Support forum for DG Tools

Box of Shame?

DGIndex? Ewww!

Then add MKV, right? Umm, no.

Curly wrote: ↑

Wed Dec 23, 2020 6:27 pm

DGIndex? Ewww!

Then add MKV, right? Umm, no.

Why not

Progress bar of life. Besides, Nattie and me are making the sweet music, know what I'm? How about you, are you busy? Sherman can teach you pointers in a jiffy.

Me, I have a Nvidia card

Good choice. I can't find a 3090 for my life, though. Got any contacts?

Canada Newegg
https://www.newegg.ca/p/pl?N=100007708%20601357248

American Newegg
https://www.newegg.com/p/pl?N=100007709%20601357248

Have you tried Amazon?

Lots OUT OF STOCK but a few are avaiable.
Recommend waiting a while, right now selection is slim and prices are inflated

EDIT
Quick question
For mpeg2 is DGIndex still the go to software or is DGIndexNV just as good?

Hope springs eternal in your breast.

gonca wrote: ↑

Wed Jan 06, 2021 6:46 pm

For mpeg2 is DGIndex still the go to software or is DGIndexNV just as good?

They're equivalent in all important respects.

Ask DG what happened at the Moose Lodge.

DG
What did you and/or Curly do?

What happens at the Moose Lodge stays at the Moose Lodge. Use your imagination.

Word from Sherman is that I already did a big speedup for DGindex in mid 2021. He's going to send us my changes now so we can have it a bit earlier.

What happens if I decide not to do it?

You think you have free will?

What happens if I decide not to do it?

Maybe you should not have said anything

Ha ha. It was too late to not say anything.

But anyway, Sherman just sent me the changes I made for DGIndex in July 2021. I tested them and the time to index MacGyver gave the following results:

DGIndex old: 110 seconds
DG index new: 31 seconds
DGIndexNV: 12 seconds

Wow, another 350% speed improvement. Thank you Rocky!

I'll do some regression testing and then make a release.

I'm curious why DGIndexNV is so much faster. Not complaining, mind you, but it would be great to further improve DGIndex if possible. I think it may not be possible, though, as DGIndex uses a very inefficient parsing architecture with lots of seeking and MPEG-level lookahead, which is very expensive. No Box of Shame for DG guys, that architecture was inherited from DVD2AVI.

Ooh, Levi just came up on SiriusXM. Best male vocalist ever! C'mon, nobody cares about opera, except maybe Mr Peabody.

Now Band of Gold is playing. Could life get better?

Sweet memories. At Haberdashers, we used to play that on the coach to our away rugby matches. Our captain and fullback Liam played it to pump us up for the game. Play it loud! Then shandy in the very muddy hot tub after the game. Who wouldn't want to relive their youth?

Hey Rocky

changes I made for DGIndex in July 2021

Are you changing the timeline, and what will the consequences be?

I don't understand your question. Any paradoxical considerations should be addressed to Albert. Remember, it's all relative.

Here's the faster version of DGIndex. Your test results will be welcome.

http://rationalqm.us/misc/DGIndex_fast.rar

VS 2019 is messing with me. I'm so confused. You tell me, is it working?

It is faster
Did a file compare on the audio and m2v demuxed, along with the d2v index
No differences

Great, thank you.

Rocky wrote: ↑

Mon Dec 21, 2020 9:20 am

The upConv, iCC, and i420 options were removed because these things can be done using script functions.

Would it be possible to add back in the "cpu2" function to the 64-bit DLL? It seems to only exist in the 32-bit DLL (and still works fine)

Having just switched to a 64-bit AviSynth+ install to try to gain a few encoding FPS, I noticed my AVS scripts from the 32-bit/2.0.0.7 version no longer load. I really like the deringing and minutiae deblocking it offered, as I felt it killed less detail than some of the outside deblocking scripts I've seen

Welcome to the forum PoeBear.

The problem is that all of that postprocessing is implemented with inline assembler, which is not supported in 64-bit Visual Studio. I do not have the skills to port the assembler code, and I have never seen any pure C/C++ code for this functionality, going all the way back to the beginning. So sorry, that won't get done, at least by me.