SEt's Avisynth 2.5.8 MT compiled for *X86_64*, Latest Build 4/16/2010

[blindbox]

Wow guys, thanks a lot! I managed to find it in megui x64.

P.S.: Just realised that was my first post. Man, I've lurked the forum for a year.

EDIT: Now I face another problem. It says

LoadPlugin: unable to load "<plugin filename here>.dll"

EDIT2: Nevermind, I figured it out, AvsPmod is 32-bit.

[levi]

Use
try { Load_StdCall_Plugin("C:\Program Files (x86)\AviSynth 2.5\plugins64\<plugin filename here>.dll") }
catch(err_msg) { Load_StdCall_Plugin("C:\Program Files (x86)\AviSynth 2.5\plugins\<plugin filename here>") }

It will fall back to 32-bit version if you are using a 32-bit app.

[azulazules]

Hey there...
Does anyone know if dup.dll v2.30 was compiled for x64 2.5.8 avisynth?
Or can someone help me in compiling it?
Source code of dup.dll v2.30 is here: http://neuron2.net/dup/dupnew.html
Thanks alot! :-)

[asarian]

Can someone please mention a frame-accurate source filter for use with SEt's Avisynth 2.5.8 MT compiled for *X86_64*? FFVideoSource yields an (obvious) error that there is no function named; and DirectshowSource craps in all directions it can whiz.

This is meant for use with MCTemporalDenoise, btw.

Much obliged.

[levi]

http://neuron2.net/dgdecnv/dgdecnv.html

64bit version included in the standard zip

[AlekseiV]

FFVideoSource yields an (obvious) error that there is no function named You have installed FFMS2 x64?

[asarian]

You have installed FFMS2 x64?

Thanks. I just did. :) I'm getting an immediate error, though ("x264 has stopped responding") when I run the following script (for 64-bit):



LoadCPlugin("C:\Program Files (x86)\AviSynth 2.5\plugins64\ffms2.dll")

SetMTMode(5,4)

FFVideoSource("f:\jobs\unter.mkv").ConvertToYV12()

SetMTMode(3)

MCTemporalDenoise(settings="medium", stabilize=true)
GradFun2DBmod(thr=1.4,thrC=1.0,mode=0,str=1.2,strC=0.0,temp=50,adapt=64,mask=false,show=false)

SupTitle("f:\jobs\unter_exp.sup", forcedOnly=false, swapCbCr=false, relocate=true, relocOffset="0,0,0,0")

[cezar]

It will be so good, if someone compiles DGAVCdecode for x64 avisynth. It's very nice frame-accurate source filter.
I ask to consider it as request.

[kypec]

DGAVCDecode has been abandoned & its support has ceased. Try DGDecNV (http://neuron2.net/board/viewforum.php?f=8) or DGAVCDecDI (http://neuron2.net/board/viewforum.php?f=6) if you don't have NVidia GPU.

[cweb]

I finally tried the 64 bit yadif, had the same problem you had.

Managed to solve it.

Just specify the path of the 64 bit yadif explicitly:
Loadcplugin("c:\apps\avisynth2\plugins 64\yadif.dll")

[eni9ma]

FFT3DGPU working on Avisynth x64?

I put in the fft3dgpu.hlsl+dll same directory

but unable to load fft3dgpu.dll

What's the problem?


LoadPlugin("C:\Program Files (x86)\AviSynth 2.5\plugins64\FFT3dGPU.dll")
FFT3dGPU(bt=1, sigma=2.0, bw=32, bh=32, ow=16,oh=16)

[easyfab]

Here's no problem and I don't use loadplugin ,
Do you have the correct x64 version (see first page of this thread ) ?

[Chris_Smith]

Hi JoshyD,

Have tested avisynth-4-16--2010 on Win7_x64, and have been using it to encode to x264, using 64-bit DGDecode and ColorMatrix. do you know if there is a version of avisynth64 that will run on 64-bit Server 2003/XP?

[TheRyuu]

Well I managed to 'fix' TemporalSoften although it's not much of a fix. Turns out several of the SSE3 asm functions don't seem to work (I don't exactly understand what's going on). The offending functions are: scenechange_sse3_asm and accumulate_line_mode2_axmm.

Basically I disabled those so now it uses the next best one by default now (ISSE). Not much of a fix but I don't understand it enough to really fix it correctly. Same with pointresize and company, I just don't know enough to fix it.

Also I noticed some rather weird settings in the build options. It kind of looks like a case of someone who doesn't know what they're doing screwing with random options. Since it's built with ICL I also included code which will cause it to not penalize AMD cpu's although there shouldn't be much dispatching with only /arch:SSE2.

Build: avisynth-x64-20110906.7z (http://warpsharp.info/avisynth/avisynth-x64-20110906.7z)
Diff: Index: src/filters/Video/focus.cpp
===================================================================
--- src/filters/Video/focus.cpp (revision 43)
+++ src/filters/Video/focus.cpp (working copy)
@@ -1708,10 +1708,10 @@
if ((!skiprest) && (!planeDisabled[i]))
{
int scenevalues = 0;
- if (env->GetCPUFlags() & CPUF_SSE3)
+ /* if (env->GetCPUFlags() & CPUF_SSE3) //FIXME: BROKEN?
scenevalues = scenechange_sse3_asm(c_plane, planeP[i], h, frames[radius]->GetRowSize(planes[c]), pitch, planePitch[i]);
- else
- scenevalues = scenechange_isse_asm(c_plane, planeP[i], h, frames[radius]->GetRowSize(planes[c]), pitch, planePitch[i]);
+ else */
+ scenevalues = scenechange_isse_asm(c_plane, planeP[i], h, frames[radius]->GetRowSize(planes[c]), pitch, planePitch[i]);

if (scenevalues < scenechange)
{
@@ -1732,10 +1732,10 @@
if ((!skiprest) && (!planeDisabled[i+radius]) ) { // Disable this frame on next plane (so that Y can affect UV)

int scenevalues = 0;
- if (env->GetCPUFlags() & CPUF_SSE3)
+ /* if (env->GetCPUFlags() & CPUF_SSE3) //FIXME: MORE BROKEN SHIT?
scenevalues = scenechange_sse3_asm(c_plane, planeP[i+radius], h, frames[radius]->GetRowSize(planes[c]), pitch, planePitch[i+radius]);
- else
- scenevalues = scenechange_isse_asm(c_plane, planeP[i+radius], h, frames[radius]->GetRowSize(planes[c]), pitch, planePitch[i+radius]);
+ else */
+ scenevalues = scenechange_isse_asm(c_plane, planeP[i+radius], h, frames[radius]->GetRowSize(planes[c]), pitch, planePitch[i+radius]);

if (scenevalues < scenechange) {
planePitch2[d2] = planePitch[i+radius];
@@ -1783,15 +1783,19 @@
int div=divtab[b_div_line[i]];
c_plane[i]=(div*(int)s_accum_line[i]+16384)>>15; //Todo: Attempt asm/mmx mix - maybe faster
}
- } else { //TODO: transfer this loop from C++ to plain ol' asm
- if(env->GetCPUFlags() & CPUF_SSE3) //FIX: does address alignment ever screw us here?
- accumulate_line_mode2_axmm(c_plane, planeP, d, rowsize,i64_thresholds, c_div);
- else if ((env->GetCPUFlags() & CPUF_INTEGER_SSE)) {
- //isse_accumulate_line_mode2(c_plane, planeP, d, rowsize,&i64_thresholds, c_div);
+ } else {
+ /* TODO: transfer this loop from C++ to plain ol' asm
+ if(env->GetCPUFlags() & CPUF_SSE3) FIX: does address alignment ever screw us here? MAYBE SO I DONT KNOW
+ accumulate_line_mode2_axmm(c_plane, planeP, d, rowsize,i64_thresholds, c_div); */
+
+ if ((env->GetCPUFlags() & CPUF_INTEGER_SSE))
+ isse_accumulate_line_mode2(c_plane, planeP, d, rowsize,&i64_thresholds, c_div);
+ else
accumulate_line_mode2_mmx(c_plane, planeP, d, rowsize,i64_thresholds, c_div);
- } else {
- mmx_accumulate_line_mode2(c_plane, planeP, d, rowsize,&i64_thresholds, c_div);
- }
+
+ /* } else {
+ mmx_accumulate_line_mode2(c_plane, planeP, d, rowsize,&i64_thresholds, c_div);
+ } */
}
for (int p=0;p<d;p++)
planeP[p] += planePitch[p];

[jpsdr]

@TheRyuu.
I don't understand where your build comes from ? Have-you been able to get the code from Joshy_D ?

[kemuri-_9]

He probably got it from the googlecode site (http://code.google.com/p/avisynth64/)

[TheRyuu]

@TheRyuu.
I don't understand where your build comes from ? Have-you been able to get the code from Joshy_D ?

The source code is publicly available at the google code site. I built it myself...

[noee]

Not sure how I can help, but I get an immediate crash in avisynth.dll with:

Command Line:
"C:\Users\Mike\Documents\Utils\Conversion Tools\HDConverToX3\applications\x264\x264-64bit_10.exe" --crf 22 --preset medium --aq-mode 2 --tune film --sar 40:33 --output "C:\Temp\job1\hdconverttox_vts_01_1.h264" "C:\Temp\job1\input.avs"

INPUT.AVS:
SetMTmode(5,4)
Mpeg2Source("C:\Temp\job1\movie.d2v")
ChangeFPS(last,last,true)
SetMTMode(2)
tfm(last,d2v="C:\Temp\job1\movie.d2v", pp=0).tdecimate()
Spline64Resize(720,352,0,62,-0,-66)


Tried it without SetMTMode(), same crash.

4.16 build (which I've been using for months), works fine.

[dannyboy48888]

I am using fluxsmooth (ssse3), TIVTC and DGindexNV all in 64 bit land with no issues. my times increased (or decreased....) as follows on a core2 duo 2Ghz p7350:

pass1 avs2avi null output
(704x384 mpeg2 source)
32 bit: 97fps
64 bit: 142fps

pass2, the actual x264 encode
(704x288 resize)
32 bit: 26fps
64 bit: 32fps

and that is all, no complaints. Just wanted to say THANK YOU!!

and of course the scripts....

pass1
LoadPlugin("C:\Users\ChuckD\Desktop\MeGUI_2028_x86\tools\dgindexnv\x64 Binaries\DGDecodeNV.dll")
LoadPlugin("C:\Users\ChuckD\Desktop\avi64\TIVTC.dll")
LoadPlugin("C:\Users\ChuckD\Desktop\avi64\FluxSmoothSSSE3.dll")
DGSource("C:\Users\ChuckD\Desktop\TODO\00000.dgi",fieldop=0)
crop( 4, 64, -4, -64)
TFM(mode=1, output="C:\Users\ChuckD\Desktop\TODO\tfm.txt")
FluxSmoothST(7,7) # Medium Noise
TDecimate(mode=4, output="C:\Users\ChuckD\Desktop\TODO\stats.txt")

pass 2
LoadPlugin("C:\Users\ChuckD\Desktop\MeGUI_2028_x86\tools\dgindexnv\x64 Binaries\DGDecodeNV.dll")
LoadPlugin("C:\Users\ChuckD\Desktop\avi64\TIVTC.dll")
LoadPlugin("C:\Users\ChuckD\Desktop\avi64\FluxSmoothSSSE3.dll")
DGSource("C:\Users\ChuckD\Desktop\TODO\00000.dgi",fieldop=0)
crop( 4, 64, -4, -64)
TFM(mode=1,input="C:\Users\ChuckD\Desktop\TODO\tfm.txt")
fluxSmoothST(7,7) # Medium Noise
TDecimate(mode=5, hybrid=2, dupthresh=1.0, input="C:\Users\ChuckD\Desktop\TODO\stats.txt", tfmin="C:\Users\ChuckD\Desktop\TODO\tfm.txt", mkvout="C:\Users\ChuckD\Desktop\TODO\timecodes.txt")
Spline64Resize(704,288) # Spline64 (Sharp)

[jpsdr]

The source code is publicly available at the google code site. I built it myself...

Interesting and nice thing, but, question : How do you download it ? Is there any other way than getting files one by one with the browser ???

[kemuri-_9]

Interesting and nice thing, but, question : How do you download it ? Is there any other way than getting files one by one with the browser ???

it's a svn repository so you can check it out.

[jpsdr]

Sorry, as i don't know about svn, i don't realy understand your answer.
If TheRyuu has been able to compile it, he has been able to download the whole files/project ?

[kemuri-_9]

Sorry, as i don't know about svn, i don't realy understand your answer.

I interpret that as "I'm too lazy to google it", so even though I'm supplying the link to the wikipedia article (http://en.wikipedia.org/wiki/Apache_Subversion), you probably still won't read it.

[TheRyuu]

Sorry, as i don't know about svn, i don't realy understand your answer.
If TheRyuu has been able to compile it, he has been able to download the whole files/project ?

If you don't understand svn, I'm not exactly sure what you have to offer that would make getting the source files worth while. If you're simply curious I'm sure you can figure out how to checkout a svn repository with the help of google. It even gives you some tips on the source->checkout page of googlecode.

[jpsdr]

Ok, finaly able to get the source got... Thanks.

[WILLIS]

Do the 64bit builds found @ this link http://code.google.com/p/avisynth64/ have a SetMTMode function?

[TheRyuu]

Do the 64bit builds found @ this link http://code.google.com/p/avisynth64/ have a SetMTMode function?

They should, everything should be pretty much the same as the 32bit version, except way more broken. And since we're talking about the 32bit MT version which is already broken, I'm surprised that this shit even works at all.

[aegisofrime]

They should, everything should be pretty much the same as the 32bit version, except way more broken. And since we're talking about the 32bit MT version which is already broken, I'm surprised that this shit even works at all.

You should be. From my experience with QTGMC'ing LOTS OF DVD videos on my i7-2600K, 64-bit is no less stable than 32-bit.

[mbcd]

Does MT work on your i7 ?

I have much problems with MT, its not usable here, avisynth always crashes (directly loaded into x264). Mostly I crashes at 80-90% of the movie, or sometimes at the beginning.

I tried 2.5.8 MT and 2.6.0 in 32 bit (is QTGMC 64bit ?).
Both with the patched plugins for better MT-stability

Now I tried converting without MT and It runs very stable, no crash since the beginning.

[MDCore]

I'm trying to setup a 64bit toolchain with AVISynth 64 and VirtualDub64. I have installed all the relevant programs and after reading through this thread I found testavsload. When I run testavsload64 with the 64bit DLL's (avisynth.dll and devil.dll) in place I get:
unable to load avisynth, error code = 126
If I put in the 32bit DLL's and run testavsload32 then it works.

I have installed the Microsoft C++2005 64bit redistributable. When I use Dependancy Walker on avisynth.dll it complains about MSVCR80.DLL. The 32bit dll doesn't complain about this.

Any ideas?

[TheFluff]

You should be. From my experience with QTGMC'ing LOTS OF DVD videos on my i7-2600K, 64-bit is no less stable than 32-bit.

This, kids, is a good example of the logical fallacy called "argument by anecdotal evidence".

[aegisofrime]

This, kids, is a good example of the logical fallacy called "argument by anecdotal evidence".

I posted that because I was curious as to under what circumstances 64-bit is unstable. I understand that my experience does not cover every single plugin so I cannot say with certainty that 64-bit is as stable as 32-bit. But then, I'm also wondering based on what evidence TheRyuu says that? Has he tested every single plugin? Or is his statement based on looking at the source code and finding out that it is very poorly coded? A short explanation as such is more helpful. :thanks:

[redfordxx]

Hi.
I am now trying working with MeGui64, AviSynth64, x264_64 and 64bit filters.
But I have two problems:many crashes and not many filters.
So I am going back to 32 and see if it is better.

I have question, whether I can have mixed chain...something like Avisynth32+x264_64...or so?
After some browsing I understand, that there is no way to put filter32 in avisynth64, right?

Tnx

[WILLIS]

Can't seem to get SetMTMode to work...it's unrecognized...any ideas?

[LeXXuz]

Anybody got a working sharpening filter for Avisynth64?

I need some working sharpening filter to slightly resharpen the image after noise removal.

I don't like the results from warpsharp or the default sharpening filter from Avisynth.

[Didée]

@LeXXuz: What, how, and why.

[LeXXuz]

I tried warpsharp and it looks okay on anime but removes too much texture detail on normal movies.

[Didée]

Now we've changed from "poking in the blue" to "poking in the green". A major improvement.

[levi]

lsfmod works

[zmaster]

Is there any way to make anti-aliasing?
p/s: «Anti-aliasing x64» of course :)

[tomcat_]

i have installed avisynth the way it was described on the first post however i have noticed that:

1. the time on the analysis it takes quite longer than the 32bit before (for the same movie)

2. After i press save on the avs script creater a video preview window opens while before it wasn't.

am i doing something obviously wrong or other users experienced the same?

[DarkDream]

Windows XP 64-Bit

This is my first 64 bit operating system and I cant get this to install.

It says I need Administrator privilages to install it. I am the damn administrator and the only account on the computer so thats fucked up.

I go to right click it and choose Run as Administrator and that doesnt even show up in windows, its not there to even click like it is on my other computer.

What should I do? i tried editing that out of the batch file where it checks for administrator privilages and it doesnt even work then.

[DarkDream]

NVM i edited the batch file again and got it to work this time without needing admin crap

[DarkDream]

Windows XP 64-Bit

I try to run MeGui in 64 bit mode enabled and thats the error I get when it starts to encode my video. avs4x264mod.exe

I have AVISynth 64-Bit installed

Why is MeGui giving me this error? how would I go about fixing this?

[XMEN3]

Sorry, would be possible medianblur x64 and removegrainHD ? thanks anyway

[Jou]

This is the first time in 14 years I need to post a question here, previously everything I needed to ask has already been answered somewhere else or I had to post in another forum...

Request: Multidecimate for Avisynth in x64 http://neuron2.net/multidecimate/multidecimate.html
I would only need the MultiDecimate.dll in x64, the rest of the filter does not really need to be x64. If possible without too much work: Make it accept RGB directly so ConvertToYUY2() is not needed.

Reason: I record some game videos in 3968x2480 with pseudo-60-fps in fraps. Generates huge amount of duplicate/identical frames in an irregular patten. Multidecimate is the most simple way to get rid of them.

BUT: I cannot encode that resolution with 32 bit virtualdub/avidemux/x264.exe or any other codec I tried. I HAVE to use x64 to just encode the video.
My ugly workaround was: Since the videos were rather short and I had enough space I could save the multidecumated video as uncompressed video, 2 Minutes 40 Seconds made a nearly 400 GB .avi.
But it would be SO much more elegant to have multidecimate for x64 avisynth to spare that workaround.
(Results at http://www.youtube.com/watch?v=-y335WQvGzY and http://www.youtube.com/watch?v=0FhUKndMrAY , including links to the original encodings which require an OC'ed i7 just to play...)

[thewebchat]

This is the first time in 14 years I need to post a question here, previously everything I needed to ask has already been answered somewhere else or I had to post in another forum...

Request: Multidecimate for Avisynth in x64 http://neuron2.net/multidecimate/multidecimate.html
I would only need the MultiDecimate.dll in x64, the rest of the filter does not really need to be x64. If possible without too much work: Make it accept RGB directly so ConvertToYUY2() is not needed.

Reason: I record some game videos in 3968x2480 with pseudo-60-fps in fraps. Generates huge amount of duplicate/identical frames in an irregular patten. Multidecimate is the most simple way to get rid of them.

BUT: I cannot encode that resolution with 32 bit virtualdub/avidemux/x264.exe or any other codec I tried. I HAVE to use x64 to just encode the video.
My ugly workaround was: Since the videos were rather short and I had enough space I could save the multidecumated video as uncompressed video, 2 Minutes 40 Seconds made a nearly 400 GB .avi.
But it would be SO much more elegant to have multidecimate for x64 avisynth to spare that workaround.
(Results at http://www.youtube.com/watch?v=-y335WQvGzY and http://www.youtube.com/watch?v=0FhUKndMrAY , including links to the original encodings which require an OC'ed i7 just to play...)


Multidecimate is a CLOSED SOURCE plugin so nobody but the author can help you.
There is no reason you can't use 32-bit Avisynth to do this. Please read about the standard input/output interface (http://www.microsoft.com/resources/documentation/windows/xp/all/proddocs/en-us/redirection.mspx) available in all Windows versions since DOS.
64-bit Avisynth is a DEAD project so there is no point in posting anything to this thread.

[Jou]

Multidecimate is a CLOSED SOURCE plugin so nobody but the author can help you.
There is no reason you can't use 32-bit Avisynth to do this. Please read about the standard input/output interface (http://www.microsoft.com/resources/documentation/windows/xp/all/proddocs/en-us/redirection.mspx) available in all Windows versions since DOS.
64-bit Avisynth is a DEAD project so there is no point in posting anything to this thread.


Closed source - damn.

I know redirection, I tried the "avs2yuv.exe input.avs - | x264-x64.exe --stdin y4m" way which I used until about ~two years ago. Since then the x264-vfw codec and avidemux x264 implementations were good enough to get rid of that workaround. But it didn't work here. Recent versions of x264 threw me an error, older versions work, but I didn't invest much time since the uncompressed workaround was the lazier way.

[GrofLuigi]

Hi,

Any solution for the green frames yet?

My script consist only of MPEG2Source ("x.d2v"). Is the plugin buggy?

I will try to learn FFMpegSource, but currently I need DGDecode. Is there another version maybe?

And by the way, this (http://forum.doom9.org/showthread.php?p=1499263#post1499263) is vital. It should be put into the installer. DirectShow (read: MPC-HC) doesn't work without it. I banged my head against the wall until I remembered to :search: :).

* Edit: this is all on x64, of course.

* Edit2: both 4/16/2010 and 3/19/2010 releases.

GL

[cweb]

Why don't you use DGINDEX instead? It's more up-to-date.
Hi,

Any solution for the green frames yet?

My script consist only of MPEG2Source ("x.d2v"). Is the plugin buggy?

I will try to learn FFMpegSource, but currently I need DGDecode. Is there another version maybe?

And by the way, this (http://forum.doom9.org/showthread.php?p=1499263#post1499263) is vital. It should be put into the installer. DirectShow (read: MPC-HC) doesn't work without it. I banged my head against the wall until I remembered to :search: :).

* Edit: this is all on x64, of course.

* Edit2: both 4/16/2010 and 3/19/2010 releases.

GL