What are strongest features of FACAR and AutoFitCD??

[windtrader]

I've used FACAR|FACAR Detect Borders for quite some time now and it's a fine tool. A few days ago, Manolito introduced me to AutoFitCD and I used it a couple of times and did some reading.

It appears that AutoFitCD has more control and flexibility than FACAR but both do more or less the same thing - resize and increase speed of encoding with less source lines to process. AutoFit also seems to easily allow one to alter the horiz and vert pixels to process while FACAR uses fixed formats.

** For those who use both or either one,
** what features/ usage conditions makes
** one better suited than the other?

[manolito]

@ windtrader

Well, this is really r6d2's territory, and I had hoped he would jump to answer your questions. But it seems he has other things to do right now, so I'll do my best to fill in for him.

Both AutoFitCD and FACAR basically serve the same purpose, and that is saving a little time and give you a slightly better encoding quality. They mainly do this by applying block optimized (or even macro block optimized) resizing and by cropping off the TV overscan area thus saving the bitrate for the visible area of the movie.

Before getting any deeper it might be time for a little overall perspective. If your target frame size is SVCD or CVD then you need every little bit of quality improvement you can get, so AutoFitCD or FACAR are extremely useful. If your target is DVD then chances are that you do not need either one. What I am saying is that if you have plenty of available bitrate to begin with, then you probably get a very good quality encode (Q lower than 20) without these tools. Using them anyway of course won't hurt, but you probably will not be able to notice the slight increase in quality.


Now for the differences between the two:

FitCD came long before FACAR. If you start FitCD as a standalone program you will notice that it can do a lot more than what is needed in conjunction with DVD2SVCD. RB's AutoFitCD plugin integrates FitCD seamlessly into D2S and reduces its functionality to exactly those functions that are useful in this context.

Gerti67 wrote his SemiAutoCropAndResize plugin because he was not completely satisfied with the way FitCD handled the resizing. FitCD normally "zooms out" of the movie a little bit so you will see more "movie content" than on the original DVD. This requires resizing the movie by a very small amount which can hurt the quality of the encode (Nyquist effect). Gerti wanted the encode to look exactly like the source ("as seen on TV" = AsoTV) and avoid this minimal resizing. When Gerti67 disappeared (that was some time before I got involved in the forum) r6d2 took over and added a lot of useful features. "SemiAutomatic" became "FullyAutomatic" so the name had to be changed, too, and this is how FACAR was born.


So what are the strong and the weak points of the two? I have been using both of them for quite some time now, and I really need them both.

FACAR:
--------
Ease of use
As seen on TV
Not suited for AVI sources (in most cases)

AutoFitCD:
------------
Flexibility
Well suited for any sources
Takes a little more time to learn
Not "As seen on TV" by default


Conclusion:
For DVD sources I use FACAR all the time. It's just sooooo easy. I especially love the integration of AviSynth filters (I am still a big fan of QMF filtering).

For AVI sources or when I need more control than what FACAR provides I use AutoFitCD. The real nice thing here is that the preview automatically reflects all the changes you make. This makes cropping off station logos a breeze.


One other thing I noticed but I did not test it thoroughly: FACAR gives excellent results with Bilinear resizing. With AutoFitCD Bilinear often looks a little blurred, so I normally use Bicubic resizing with AutoFitCD. Might have somethig to do with this "minimal resizing" I mentioned before.


Alright, that's pretty much all I know about AutoFitCD and FACAR. If r6d2 reads this: Please don't shoot me if I wrote some total nonsense, I'd be glad to be corrected.


Cheers
manolito

[r6d2]

Originally posted by manolito
Well, this is really r6d2's territory, and I had hoped he would jump to answer your questions. But it seems he has other things to do right now, so I'll do my best to fill in for him.Thanks, manolito, pretty good and clear wrap-up.

[windtrader]

@Mano - Thank you very much. I did not want to post a "reminder" so I'm glad you volunteered. Your explanation confirms some of my thinking but also hits a concern I've been pondering.

I only backup to DVD and always use FACAR to fix the small framing errors in the source if they exist. The reason I used AutoFitCD was your suggestion that it was more flexible when I had a DVD that FACAR could not properly recognize the border. As I read and learned more about Avisynth filtering I see that Nyquist effect somehow can degrade the result when just a small amount of resizing is done (something to that effect).

I know many different outcomes can result by using the different filters but using the standard bicubic or bilinear resize filters, should one use FACAR or FitCD to apply small resizing on DVD just to adjust the minor source frame size errors or is there a risk that the result will be degraded, even to a small degree?

If that is the case, then I certainly can live with passing the small framing errors and correct them during playback by zooming in a bit to frame the picture. I don't know what type of resise algorithm is being applied in the software decoder but that's another can of worms I suspect. :-)

Thanks again for the help.

[bobwillis]

Hi,

I previously thought (prior to reading this thread), that when using FACAR to black out the overscan area (say 3% horizontal & vertical), no resizing was done whatsoever. As a result no resizing picture degradation could occur. Was I wrong?

Regards,
Bob

[manolito]

Hi Bob,

no, you were absolutely right. FACAR "just stupidly blanks the overscan area" (quote from Gerti's documentation) without additional resizing.

Cheers
manolito

[bobwillis]

Hi Manolito,

Thanks for the confirmation.

Best Regards,
Bob

[windtrader]

FACAR "just stupidly blanks the overscan area" (quote from Gerti's documentation) without additional resizing. This seems true when applying a mask in the overscan area, as directed by the OverscanH and OverscanV variables. What I find myself doing is the opposite: that is, the framed area as shown by the guide lines(bars)in FACARDetect is smaller than the original and what happens when setting of zero to both OverscanH and OverscanV a resize is performed, even for a few pixels.

This was clipped from the log today.

---AVS Begin---
LoadPlugin("C:\PROGRA~1\DVD2DVD\AVISYN~1.5PL\Mpeg2dec\MPEG2D~1.DLL")
MPEG2Source("G:\DVD2DVD\DVD2AV~1.D2V")
ResizeTo=0 global ResizeMethod=1 DebugMode=0
GammaCorrection=1.0 Hue=0 Saturation=1.0 Brightness=0 Contrast=1.0
DetectedTop=58 DetectedHeight=362
OverscanH=0.00 OverscanV=0.00 BlockOptimization=0 AsoTV=1
global Bicubic_b_Value=0 global Bicubic_c_Value=0.6
DestinationWidth=720
DestinationHeight=480
DestinationTop=0
DestinationBottom=0
LoadPlugin("C:\Program Files\DVD2DVD\Avisynth2.5 Plugins\SimpleResize\SimpleResize.dll")
# Import("C:\Program Files\DVD2DVD\Avisynth2.5 Plugins\FACAR\qmf15b1.avs")
# Import("C:\Program Files\DVD2DVD\Avisynth2.5 Plugins\FACAR\QMF_functions.avs")
# Import("C:\Program Files\DVD2DVD\Avisynth2.5 Plugins\FACAR\PreResize.avs")
Import("C:\Program Files\DVD2DVD\Avisynth2.5 Plugins\FACAR\FACAR.avs")
# Import("C:\Program Files\DVD2DVD\Avisynth2.5 Plugins\FACAR\PostResize.avs")
ConvertToYUY2()
----AVS End----

In this instance, resize does seem to take place, so the original question stands: Does degradation occur due to Nyquist effect or other resizing artifacts and if so, is it enough that I should discontinue the practice?

thx

[r6d2]

Originally posted by windtrader
In this instance, resize does seem to take place, so the original question standsAs far as I can see, resize does not take place here.

You may want to try Debug mode to see what FACAR is doing.