Curious about H.263 in ffvfw
 

marcellus stated:
Kwag responded:
marcellus then asked:

marcellus reported some positive effects being produced by using H.263. I don't think that you can deny that it must have some effect one the stream, why else would marcellus report that it did?

I am curious if Kwag (or anyone else) has tested this...

Thank You....

[/quote]

I don't get why people post something like this... :evil:

If you really are curious, why don't you test it yourself then?? :roll:
There's no way that ffvfw will create that H.263...

That's EXACTLY what I thought before I saw you already answered that.

What is the purpose of post like "did someone try ?" "I'm curious about..."
Who do you thing we are ? Your servants doing dirty works for you to enjoy the results ?

The conversation in the forum was left open ended, I am curious if anyone experimented with this setting yet. A simple "no I have not" would of sufficed...The servant thing is taking it WAY to far...

Obviously, marcellus has been using it for quite some time...Either the program simply rejects the H.263 and uses the traditional setting, or making the traditional setting is some how incorporated into H.263...

@poerschr,

If you READ the H.263 specifications, you'll understand WHY an H.263 stream can't play on WMP, WinDVD, PowerDVD, or on your standalone player.
So really, there's no sense for me (or anyone!) to test it.
If marcellus was able to play it, it was because he created a standard MPEG stream. Not a H.263 stream.

Edit: And yes, I did test it just for the hell of it and to satisfy my curiosity, and verified that the produced MPEG stream is identical, viewed on Bitrate Viewer. So even if you set to encode as H.263, you're still encoding MPEG.

-kwag

Thanks kwag...I just wanted to make sure that H.263 was not in some way influencing the mpeg stream.....


I understand now why you are truly the master of this site...I appreciate your direct answers to my questions...

I'm sorry that I react so slowly, but in my rush and lazyness I didn't read this thread at his time.
So:
Thats true, Kwag didn't bother to answer me and I got tired to ask again. Even though he somehow answered in this thread he is doing it in an unsatisfactory way for me.
Well, I never talked about a H.263 stream, but about the H.263 quantization method. I asked Kwag to try it because in my opinion with FFVFW the quality, (at 352x288, PAL, low bittates) is as follows:
Mpeg standard matrix<Mpeg Kwag's matrix<Mpeg H.263 method
Since when the bitrate viewer could say anything about visual quality? And again, I always been sure that you'll obtain a mpeg stream, playable on a dvd-player, but quantized with H.263 method, not a H.263 stream. I don't know what part of this encoding standard could apply to mpeg encoding, but that's about the magic ffvfw does, since it beats at quality any other mpeg2 encoder known to me.
Marcellus

Hi marcellus,

When the one encoded as H.263 and the one encoded as MPEG produce identical foot prints. That's what I verified, and the result of both samples was a mirror image one from the other. There's no difference. File size, average bitrate, GOP pattern. Everything was the same.

-kwag

Ok, to cut it short let's say that H.263 setting has a placebo effect on me :D :D
You still didn't check it visually but let's drop the issue because it starts to be boring

Best regards
marcellus

The quantization type in ffvfw, h.263 vs. mpeg, defines the default quantization matrix.

But is it possible to test the H.263 matrix in MPEG-2?
In MPEG-4 it's really good for anime, avoid blocks on edges, but also cleans the image a bit too much.

I'd like to see how it compares to the Notch matrix if possible :roll:

Bilu

Selecting h.263 does nothing but select the default mpeg4 matrix, this does not mean it cannot be used for mpeg1/2 as well.

I know, but I'd like to see feedback from others.

I'm setting up my first MPEG-2 encoding stuff and I'm on a mencoder phase right now ;)


Bilu

Hi bilu!
Well, if you ask me I'd say that I get better results with H.263. If you ask kwag or jellygoose they will say that with H.263 you will obtain the same result as with standard mpeg matrix. I dropped dead this issue since I'm not 100% sure that what I see isn't a placebo effect, I have to do some more testings. Untill then I have to keep my opinions to myself.
Happy encodings anyway
marcellus

@marcellus

Maybe you ALL are right !!! :lol: :lol: :lol:

From mencoder docs (also libavcodec based):
http://www.kvcd.net/forum/viewtopic.php?p=64535#64535

Maybe H.263 matrix is the default in FFVFW too ;)
Please try forcing MPEG matrix to see if there are differences.

Bilu