just 18 points Kwag and we all gonna test your Beer-O-matic machine in puerto (que)rrrrico! :lol:

@Incredible
Have I missed something?

watch his posting-counter ... and you can smell the beer!! :lol:

And here you can see how he already prepared himself for the "gran KVCD reunión"!
http://www.kvcd.net/forum/viewtopic.php?t=9021

Are you reading my mind 8O
I just came from the kitchen with a mug full of my last brew, and I read this post :lol: :rotf:

-kwag

Aha... 17 to go. :wink:

EDIT: Of course! Now I see why Kwag's always talking about beer... :wink: I just thought he was raving alcoholic :!: :D

Maybe... but I'm german and so I smell the beer! :lol:

Do you know "Kölsch" Beer from my city Cologne?? They serve it in long glasses! DELICIOUS!! :wink:

The way this thread is going you should be there in the next hour :D

10 000

quite a milestone :ole:

I think if Kwag will go to brew the beer for ALL of us, this will take time! And he will be occupied by doing that (how many members are we here??) .. so it will take a while until 10000 is reached :lol: :lol:

:rotf: Yes, I believe I do. Oh, yes. Remember, I was in the ARMY in Germany, so I had A LOT of German beer ;)
( Licher, Henninger, Becks, and many brands I can't remember now )

-kwag

That's the only thing I don't like about the micro brewery. I have to wait 4 to 5 days for the fermentation process to finish, then another 4 to 5 days in the frig to stabilize the beer. So on average, I have to wait 8 to 10 days, before I can pour a first glass.
And then it's empty in about a week :mrgreen:

Edit: Went off topic :!: Back to scripting :cool:

-kwag

I just tried this script and it's giving excellent results.
Am a newnie to kvcd's , don't understand anything in the script too.

Just copied the script , removed the crop line as it was giving an error , resized the resolution to 352x480, min/max bitrate is 300:1800 getting a cq of 62.6 and widescreen video.
The source is dvd , runtime 131 mins.

Any thing else i can change to get a better cq , fitting it to one cd kvcd.

My question seems to have gotten lost in the indepth discussions (on scripting and beer 8) ).

Anyway, for those who are testing the new script - are you using "detect scene change" or not?

Thanks,
Icarus


EDIT: Wow! My 100th post. Do I get a free beer for that?

I do not personally. But I just did a video yesterday and no had time to do real comparison for the moment.

Hum. For 100 the prize is a Stawberries/Milk. Do you want it warm or cold ? :-PPP

.......maybe......water!
:lol:

just kiddin Icarus3000,
... number of posts don't show knowledge,
compare mine posts with many others members here....i loose in quality!
:wink:

Don't do it! Why? Very easy:

The scene change detection of TmpgEnc is based on an internal threshold, and that threshold is not known, neither adjustable.

So as we know I frames do need the most of bits and seen in a whole, you could have a little quality rise in scene changes, but on the other hand the overall CQ will drop. And thats the problem on oneCDencodings where we do gain from every bit.
I saw one case where SceneChangedetec. was enabled and in some high action scenes I got crazy arbitary GOP counts:

IBBPIBIBBPBBIBIBIBI ... you see what I mean?

IF you want to do encode using Scene change, then TmpgEnc does got an option where you can enter the "Forced Picture Type settings" which is a bit tricky but you can set there the option to "Control" tmnpgEncs automatic SC findings.
;-)

PS: IMHO a KVCD specs GOP of 24/25 is tight enough to do good encodings without scene change detection ... seen in relation to our very compressing way of KVCD. I did a Template where I did set a GOP to 84 frames in effective! There it gave me a smaller quality increase when doing the SceneChange way BUT using "Forced Picture Type Settings" and its "sensivity" threshold determination option.

BTW: we should do a test and encode a KVCD using a GOP of maybe crazy 250 frames BUT with a "right" setting in SceneChange detection in "forced picture type settings" .... Means only I frames at Scene Changes ..."sounds" cool and logic but could result in scenes using really 250 frames in a GOP .... and that should we find out as this could result in a Mega compression advantage ... but on the other hand in a very risky stream according to SAP GOP sensibilities :lol:

Ill do it tonight! 8)

That won't work Inc. We've already tried long GOPs, and the result is a "Flashing" or "Pulsating" video.
Search the forum on "Flashing" or "Pulsating" and you'll see.
When you have a very long GOP, the video degrades as time goes by, and when a new I frame is inserted, it's like all of a sudden you turned on a light :lol:

-kwag

ok, I made a lot of OneCDencodings using a GOP length of 84 and everything was playing fine and quality-linear. :wink: And you know my taste according to quality

If you like a taste Ill give you the link of a 400kbit AVG sample of mine of IndianaJones1 @480x576.

But back to Hyb.Fupp() ....
Did you check the avsi inside? Its build very logical and if you look closer everything builds on that edge detection ... I have to say that this is a great work of Fupp as you also can set every affecting parameter means total control of everything which does that routine.

I do recognise that the spatial work does its job very heavy if left to the presets defaults ... actors become "shaved" very fast even in non motion areas ... so watch out ... but that should only be a parameter setting job to ged rid of that :lol:

But what I do not understand that much is why he uses that complex diff. resizers? I mean you can just set a clear Bicubic setting as default which does treat the WHOLE image and the edgedetection is only working for spatial/temporal/sharpen/blurring ???

And that is what we should try on MA ... the implementation of masktools and the edgedetection.

EDIT:

Masktools (download is mentioned in here for vers. 1.49) works very nice, just apply for a test at the end of your script BEFORE adding the borders:

If you change MaskedMerge(a,b,c) into MaskedMerge(a,c,c) .... you can see what will be softened. :)

one little question i am getting the same error Kwag was getting about the even numbers and have got a question.

here is my line from moviestacker

Now i know the 11 is the problem but what should i change it to 12 or 10 :oops: o or it doesn't matter

thanx

If it's action scenes, you won't notice it.
That's one of the reasons we dropped the GOP size down to 24(25).
The original KVCD templates had a GOP size of 48, and people were reporting the flashing effect.
It's only visible on long static scenes, like this: http://www.kvcd.net/test.m1v.mpg
That's a GOP of 80.
Look closely, and you'll clearly see when the new "I" frame strikes ;)

-kwag

OK, but we should still UNCHECK detect scene change?

Scene change in the encoder has nothing to do with a script itself (maybe SCD in Temporal Filters), its a method to archive best effective Quality refreshment in the encoding.

BUT read my explanation above why its very risky to set SCD in low avg bitrate encodings like 3 Movies to one DVD-R or 6 vcd movies to one DVD-R. And shurely on OneCDencodings.

But thats all only IMHO :wink: Ind I proofed it only for myself.

What is, at the moment, the better script for use with HybridFupp?

Happy if HybridFuPP can help 8)
 

Hi here

Thanks a lot for spending time testing HybridFuPP. Next version coming very soon...

@kwag
Could you give me a link to a little part of your source, so that I can test further ?

Regards,

FuPP

?

HybFupp can be "seen" almost as an all in one function for DVD sources, as it gots internal resizing and filtering.
And its risky to apply for example another spatial softener etc. before or after the HybridFupp() call.

Re: Happy if HybridFuPP can help 8)
 

Welcome FuPP :D
I'll send you a PM with a link, so you can download a piece of the VOB.

-kwag

incredible,
Thanks for bringing HybridFuPP to our attention. I would never have known about it if you had not posted information about it here. I have been trying HybridFuPP_071_b and it really does give amazing results.

I'm sorry Inc, I supose that avs that calls hybfupp maybe should have
anothers sentences or functions.
Then how can I use hybfupp. I must to use it alone (not to call it from another avs)?

Prodater64
Just put the HybridFupp function in your script and call it like any other function. You may need to read the previous posts here to get it working, but all the information has already been given.

I was read it. But is to hard for me. To much tech for mi english. Sorry.

Prodater64,
Then it is probably best to wait a while. Step by step guides should be available in a few weeks. Right now the function is still being developed.

Greetings Fupp,
I can only congratulate you for your wonderful script.
It's the sharpest possible images while mantaining the compressability near to MA.
Usually MA compresses a tiny bit more than Hybrid but the resulting image is always better than MA.
Oh, boy! I hope you can join the best of MA with Hybrid with Kwag's help.
Cheers

PS: BTW it works ok with mencoder4win32 on a fake avi built with makeAVIS that comes with ffvfw.

Hi rds_correia !

Thanks for your kind words. To be honest, I don't know MA a lot. I previously used Homie FR's FMF. I will of course have a look to MA very soon.

Cheers,

FuPP

Hi FuPP ,I would also like to say Welcome and we are glad to have you here

My fisrt attempt at your script produced good results,takes a little longer but who cares :D

@kwag (and all...)

Good news ! I've had a look to the sample you kindly provided to me and the ghosting effect you've seen will no more be present in the next version (it had already been corrected). actually, there were two phenomenons : first, SCD is not perfect at this time. So, the motion mask can sometimes interfer. BUT, it should normally not be visible.

The problem is that I use Unfilter for moving areas and this one has a bug with luma. As a solution, I changed a little bit (too much :wink: ) luma after unfilter. This has been removed in next version :)

++

FuPP

Hey, thanks a lot !

Cheers,
FuPP

Thanks FuPP :)
Can't wait for your new version :cool:

-kwag

yes, we all (Can't wait) too FuPP :!:
great work.

your presence here is better than news , welcome :!:

:D

Well, I just say as all the other: Welcome FuPP and congratulations on an excellent function/script! Looking forward to the next version. :wink:

Here it is !

Be careful : need latest masktools 1.4.14 and latest avisynth 2.55

link (click right on link and save target) : http://fupp.chez.tiscali.fr/HybridFu...FuPP_084_b.zip

Change log :
- 0.84b:

- It is now possible to create personal processing chains
- advanced debug mode (displayed with colors, modes can be now combined)
- added a strength parameter for dark areas processing
- more speed
- no more limits concerning final size for chroma an dark (called ‘special’ in 0.7x versions) modes
- deringing (optional) is now done after resizing. Mod 16 size limit concerns now only final size.
- be careful : some parameters have been renamed (ex : special -> dark).
- parameters order has changed ! so be careful if you used unnamed parameters in previous releases :
ex hybridFupp(448,320,12,7,10,30,9,32,17,1,5,true,tru e,true,30,0)

Please read joined readme files for a lot more details and... enjoy 8)

FuPP

You got it FuPP :!:
I just scrolled through the section where the artifact was showing, and it's gone bye bye :D

Thanks,
-kwag