a ~sansgrip/extras , etc., and then that sub directory can be accesed for your non avisynth related programs.

-kwag

This is on purpose. There's a whole bunch of stuff in there that you wouldn't want to mirror :).

I was actually going to revamp my page into two, one for Avisynth filters and one for everything else. I'll probably get round to that today :).

@kwag:

Any updates on the video+system/ES video strangeness?

It's a BIG bug on TMPEG. I tried my older version 2.57 PLUS, and it works correctly using Video only and System video only. So TMPEG 2.59 is broken on System (Video only) 8O

-kwag

Ouch!

So... Should sample strips be encoded with a system stream or without? :)

A thought that struck me last night:

If KVCDP only accepted .m1v and .m2v files it would eliminate the possibility that someone might feed it a file with audio too. But I still think we need to take the system stream into account, as it adds ~1.5% to the final result.

So, what if we encode the sample strips as an ES, but I figure the system stream into the final calculations?

Hi Kwag and SansGrip,

Kwag wrote:
I use Tmpgenc Plus 2.59. Does this mean I should go back to 2.57 or
stick with what I have. I'm confused as what to use. :? I am getting
accurate file size prediction with the manual process. What to do,
what to do :?:

-black prince