Would installing 32bit only be best without the 64bit OS? :hmm:

VirtualDub = 32-bit
Virtualdub plugins = 32-bit
Avisynth = 32-bit
Avisynth filters = 32-bit
lossless codecs =- 32-bit

Use either 64-bit or 32-bit O.S.

Lagarith is both 32 and 64, and is what allows jumping between 32 and 64 software restoration workflows. It's all 32, or all 64. For example, I've been using QTGMC and KNLmeansCL in 64-bit Avisynth a good bit lately. But almost everything else is 32-bit. So I have to process in stages.

This isn't for capture. Capture is just 32-bit.

^ Not really aimed at sanlyn, just quoting him. He knows this already. :wink2:

Hey sanlyn you were right about installing XP coz it did the trick of fixing the error. Thanks for the help. :D

Btw, how do I load .m2ts files in Avisynth? :)

.m2ts is a container similar to .ts. Open the same way your other .ts samples were decoded, one of two ways:
- FFMS2 (sometimes troublesome)
- DGindex and .d2v project files, audio with NicAudio (more reliable frame index)

See post #28 (http://www.digitalfaq.com/forum/vide...html#post53891)


How do I load clips into AviSynth? http://avisynth.nl/index.php/Importing_media

How do I find the codec and format of video files?
MediaInfoXP (for all versions of Windows): https://www.videohelp.com/software/MediaInfoXP

It's a h264 .m2ts file which was not able to get loaded onto DGindex! :question:

What about the DGAVCDec? :hmm:

For h.264 encodes, yes. DGAVCDec

Hey sanlyn I'm getting invalid arguments to function RemoveSpotsMC2 error message. Please help. I've created the FFTW library 3 as well as moved the necessary dll files to the plugins folder.

Can't answer that one unless I can see the way RemoveDirtMXC2 is written in your script.

It's the same script as earlierly posted.

In lines ,like these:
there are r functions or commands. You need a period to separate each of them:

Maybe you haven't used NeatVideo recently or have a properly created profile. There are not any filters in existence for Virtualdub that can out do NeatVideo. The results from NV are amazing. Clear with no blurring and no softness if done correctly.

NV takes advantage of the CPU and the GPU and processes quite nicely. You shouldnt slam a product that works amazingly. I've used some of the best scripts for Avsynth and NV does a much better job than those as well. The biggest obstacle is getting a precise profile built within NV.

These blanket statements or "mostly blurs, nothing else" are just plain wrong and giving poor advice.

No.

NeatVideo does mostly blur and smear. That's been widely known for at least a decade now, and it's as true in 2018 as it was in 2003. It's the poor man's (lazy man's) restoration tool. People flocked to NeatVideo only because of how NeatImage was (at the time) a near-miraculous tool for photo editing. But the video tool was a flop, nothing like NeatImage. And NeatImage was long ago buried by Photoshop and Lightroom; NeatImage filled a temporary niche, not a long-term one.

A newbie at Avisynth can easily outperform what NeatVideo can do, and those of us with a decent grasp of Avisynth can make NeatVideo look like a tinker toy or cheap Chinese app.

Worse yet, it is literally throwing away $75 when the superior tool is 100% freeware. I have nothing against paying for software, but not when it's inferior to the freeware.

GPU encoding is often fast and rough, not usually quality. So that's actually not a point in its favor.

If you'd like to have a discussion on NeatVideo, perhaps showing some examples of your claims that it has improved, then do so. Perhaps then I'll give it another look. But not just generic claims of "no, you're wrong, it looks amazing". Until then, I stand by my experience, and it's not at all poor advice. I have no problem being proven wrong, but proof is required. I'm all for adopting or re-adopting tools/software if it gives good results. The goal is good video, not ego.

I do remember you from VH days of yore, so I would be interested in seeing/reading what you have to say.

I don't know about all that lordsmurf, the latest version of neatvideo is good imo (v4), i've tested on camcorder footage i was pretty pleased unlike knlmeans in avisynth. One thing i could agree on is that you'd have to fine tune the parameters ( advanced mode)

Then I'll give it another demo when I get some time. :congrats:

KNLmeansCL is pretty decent, under the right circumstances. I've successfully used it twice, when nothing else would work as well. Both times the final products were pretty amazing. poison at VH helped me on one of them, though his script took lots of tweaking and doom9 posting (sometimes arguing, being insulted, not the nicest of places). The other was fully on my own.

Hey sanlyn I tried DGAVC for my .m2ts file and got a error message 'found nalu 13, len 2 undefined'. Please help.

I have no idea what that message means and never saw it before. If DGAVC gives you the choice to ignore the error, proceed anyway. If you are given no choice and DGAVC simply won't output at all, you have an unusual frame organization in your capture. Use FFMS2 with ffAudioSource anf FFVideoSource as in post #28.

Would ignoring the error create any problems? :hmm:

And which one is better DGAVCDec or FFMS2? :question:

DGAVCDec is reported to be more frame-accurate than FFMS2. But there are two considerations: (a) DGAVDec is more exacting and reports even minor errors that don't noticeably affect the results. I get DAGAVCDec errors almost every time on HD PVR recordings from HD TV, because those captures are typically and purposely slightly out of exact spec for BluRay, done because Hauppauge doesn't want to look as if like they're making it easy for you to pirate videos off of cable. I rework them by making new segment identifiers (no re-encoding) with TSMuxer and editing out unwanted sections and resycning audio in TMPGEnc, and then reorganizing and authoring them with MuliAVCHD for BluRay disc.

(b) with FFMS2 you pretty much have to have a disaster on your hands to get fatal errors from FFMS2. Either way, if you want to do some scrolling back and forth in VirtualDub while either utility is in the process of decoding, both will eventually crash. But, then, neither utility is designed for a lot of to and fro scrolling.

It depends on how closely your recording conforms to industry specs. Both utilities will insist on close conformance, and both might give very slightly different results with problem recordings. I'd try DGAVCDec first. If it totally bombs, try FFMS2.

Ok i got it. :)