06-17-2004, 12:56 PM
|
Free Member
|
|
Join Date: May 2004
Posts: 84
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
@Dialhot
Not if u have a huge LCD tv... 60"
Wont matter much at all i dont think... just black borders.. which i sed i dont like, but if the pictures big then thts gr8
__________________
im a noob, sorry!
|
Someday, 12:01 PM
|
|
Site Staff / Ad Manager
|
|
Join Date: Dec 2002
Posts: 42
Thanks: ∞
Thanked 42 Times in 42 Posts
|
|
|
06-17-2004, 01:04 PM
|
Free Member
|
|
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by Dialhot
PS: and let image that in 3 years the majority will have a LCD (or plasma) screen, that DON'T have any overscan, and all our beautifull KDVD will be good to be trashed
|
Not exactly
The ones doing KDVDs, the proper and correct way with precise resize/crop/overscan, will probably see slight black bars on the left and right of their Wide Screen TV. The good thing is that most (all?) plasma/digital, etc., wide screen (non-CRT) TV's have a little button called "zoom", (usually a linear zoom, non stepping) which will let you just expand your movie, again flush with all four edges.
Anyone using FACAR's method, will be able to do the same, but with precious movie completely chopped off on the sides, because it was cropped off when encoded :P
And that my friends, is a waste of movie, just like a 16:9 panoramic movie that has been converted to 4:3 Pan&Scan.
-kwag
|
06-17-2004, 01:09 PM
|
Free Member
|
|
Join Date: May 2003
Location: Germany
Posts: 3,189
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
If A Plasma would not be able to compensate such overscan areas, that would mean that 80% of all broadcastings will be shown including these nice errors at the edges. Some do come broadcasted with "black bars", and some ones are almost filling the 768x576 .... but show me one which gots clean edges (and so much DVDs do get cropped in my cases also as often they got some alwful edges).
|
06-17-2004, 01:34 PM
|
Free Member
|
|
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by incredible
If A Plasma would not be able to compensate such overscan areas, that would mean that 80% of all broadcastings will be shown including these nice errors at the edges.
|
Exactly
That's why they have the "linear" zoom" feature, so you can correct the picture
Not like a regular CRT TV, where the overscan is fixed, because of horizontal analog raster, and you can't make adjustments. Well, you can, but being analog in nature, it's prone to temperature, etc., drifts, which produce errors with time.
BTW Inc, here's a nice document about Analog Video Signals that I'm sure you will add to your bookmarks. It's titled: Bandwidth Versus Video Resolution
http://www.maxim-ic.com/appnotes.cfm/appnote_number/750
-kwag
|
06-17-2004, 05:33 PM
|
Invalid Email / Banned / Spammer
|
|
Join Date: Aug 2002
Location: Brasil - MG - third stone from the sun
Posts: 5,570
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by Dialhot
Quote:
Originally Posted by incredible
But thats not such an error
|
Inc, personnaly I don't give a s... about all this. This is Jorel that continue to hurt his head against a wall asking "what's better" without understanding that 2 pixels more or less won't change the face of the world.
As I said, the diff between Moviestaker and Fitcd is 0.3%. I really don't care about that.
Why ? Find me a TV set that you are ABSOLUTLY SURE that it is correctly setted and don't distort the A/R by 0.3% itself
In other words: there is really no need to torture our brain for such little diff as you will always have a weak link in your chain...
But Jorel seems to love brain torture
|
not true Phil, sometimes that little difference give error and the script DON'T open in vdub or any other player....for this reason i posted that the last MS is buggy! you can test all with overscan and you'll get errors too!
if in each detail i got 0.3% after some months with more adictions i will get 10 or maybe 20% of gain!
for me all details are importants. i think that i am the only one in the forum that delay 6 months searching the best adjusts to encode a simple musical.....searching for that details,i found that using the sharpen filter (unfilter,asharp,etc) before the resize give more quality and short final size....and that the image was pulsing and the dark parts was with "dark green"....remember? i "discover" this all cos i'm lucky cos i have big atention with "everything"
give me 0.3% of gain in my encodes every month and i will be very thankful!!!!
the "brain torture" from my point of view is called "searching perfection".
about tvs (trc)
the image is like a "rubber". if the scene is dark,you will get more width(less current ),if is very bright you will get less width(more current),than never you will get "perfect" size in trcs! but all parameters with good adjusts add quality in the result....just my point of view but it's nothing against your point!
|
06-17-2004, 05:46 PM
|
Free Member
|
|
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by jorel
about tvs (trc)
the image is like a "rubber". if the scene is dark,you will get more width(less current ),if is very bright you will get less width(more current),than never you will get "perfect" size in trcs! but all parameters with good adjusts add quality in the result..
|
Isn't it the other way around
The brighter, the wider. The darker, the narrower
Brighter means more "bias" (less difference between plate and grid), conducting more electron flow, and causing wider deflection.
Darker means a more negative bias to the grid, reducing electron flow, and the opposite occurs.
Or did you swap the "more width" with "less width"
-kwag
|
06-17-2004, 07:05 PM
|
Free Member
|
|
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by jorel
not true Phil, sometimes that little difference give error and the script DON'T open in vdub or any other player...
|
That's completly wrong !
The only thing that can make a bad resize not open in vdub is that and the user didn't set correctly Moviestacker (or FitCD) to produce only correct values ! That is NOT the fault of Moviestaker or even fitCD (you can have this problem on both)
Quote:
the script needs a dimension mod.for this reason i posted that the last MS is buggy!
|
NO IT IS NOT You just didn't set the slider that is under the crop area to the correct value that's all. Moviestacker does only what you asked it to do.
If you had set the slider on the correct value, the two softwares would have give you the same resize parameters.
Quote:
you can test all with overscan and you'll get errors too!
|
Sorry, I NEVER have this because all my sliders are on the 16 value
(sometimes I put them on 8 but it's rare).
Quote:
if in each detail i got 0.3% after some months with more adictions i will get 10 or maybe 20% of gain!
|
We are not talking about cumulative gain here but about less distort. You wil never have a gain of 20% in this domain
Quote:
for me all details are importants.
|
But in this case is not a detail, is a nonsence. Just like I said just above : WHAT GIVE YOU THE PROOF THAT YOU TV ISN'T SCREWING THE A/R OF YOUR SOURCES ?
You are the best guy here to tell that is impossible to have a perfect ( and stable in the time) circle drawn on a TV set. Am I wrong ?
So WHAT is the matter of having a PERFECT A/R that you TV will NEVER display ? I call that losing your time...
Quote:
and that the image was pulsing and the dark parts was with "dark green"....remember? i "discover" this all cos i'm lucky cos i have big atention with "everything"
|
Excuse me but this problem of pulsing had NOTHING to do with a 0.3% of distort in the A/R (let me remember you that 0.3% of distort in the A/R on a 28" TV represents a diff in the height of the image of 0.11 centimeters, ONE MILIMETER)
Quote:
the "brain torture" from my point of view is called "searching perfection".
|
Stop watching your video on a NTSC TV then
Quote:
but all parameters with good adjusts add quality in the result....
|
one milimeter on a 28" TV... let put the things in their real dimension
|
06-17-2004, 08:05 PM
|
Invalid Email / Banned / Spammer
|
|
Join Date: Aug 2002
Location: Brasil - MG - third stone from the sun
Posts: 5,570
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Phil,
after you posted:
Quote:
Originally Posted by Dialhot
Quote:
Originally Posted by jorel
not true Phil, sometimes that little difference give error and the script DON'T open in vdub or any other player...
|
That's completly wrong !
The only thing that can make a bad resize not open in vdub is that and the user didn't set correctly Moviestacker (or FitCD) to produce only correct values !
|
i still don't read the rest of your post to do pictures to show you.
the site where i can post the pictures don't accept big sizes and was needed to cut part of the pictures to show the details needed,ok?
fitcd: http://img10.imageshack.us/img10/8690/fitcd.jpg
moviestacker(resize/video stream): http://img10.imageshack.us/img10/9844/ms27.jpg
mpviestacker(avisynth script): http://img10.imageshack.us/img10/7382/ms28.jpg
see that importants details:
no matter if i uncheck the "anamorphic" or "ITU",WHEN i load the source they are checked ALONE!
see if have differences from the scripts and adjusts in each program.
WHAT I DID WRONG TO GET DIFFERENTS VALUES BEFORE
after you see the picture,look here:
Quote:
Originally Posted by Dialhot
That is NOT the fault of Moviestaker or even fitCD (you can have this problem on both)
|
see my old post:
Quote:
Originally Posted by jorel
the scripts from:
--->moviestacker
BilinearResize(688, 350, 6, 0, 708, 480)# moviestacker overscan2
AddBorders(16, 65, 16, 65)
|
http://www.kvcd.net/forum/viewtopic....r=asc&start=32
HOW I GOT THIS VALUES FROM MOVIESTACKER
change everything and you never get this values...this is buggy!
as i wrote before, i get sometimes wrong scripts from the last moviestacker that don't open in "any" place !
if you never get this problems not means that they don't exist!
PLEASE, DON'T CALL ME LIAR, NEVER MORE
here you are not posting about mistakes or problems from programs,
you are posting about MY PERSON! THIS IS INADMISSIBLE
for WHAT KIND OF REASON i need to LIE about a program build by muaddib that is my friend(yes,i send little gifts to his kids and talk in telephone sometimes) to be RIGHT in "something"
IF i posted that i buggy and some scripts don't open THEN this is TRUE cos i'm posting the problems that i get using the program from my FRIEND and if you don't know,i was his BETA TESTER in his second moviestacker version.
muaddib knows that IF i found something wrong in moviestacker i will post about this problem and he TRUST ME!
i never will lie about anything cos i'm not this way and think that i'm posting buggys from the program of my friend...IS UNADMISSIBLE
the remainder of your post i will read in few minutes!
|
06-17-2004, 08:17 PM
|
Invalid Email / Banned / Spammer
|
|
Join Date: Aug 2002
Location: Brasil - MG - third stone from the sun
Posts: 5,570
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by kwag
Quote:
Originally Posted by jorel
about tvs (trc)
the image is like a "rubber". if the scene is dark,you will get more width(less current ),if is very bright you will get less width(more current),than never you will get "perfect" size in trcs! but all parameters with good adjusts add quality in the result..
|
Isn't it the other way around
The brighter, the wider. The darker, the narrower
Brighter means more "bias" (less difference between plate and grid), conducting more electron flow, and causing wider deflection.
Darker means a more negative bias to the grid, reducing electron flow, and the opposite occurs.
Or did you swap the "more width" with "less width"
-kwag
|
i understood what you mean, it's right the aspect that you posted cos you talk about the width of deflection and i was talking about the width of the "image" that we see...with dark scenes you see more image than with very bright scene
...more bright-more deflection-less image(like little zoom),
dark scenes-less deflection but more image..
...you understood what i mean?
or it's worse?
|
06-17-2004, 08:40 PM
|
Invalid Email / Banned / Spammer
|
|
Join Date: Aug 2002
Location: Brasil - MG - third stone from the sun
Posts: 5,570
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
@ Phil
you quoted this phrase from my post? :
"the script needs a dimension mod."
but i don't wrote this. where did you found it?
i forgot to ask:
do you want that i cut part from the vob and send to you
maybe you trust in your own scripts using fitcd and moviestacker.
this was the better idea that my poor brain can think to remove YOUR doubts!
just choose the size!!!!
....and you can choose what i have to do to cut the vobs too, i will follow your way without lie!
ah...you don't understand "a thing" how we see our movies in Brasil.
here is pal-m and not ntsc...we can choose PAL or NTSC but the standard still "M"! you understood what i mean?
i'm still reading your last post!
|
06-17-2004, 09:52 PM
|
Invalid Email / Banned / Spammer
|
|
Join Date: Aug 2002
Location: Brasil - MG - third stone from the sun
Posts: 5,570
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by Dialhot
But in this case is not a detail, is a nonsence. Just like I said just above : WHAT GIVE YOU THE PROOF THAT YOU TV ISN'T SCREWING THE A/R OF YOUR SOURCES ?
You are the best guy here to tell that is impossible to have a perfect (and stable in the time) circle drawn on a TV set. Am I wrong ?
|
yes, you're wrong if the image from dvd source and the encoded from the script show the exact AR....not perfect but source and target are EQUAL this is what i posted about Facar EQUAL.
if the tv have error will be small and the same using all sources!(or the tv have big problems but it's another problem)
the error is equal in the tv, you watch the SAME,
then the error IN THIS CASE is irrelevant
|
06-18-2004, 03:08 AM
|
Free Member
|
|
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by jorel
i still don't read the rest of your post
|
You should have because all the explanation was under that.
Quote:
BilinearResize(688, 350, 6, 0, 708, 480)# moviestacker overscan2
AddBorders(16, 65, 16, 65)
|
350 is NOT A MOD 16 VALUE. Moviestaker gives it to you BECAUSE YOU DIDN'T SET CORRECTLY THE SLIDDERS that are under the crop zone.
So the error is YOURS, not Moviestaker.
If you don't tell to Moviestacker that you want mod16 only value, it gives to you what you have asked !
Quote:
HOW I GOT THIS VALUES FROM MOVIESTACKER
change everything and you never get this values...this is buggy!
|
NO IT IS NOT
Please Jorel, first try to use correctly a tool before to say it is buggy.
To be precise, Moviestacker is not as much accurate in its value than FitCD but it is definitely NOT bugged.
Quote:
if you never get this problems not means that they don't exist!
|
If you can't read my explanations on how to avoid the problem FOR SURE YOU WILL HAVE IT FOREVER.
I never tell you don't have the problem, I never tell you invented the problem, I tell you that you don't do what must be done to NOT have the problem.
And I spent time to post an answer to you that you didn't even read.
Flame me about that !
Quote:
PLEASE, DON'T CALL ME LIAR, NEVER MORE
|
Where did I say that ? Do you see in my line a place where I say "Jorel, stop lying, you can't have these values with movie stacker ?".
I said exacly : " the only way to have bad values...". So in other words, I say here that you CAN have bad values, isn't it ?
The problem is still taht you stop reading when you want and obviously don't see the explanation that are under the problem.
For you question about a bads quoted line : this is surely a problem I had in posting because I started to type a sentence without looking at the screen and it seems the the cursor wasn't in the correct place
So this was a phrase for me
|
06-18-2004, 03:25 AM
|
Free Member
|
|
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by jorel
yes, you're wrong if the image from dvd source and the encoded from the script show the exact AR....not perfect but source and target are EQUAL this is what i posted about Facar EQUAL
|
You do not answer to my question. Forget FACAR, forget DVD, forget your PC.
I'm asking you if, yes or not, it is possible to have on a TV set a perfect circle drawn on the screen, that stays a perfect circle in the time.
This has nothing to do with DVDs, I ask it to the TV technician that you are.
Now, please, answer that question.
Quote:
if the tv have error will be small and the same using all sources!
|
Of course, I agree with that. So why look for the perfect source if this perfection will be screwed by the TV ?
Note: FYI FACAR does nothing else that I said to you I do on all my DVD aand you answered to me that it was'nt acceptable for you.
With FACAR you just CUT parts of the image and don't do ANY resize.
In fact the problem is that you have in mind a idea where all the defaults are cumulative. But NOT AT ALL !
You can have a sourec that is, let set, shrink in the width by 0.3%. And, by chance, a TV set that expand the width to much by... 03% The 2 defaults will give you a "perfect" image.
Now, if your source was "perfect" (0.0% of error). On the same TV set your image will have an error of 0.3%
Morality : you lose a lot of time to have a perfect source where a little "not" perfect can give you in the end a better "perfect" image.
Quote:
the error is equal in the tv, you watch the SAME,
then the error IN THIS CASE is irrelevant
|
So you wish is NOT to have a perfect image BUT to have an image with the SAME DEFAULTS that your DVD have ?
The problem is that all this work with moviestaker is not to have the same image than you have with your DVD, but to have MORE image, like Kwag explained to you. And this is what introduce this little 0.3% error of Moviestacker.
(by the way, we are still talking about 1 milimeter... It seems that you dropped also this part)
EDIT: I forgot this part :
Quote:
do you want that i cut part from the vob and send to you
|
I don't need that. I just need the film area of your picture that Karl and I asked to you in top of the thread but I can't find the aswner.
So if you can just sday :
source is : ... x ...
image area is : ... x ...
target size is : ... x ...
And I will give you the line given by Moviestacker WHEN IT IS SET CORRECTLY and also the EXACT SAME LINE that you received with the explantion on HOW you have obtained it.
(just to prove to you that I also receive bad values from moviestaker if I do not set it correctly)
Note: Some weeks ago we had a thread where I told you taht, as i'm doing myself, you should NOT resize the DVD sources but only do a crop AND addborders (or do a letterbox that gives the same result in a faster way). You answered to me that it was not acceptable for you and wanted to have THE MAXIMUM of the image on your screen.
Today you "discover" FACAR and can you tell me what it is doing ? CUT and CROP with NO RESIZE
I don't bother to explain to you how to do things but please, do not change your mind anytime
|
06-18-2004, 03:49 PM
|
Invalid Email / Banned / Spammer
|
|
Join Date: Aug 2002
Location: Brasil - MG - third stone from the sun
Posts: 5,570
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
answering your question about "perfect circle" scan lines never give perfect circle)
it's impossible, no matter what you do to get it and, IF you got, a simple different adjust in brightness or contrast will "egg" that "perfect circle"!
but this is OT! the topic is the script used to get the same AR just like the source in 4:3 tvs!!!! if the tv have little error, that error will be the same using the dvd or the movie encoded with the script that is the point! for this reason r6d2 build: FACAR - As seen on TV! (this is the title of his script) http://forum.doom9.org/showthread.ph...ighlight=facar
another:
as you wrote:
"350 is NOT A MOD 16 VALUE" and "Sorry, I NEVER have this because all my sliders are on the 16 value" :
did you saw that picture from moviestacker?
http://img10.imageshack.us/img10/9844/ms27.jpg
see the adjusts used PLEASE!
loading the source i get:
BilinearResize(688, 350, 6, 0, 708, 480)# moviestacker overscan2
AddBorders(16, 65, 16, 65)
loading the SAME source later(or in another day,another hour,later):
BilinearResize(688, 352, 8, 0, 704, 480))# moviestacker overscan2
AddBorders(16, 64, 16, 64)
without change anything in MovieStacker LIKE THE PICTURE POSTED!!!
I DON"T CHANGE ANYTHING, MS do the changes ALONE in the final script!
IT'S NOT MY FAULT!!!!!
that adjusts give me wrong scripts SOMETIMES, not all the times, then it's buggy!!!
I DON'T DID ANY DIFFERNTS ADJUSTS IN MS OR FITCD BUT GOT THAT WRONG SCRIPT FROM MS CLEAR
or i'm still UNCLEAR?
|
06-18-2004, 04:51 PM
|
Free Member
|
|
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by jorel
|
He should change the title to:
"FACAR - As seen on regular TV's!, that are not Wide Screen HDTVs"
Or maybe: "Cut & Scan" , so your Wide Screen movie is converted to almost a "Pan & Scan"
-kwag
|
06-18-2004, 05:28 PM
|
Free Member
|
|
Join Date: May 2004
Posts: 84
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
I have a widescreen tv... still has overscan!
but before u say - i dont know about widescreen hdtv's
but i m 100% sure of widescreen tv's they sure do have overscan
ps... i think this topic has really gone overboard lol
i dont even understand half the stuff.. and before u say y i was also "defending" facar - i was just trying to explain wot it does.
my overall word is tht after reading so much... FitCD is great - fair enough, and so is FACAR. MovieStacker to me is just - bLaH - lol.
using anyone one of the methods is fine.
both have bad n good points
and depends really on ur taste.
__________________
im a noob, sorry!
|
06-18-2004, 06:23 PM
|
Invalid Email / Banned / Spammer
|
|
Join Date: Aug 2002
Location: Brasil - MG - third stone from the sun
Posts: 5,570
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
@ bazzy2004 (Kwag and Phil too) and all
some clarifications:
what Kwag posted you can true believe,in what Phil posted,the same!
we are only having discutions cos we all(you too) always search good results...and the most important:
if Phil or Kwag will post bad things about my mother,i will still love this friends,i really like his persons...for this reason i have freedom to post things like: "PHIL,WE NEED TO KNOW WHO IS MORE STUP...YOU OR ME?" (this phrase with "caps lock" is an example) for me,no matter what words they post i always will like this friends and his opinions are full of wise...this is the real reason of that discutions:
first, i don't know everything about "everything" and they are the best teachers that i have BUT i always try to show my point of view cos,soon or later in all problems get more friendship,respect and knowledge.
all references caming from this guys always shine like gold! trust me!
the real problem is that THEY HAVE TO FIND A WAY TO PROOVE IT FOR ME COS I THINK THAT I KNOW "SOMETHING"
this great thread is still running and in the end will get full clarifications...this is the target!
|
06-18-2004, 10:08 PM
|
Invalid Email / Banned / Spammer
|
|
Join Date: Aug 2002
Location: Brasil - MG - third stone from the sun
Posts: 5,570
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by Dialhot
Note: Some weeks ago we had a thread where I told you taht, as i'm doing myself, you should NOT resize the DVD sources but only do a crop AND addborders (or do a letterbox that gives the same result in a faster way). You answered to me that it was not acceptable for you and wanted to have THE MAXIMUM of the image on your screen.
Today you "discover" FACAR and can you tell me what it is doing ? CUT and CROP with NO RESIZE
I don't bother to explain to you how to do things but please, do not change your mind anytime
|
ok, ok Phil!
i'm not changing my mind cos i still don't have a final position about it!
look that scripts from MS and fitcd with ITU and anamorphic source
(if needed, i can post without ITU and anamorphic checked)
MS (blocks tv overscan=0 overlap)
Mpeg2Source("F:\a.d2v")
BilinearResize(720, 368, 8, 0, 704, 480)
AddBorders(0, 56, 0, 56)
LetterBox(0, 0, 0, 0)
fitcd (no resizing)
Mpeg2Source("F:\a.d2v")
Crop(8,0,704,480)
AddBorders(8,0,8,0)
#Trim(0,399 .FadeOut(150)
choose and change what needed and i will do a sample to watch in my tv(4:3)...can you do it please? thanks!
|
06-18-2004, 10:19 PM
|
Invalid Email / Banned / Spammer
|
|
Join Date: Jun 2003
Location: Voyager II
Posts: 31
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Hi guys,
I had not visited this site for a while. I was invited to this thread by jorel at Doom9 forum. I thank him because this is an interesting thread indeed. I’ve seen a lot of strong statements made by several people. Some of them are correct, others are very misinformed, IMHO.
I would rather not get into a one to one discussion to sort out which is the best resizer/cropper, since I think that in general "what's best" is not a valid question and has lots to do with personal taste. Also, because of some previous personal experience at this forum (debating very simple topics with direct questions I’ve made), it has been fruitless.
I try to use reason to accomplish an understanding, but my gut feeling is that some people here do not like that. They just want to find ways to justify things that resist no analysis whatsoever.
But I'd like to clarify some topics to alleviate some “misinformation”.
First, yes, I did check the Jukka Aho’s page Kwag quoted when I was building FACAR on top of Gerti67’s GPL code. That’s how I discovered one of the bugs of FitCD which shh corrected (and muaddib did not).
I even changed some e-mail with Jukka Aho for specific questions about the kind of resolutions some of KVCDers like to use, like 528x480 or 544x480. In fact, he had no clue about those weird resolutions at that time, and he suggested me to do some experimenting. I did, and in fact created my own aspect/resolution/overscan measuring tool, which then I used to build FACAR.
Rest assured that the decisions I made building non-standard resolutions into FACAR are backed up by facts and not by mere intuition.
Second, incredible is incredibly right about some of his statements. He is right when he details the overlay cropped mode used in FACAR. He is not right, though, when he says that it is equivalent to use the Letterbox command. Part of the optimization FACAR does is to reduce the resizing overhead, which you don’t get if use plain Letterbox. (This one was even hard to get by shh, but finally got it.) I also agree with him that FACAR is more suitable for 2.35:1 sources. With full screen sources I prefer to use AutoFitCD.
He is also correct by pointing out that overlay cropping does not help compression as much as full film pixel resizing.
Third, any aspect ratio error below 1% or so is not noticeable by most humans. Some experts say that even up to 2% is acceptable.
Fourth, the Grip suite has the same bug that FitCD 1.1.2 does. Those of you which use 3/4 DVD frame sizes are probably not aware, but your AR is likely to be wrong in most of your encodes. But don’t worry since as stated on my previous point, you won’t be able to notice. (Unless you are not human, of course.)
Fifth, the number of overscan blocks depends on the frame size of the output. There is no magic value for all resolutions. A percentage will do instead (which is what FACAR does).
Sixth, FACAR is not perfect. It still has some bugs. In particular, non standard resolutions with subtitles don’t work in DVD2SVCD. Also, the postresize script won’t take every filter out there in Debug mode. It also lacks some nice to have features, like interlaced resizing. But what it does, “As seen on TV” mode, does it well. So I don’t think I’ll take Kwag’s suggestion on a new name.
Also, since I’m more quality driven instead of space driven, I prefer to create the least resizing artifacts possible. This is achieved better by the overlay cropping approach, since resizing is done in only one dimension (this is the greatest asset of the anamorphic sources technique, BTW).
Seventh, you indeed may have proper AR preview with DVD2SVCD. Just use FACAR and select preview with debug mode set to 3.
Cheers,
r6d2
PS: Please let jorel change his mind whenever he wishes. It does not hurt to have an open mind.
|
06-18-2004, 10:41 PM
|
Free Member
|
|
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by r6d2
But what it does, “As seen on TV” mode, does it well. So I don’t think I’ll take Kwag’s suggestion on a new name.
|
Hi r6d2,
Long time no see
Really, you're making a 16:9 movie almost a Pan&Scan
And I'm not kidding
My question is, why so much crop
I understand perfectly that your resize/crop routine creates: "As seen on TV", but the problem of "As seen on TV", is that much of the pixel area is being lost on the sides, and that's a fact.
So why would you settle for creating a script that does that, instead of creating a script that does "Accurate and Precise" cropping, to maintain the correct source to target ratio as accurately as possible (FitCD already does that), so people don't loose active picture area, and then resize correctly with overscan so the movie reproduction fills the screen with the "true" director's perspective and vision
That's what this is all about Quote:
Also, since I’m more quality driven instead of space driven, I prefer to create the least resizing artifacts possible. This is achieved better by the overlay cropping approach, since resizing is done in only one dimension (this is the greatest asset of the anamorphic sources technique, BTW).
|
This is achieved by correctly calculating the most precise crop, so when the scaling is done, you have the minimum % difference from source to destination.
This is plain math, and there's no other way around it, as described in the documents posted in the links, related to aspec ratios
-kwag
|
All times are GMT -5. The time now is 02:45 AM — vBulletin © Jelsoft Enterprises Ltd
|