Interesting thread

[DDogg]

http://rilanparty.com/vbb/showthread.php?s=&threadid=7875

I do think folks should change the bicubic value to .60. It is presently set at .75. This thread tends to confirm that.

[ScooterMyth]

Shouldn't that be .5 instead of .6?

[mrbass]

http://www.cs.sfu.ca/~torsten/Teaching/Cmpt878/LectureNotes/PDF/09_interpolation2.pdf
#12 screen shows the image....entitled "Mitchell, Netravali (contd.)"
I thought the above test was done on humans and below is what is mathematically more correct perhaps for animals.

Steady
"Let me just say that B=0, C=0.5 is a special case filter called Catmull-Rom. It is well known by researchers to be the most numerically accurate of the BC splines."

Since we all know our ancestors are from Planet of the Apes I do think 0.6 is a good compromise. Especially for those with hairy backs and those who enjoy banana splits.

Sounds like when/if .avi import is an option BilinearResize would be better to user rather than BicubicResize or perhaps make that an option in the near future.

also note this comment concerning BilinearResize vs. BicubicResize
http://www.math.berkeley.edu/~benrg/avisynth-reference.html#BicubicResize
If you are magnifying your video, you will get much better-looking results with BicubicResize than with BilinearResize. However, if you are shrinking it, you are probably just as well off, or even better off, with BilinearResize. Although VirtualDub's bicubic filter does produce better-looking images than its bilinear filter, this is mainly because the bicubic filter sharpens the image, not because it samples it better. Sharp images are nice to look at--until you try to compress them, at which point they turn nasty on you very quickly. The BicubicResize default doesn't sharpen nearly as much as VirtualDub's bicubic, but it still sharpens more than the bilinear. If you plan to encode your video at a low bitrate, I wouldn't be at all surprised if BilinearResize yields better quality.

[markrb]

I am a little lost here, but I really want to learn, especially where the quality of video is concerned.
In everyday language what does this all mean for the average DVD to SVCD conversion?
Does this increase the video quality or only the ability to compress?
From what I read the lower the number the better it compresses, but wouldn't that comprimise video quality?

I am posting here instead of the original thread because I am not interested in how it affects Divx or avi files only where DVD2SVCD is involved.

I wonder what DVD2SVCD thinks.

Thanks,
Mark

[dvd2svcd]

I don't think. I have no opinion on the issue.

[DDogg]

I have a lot of DIVX background but thought it would be good to move a svcd version of this thread here. Good thread for the advanced forum and an example of the type of discussion and learning opportunity we can try to get going here. These guys are very, very sharp.

What you can do is set the script for editing and then change the bicubic line to just Bilinear(480,480) without the bicubic parameters. I have done this occassionally when I was trying to push the limits of over 60 mins on a 800 cdr.

What we really need, from a testing standpoint, is somebody that owns a killer bigscreen hi def TV, an excellent pair of eyes and the experience to do several encodes of a difficult piece to compare.

Scooter, the good eyes part leaves out you and I :)

[DDogg]

Old thread but interesting

[Vespers]

Hi all, though I don't post that much I do read the forums pretty religiously. I have learned a great deal about encoding in the last few weeks and wanted to add to this thread. First off I have a 53 inch Hitachi HDTV, and have always had problems getting encoded movies to look well on it. Thru much trial and error these have become my standard settings for most encodes(not including anime)
Conversion 16:9(Unless its a really special case)
Encoder:1 Pass VBR, with antinoise filter-2,Linear quantizer,upper field first,zigzag scanning order, & progessive frames checked.
Image quality priority of 1,Q. factor-4.
Bitrate: Max-2595,Min-1000, Max Avg-2500
Frameserver: No editing, Bicubic resize,Avisynth bicubic b value-0
Avisynth c value .70
Temporal Smoother - Checked at normal values
Audio - No downsample, 128 Bitrate, Autodetect Azid Gain-Checked
DVD2AVI - Forcefilm, automatic

My results look extremely good to me, though I know this is subjective at best, I have experimented with many movies action and drama. I tend to keep the sound at 128(my ears aren't that good anyway) so I can squeeze more bits into the encode.

I have noticed that if I go above .70 on the Avisynth c value that you will sometimes get strange lighting effects when there is lightning or electrical effects going on on the screen, ie during The Green Mile when the french guy was electrocuted and the big guy was in his cell feeling it and the lights started popping with sparks. Otherwise most movies seem okay at the standard .75 Avisynth c value, just to be safe DDogs advice is right on the money, .65 should work for most people.

I know 3 or 4 pass runs yield better results but from what I have seen the viewing differnce is nominal but the time differnce is the huge 5 hours vs 14 hours........

Anyway hope this helps someone, keep up the great work DVD2SVCD, DDog and others......