Sansgrip Filters: Blockbuster and Company is Killing Me!!!

[deltaboy]

these filters work great and are so fast, i just wish they would work for me! it only allows me to encode about 90% of my video. damn access violation is getting on my nerves. i dont know if its my script or my video or my version of tmpgenc or my versions of the filters. i cant figure it out!

I FEEL LIKE SUCH A NEWB AGAIN...!!!

using:
fluxsmooth + blockbuster + legalclip
video: RGB PC scale
tmpgenc 2.58
blockbuster: variance
...anything else?

everyone can start making fun of me now.

[kwag]

Hi deltaboy,

I have reported to SansGrip the access violation when using high resolution modes and using Blockbuster. Specially the "Noise" method. I had exactly the same errors you do on the KVCDx3 and the KVCDx2 704x480, bombing towards the end of the encode. Don't use Blockbuster on these resolution. You can use LegalClip() and FluxSmooth() as they do work properly.

-kwag[/u]

[deltaboy]

what resolutions wont work? i thought 352 x 480 was quite meager...

[kwag]

Quote:

Originally Posted by deltaboy

what resolutions wont work? i thought 352 x 480 was quite meager...

I've done several movies at 352x240 without problems. I don't remember if I tried 352x480 with Blockbuster. 704x480 definitively blows up with Blockbuster

-kwag

[deltaboy]

k, i think i may have solved it, however, i may be messing it all up! we'll see...

i changed my variance=1 to cache=1024. putting in a value of cache=8 almost immediately causes the violation. bumping the value to a high number will minimize the possibility of a violation, apparently. one drawback is that it appears to slow down the encoding.

i will know if it works when i wake up tomorrow morning.

fingers crossed!!!

[mfb]

i've done lately a couple of movies using the x2-704x576-, 352x576plus-,
x3@480x576 templates with all of the filters (blockbuster noise & sharpen,
fluxsmooth and legalclip)...good quality and no problems

[kwag]

Quote:

Originally Posted by deltaboy

k, i think i may have solved it, however, i may be messing it all up! we'll see...

i changed my variance=1 to cache=1024. putting in a value of cache=8 almost immediately causes the violation. bumping the value to a high number will minimize the possibility of a violation, apparently. one drawback is that it appears to slow down the encoding.

i will know if it works when i wake up tomorrow morning.

fingers crossed!!!

I will try that today. It makes sense, because If I encode from a couple of minutes before the error using a selected range, it encodes all the way to the end. So it seems it could be a memory/caching problem. Will try it

Thanks,
-kwag

[deltaboy]

well... it works, but... bad results!

my encode produced a substantialy larger file size. we are talking about 75MB larger! actually, i got my best results from directly placing the d2v file into tmpgenc. wierd. using the standars temporal smoother was about 5MB larger.

what a burn!!!

[Jellygoose]

Blockbuster works for me everytime I use it...
I only tried it with high resolutions (704x576 and 480x576).
I use TMPGEnc 2.58 Plus...

[kwag]

Quote:

Originally Posted by deltaboy

well... it works, but... bad results!

my encode produced a substantialy larger file size. we are talking about 75MB larger! actually, i got my best results from directly placing the d2v file into tmpgenc. wierd. using the standars temporal smoother was about 5MB larger.

what a burn!!!

Use KVCD Predictor. This way you'll always fit the movie on your CD-R

-kwag

[kwag]

@deltaboy,

Your tip was 100% correct
Setting cache=1024 fixed the problem. I finished encoding a two hour movie targeted for two CD's, and came out great. Previous attempts failed every time. The correct line I uses was this for a 704x480 encode:
Blockbuster( method="noise", detail_min=1, detail_max=10, variance=.5, cache=1024 )

Thanks,
-kwag