AviSynth+ thread Vol.2

[tormento]

for x64 see kedautinh12 #642
LOL

I should open my eyes wider when reading.

10x

[kedautinh12]

No wonder that the unofficial 3.6.2 test2 build is not featured, because I found that it is thoroughly broken...

During a few tests with interlaced sample clips this one crashed reliably using the test2 build:
http://www.nepadigital.com/reencode/avidv.avi

First of all the SSRC audio filter seems to be missing. Plus the encode crashes because a VC++ redistributable is missing. The missing file is VCRUNTIME140D.dll, and this redistributable is not included in the very latest Abbodi VC++ AIO package. Reverting to version 3.6.2 test1 fixed the problems.
I check vcredist downloaded from there errors with some softwares only support visual c++ 2015, 2017, i checked and package from there only name visual c++ 2019 after install
https://github.com/abbodi1406/vcredist/releases
When i downloaded from there, visual c++ full name 2015-2019 after install and softwares work correctly and don't crash anymore
https://downloadly.net/2020/13/3230/03/microsoft-visual-c-redistributable/15/?#/3230-microsof-082013094212.html

[kedautinh12]

for x86 see my post in #648

for x64 see kedautinh12 #642

your built and pinterf can't use with megui 32bit

Edit: i know what wrong, just download this file and copy all file from library folder, after paste to windows/system32 if you use win x86 or windows/sysWow64 if you use win x64
Edit2: https://drive.google.com/file/d/13O9HsY1y93DM16Z75ZXU4G63esEHD2OE/view?usp=drivesdk

[kedautinh12]

Everyone can try 3.6.2 test 2b: https://github.com/AviSynth/AviSynthPlus/issues/190#issuecomment-694731296

[pinterf]

Previous not-so-public test2 version contained debug builds of the x86 bundled plugins so people with no VS development environment had bad luck.

[real.finder]

I did update those old posts with this Reply (note the colors)
ok, so I will mention what not has HBD yet

1st is plugins that has VS ports with HBD:-
1 - EEDI3 (https://github.com/pinterf/EEDI3) (no HBD yet (https://github.com/HomeOfVapourSynthEvolution/VapourSynth-EEDI3))
2 - Dither_resize16/fmtconv (no yet) here https://github.com/EleonoreMizo/fmtconv the vs port
3 - SangNom2/SangNomMod (done by Asd (https://github.com/Asd-g/AviSynth-SangNom2))
4 - TTempSmooth (done by Asd (https://github.com/Asd-g/AviSynth-vsTTempSmooth))
5 - EEDI2 (done by Asd (https://github.com/Asd-g/AviSynth-EEDI2))
6 - TCanny (there are TCannymod by chikuzen but I think with no HBD) (done by Asd (https://github.com/Asd-g/AviSynth-vsTCanny))
7 - Yadifmod (there are Yadifmod2 by chikuzen with not completed HBD) (done by Asd (https://github.com/Asd-g/yadifmod2))
8 - DeblockPP7 (done by Asd (https://github.com/Asd-g/AviSynth-vsDeblockPP7))
9 - median (not yet) here https://github.com/dubhater/vapoursynth-median the VS port
10 - tbilateral (done by Asd (https://github.com/Asd-g/AviSynth-vsTBilateral))
11 - tedgemask (done by Asd (https://github.com/Asd-g/TEMmod) but no HBD yet) here https://github.com/dubhater/vapoursynth-tedgemask the vs port
12 - MSmooth (done by Asd (https://github.com/Asd-g/AviSynth-vsMSmooth))
13 - ASharp (no yet) here https://github.com/dubhater/vapoursynth-asharp the VS port
14 - degrainmedian (no yet) here https://github.com/dubhater/vapoursynth-degrainmedian the VS port
15 - TDeint (done by pinterf (https://github.com/pinterf/TIVTC))

plugins that has VS ports but no HBD:-
1 - scxvid (there are no x64 in avs yet)
2 - hqdn3d (there are avs port but with 16bit hack)
3 - ssiq (there are no x64 in avs yet)

and there are some filters seems has no VS ports

1 - VariableBlur (it was planed by tp7 (https://forum.doom9.org/showthread.php?p=1658610&highlight=VariableBlur#post1658610))
2 - frfun7 (useful for Dot Crawl Removal, used in DDComb)
3 - GradFun2db (maybe it can be replaced with f3kdb?)
4 - LGhost (HolyWu Recently did the VS port, avs+ done by Asd (https://github.com/Asd-g/AviSynthPlus-vsLGhost))

so a lot of those are done now by Asd here https://github.com/Asd-g?tab=repositories and there are some that I didn't mention like https://github.com/Asd-g/AviSynth-CAS and https://github.com/Asd-g/AviSynth-FillBorders and https://github.com/Asd-g/AviSynth-vsTMM and https://github.com/Asd-g/AviSynth-BWDIF and https://github.com/Asd-g/AviSynth-JincResize and https://github.com/Asd-g/RawSource_2.6x and https://github.com/Asd-g/MTCombMask and https://github.com/Asd-g/ReduceFlicker and there are many others!

note: the ones that has "vs" in their names because they are not same as old avs one (Missing things from avs old one or different parameters)

[kedautinh12]

I think FFT3dGPU (no HBD yet) and add more plugins work very good but only have in vapoursynth: BM3D (https://github.com/HomeOfVapourSynthEvolution/VapourSynth-BM3D) and CUDA-accelerated implementation of BM3D image denoising method (https://github.com/DawyD/bm3d-gpu), fvsfunc (https://github.com/Irrational-Encoding-Wizardry/fvsfunc) have AutoDeblock function (trying to detect blocked frames and won’t touch non-blocked frames. It’s recommended for MPEG2 input, but also works with h264 input. You can use it on the whole clip)

[real.finder]

I think FFT3dGPU (no HBD yet) and add more plugins work very good but only have in vapoursynth: BM3D (https://github.com/HomeOfVapourSynthEvolution/VapourSynth-BM3D) and CUDA-accelerated implementation of BM3D image denoising method (https://github.com/DawyD/bm3d-gpu), fvsfunc (https://github.com/Irrational-Encoding-Wizardry/fvsfunc) have AutoDeblock function (trying to detect blocked frames and won’t touch non-blocked frames. It’s recommended for MPEG2 input, but also works with h264 input. You can use it on the whole clip)

seems there are AutoDeblock2 for avs but I can't find it, it will be easier for me to update it instead of read python code and back port it

[kedautinh12]

seems there are AutoDeblock2 for avs but I can't find it, it will be easier for me to update it instead of read python code and back port it

I only find out here related avisynh
https://pastebin.com/JqBvNp32

Edit: Script related
vweakdeblock = Deblock_QED().dfttest(sigma=1 ,y=true,u=true,v=true,tbsize=1).smdegrain(blksize=16,tr=2,pel=1,subpixel=3,prefilter=2,thsad=50,thsadc=25,refinemotion=true)
weakdeblock = Deblock_QED().dfttest(sigma=5 ,y=true,u=true,v=true,tbsize=1).smdegrain(blksize=16,tr=2,pel=1,subpixel=3,prefilter=2,thsad=100,thsadc=50,refinemotion=true)
mediumdeblock = Deblock_QED().dfttest(sigma=9 ,y=true,u=true,v=true,tbsize=1).smdegrain(blksize=16,tr=2,pel=1,subpixel=3,prefilter=2,thsad=150,thsadc=75,refinemotion=true)
strongdeblock = Deblock_QED().dfttest(sigma=13,y=true,u=true,v=true,tbsize=1).smdegrain(blksize=16,tr=2,pel=1,subpixel=3,prefilter=2,thsad=200,thsadc=100,refinemotion=true)
vstrongdeblock = Deblock_QED().dfttest(sigma=17,y=true,u=true,v=true,tbsize=1).smdegrain(blksize=16,tr=2,pel=1,subpixel=3,prefilter=2,thsad=200,thsadc=100,refinemotion=true)

[real.finder]

I only find out here related avisynh
https://pastebin.com/JqBvNp32

Edit: Script related
vweakdeblock = Deblock_QED().dfttest(sigma=1 ,y=true,u=true,v=true,tbsize=1).smdegrain(blksize=16,tr=2,pel=1,subpixel=3,prefilter=2,thsad=50,thsadc=25,refinemotion=true)
weakdeblock = Deblock_QED().dfttest(sigma=5 ,y=true,u=true,v=true,tbsize=1).smdegrain(blksize=16,tr=2,pel=1,subpixel=3,prefilter=2,thsad=100,thsadc=50,refinemotion=true)
mediumdeblock = Deblock_QED().dfttest(sigma=9 ,y=true,u=true,v=true,tbsize=1).smdegrain(blksize=16,tr=2,pel=1,subpixel=3,prefilter=2,thsad=150,thsadc=75,refinemotion=true)
strongdeblock = Deblock_QED().dfttest(sigma=13,y=true,u=true,v=true,tbsize=1).smdegrain(blksize=16,tr=2,pel=1,subpixel=3,prefilter=2,thsad=200,thsadc=100,refinemotion=true)
vstrongdeblock = Deblock_QED().dfttest(sigma=17,y=true,u=true,v=true,tbsize=1).smdegrain(blksize=16,tr=2,pel=1,subpixel=3,prefilter=2,thsad=200,thsadc=100,refinemotion=true)

I got AutoDeblock2 from a friend, and did add HBD, you can got it from Thread in my signature

[kedautinh12]

Thanks

[tormento]

I did update those old posts with this
There is this thread (https://forum.doom9.org/showthread.php?p=1655989), why don't update that instead of polluting an AVS+ developement thread?

Please moderator, move all the pertinent messages there.

[real.finder]

There is this thread (https://forum.doom9.org/showthread.php?p=1655989), why don't update that instead of polluting an AVS+ developement thread?

Please moderator, move all the pertinent messages there.

it was start from

To whom redact AviSynth+ Wiki: what about starting to divide plugins into AVS supported interface versions?

It could be easier for programmers to start a revision of the missing plugins.

[tormento]

it was start from
Nonsense. I was asking about Wiki.

[real.finder]

Nonsense. I was asking about Wiki.

if so, then why you didn't say so when I reply using your post https://forum.doom9.org/showthread.php?p=1910354#post1910354 ? Also there is this thread https://forum.doom9.org/showthread.php?t=168012

and this one?

Real men code AVS+ plugins.

So go back to work. :)

[szabi]

Hi

I tried to install AviSynthPlus_3.6.1_20200619.exe but unsuccessful due to Microsoft Defender SmartScreen deny it.
I got the same when attempt to install AviSynthPlus_3.6.1_20200619_vcredist.exe.
Any suggestion would be welcomed.

Regards
szabi

[manolito]

You can disable SmartScreen with this RegEdit script:

DisableSmartScreen.reg
Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SOFTWARE\Policies\Microsoft\Windows\System]
"EnableSmartScreen"=dword:00000000

[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Explorer]
"SmartScreenEnabled"="Off"

[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Internet Explorer\PhishingFilter]
"EnabledV9"=dword:00000000

[wonkey_monkey]

Or just click on "More info" and then "Run anyway."

[szabi]

Or just click on "More info" and then "Run anyway."

It helped, thanks.

[FranceBB]

I wanted to check whether input is 411 or not, but if I try "clp.is411" where "clp" is my source, it says that there's no function called "is411" however there are functions called Is420, Is422, Is444.
What about 411?

EDIT: never mind, it's clp.IsYV411, I just found out.

[StainlessS]

There are a range of depths for 422, 420 and 444, but only YV411 at 8 bit.
Usage of YV411 is not to be 'encouraged', and only 8 bit IsYV411() is implemented.

EDIT: OK, I see above that you got it.

[FranceBB]

There are a range of depths for 422, 420 and 444, but only YV411 at 8 bit.
Usage of YV411 is not to be 'encouraged', and only 8 bit IsYV411() is implemented.

EDIT: OK, I see above that you got it.

Well yeah, it's just to make a script I'm writing bullet-proof.
What about 410, though?
Is410 and IsYV410 don't seem to work...

[real.finder]

is https://publishwith.me/ep/pad/view/ro.rDkwcdWn4k9/latest locked?
edit: I think I know why, I note there were some CCP writing spam text
I was trying to add
SetFilterMTMode("IT", MT_SERIALIZED)

anyway, here my MtModes.avsi https://pastebin.com/tmw7J2mj

[real.finder]

is frame properties only for video? (I think) but what about the audio?

[StainlessS]

is frame properties only for video? (I think) but what about the audio?

I know nowt about this, but how would one attach properties to audio ?
By entire Clip only ???, Sample ??? [lots & lots of properties].

EDIT:

What about 410, though?
Is410 and IsYV410 don't seem to work...
What is 410 ? ( an Avs colorspace ??? )

[FranceBB]

I know nowt about this, but how would one attach properties to audio ?
By entire Clip only ???, Sample ??? [lots & lots of properties].

EDIT:

What is 410 ? ( an Avs colorspace ??? )

Well, it seems so, according to the Wiki at least: http://avisynth.nl/index.php/Avisynthplus_color_formats

Group: YUV410, Pixel Type: YUV9
although I never really heard of such a thing before

[StainlessS]

YUV9, Is read only, ie converted to something else on input [EDIT: by source filter].

I think is 1 chroma sample per 3 luma samples, both horizontally and vertically.
Will maybe need be mod 6 (or maybe mod 12) both H & V.
EDIT: Wrong, See Later. END EDIT:

Even more ghastly than YV411.

EDIT:
You can totally ignore YUV9, only of interest to source filter guys.

[wonkey_monkey]

YUV9, Is read only, ie converted to something else on input.

I think is 1 chroma sample per 3 luma samples, both horizontally and vertically.
Will maybe need be mod 6 (or maybe mod 12) both H & V.

Even more ghastly than YV411.

Isn't it 1 chroma sample per 8 (4x2) luma samples? That's what the "YUV410" notation suggests. 8+1 = 9, hence YUV9.

There are some 3:1 codecs out there as well, I think, but only in professional equipment as I recall.

[StainlessS]

Maybe you're right there,
all I know for sure is that its a bloody awful colorspace.

Thanks for keeping an eye on me Wankey_Monkey :)

[qyot27]

2.6 had a few stubbed color formats that were clearly on the roadmap to be used, but never were (this is the section (https://github.com/jeeb/avisynth/blob/master/src/core/avisynth.h#L420) of 2.6's source of avisynth.h). Some of them were implemented under a different, more standardized naming scheme in Plus (some high bit depth YUV formats were slated, as there were tags like 'YV48', 'YV96', etc.).

YUV9 is not one of these stubs, but I don't know any core filters that use it. It was added at the same time as the stubs, way back in the early CVS commits for 2.6.

[StainlessS]

Found this

https://www.fourcc.org/pixel-format/yuv-yuv9/


YUV9 yuv pixel format

Intel's web site states that YUV9 is "the color encoding scheme used in Indeo video technology. The YUV9 format stores information in 4x4 pixel blocks. Sixteen bytes of luminance are stored for every 1 byte of chrominance. For example, a 640x480 image will have 307,200 bytes of luminance and 19,200 bytes of chrominance." This sounds exactly the same as YVU9 to me. Anyone know if there is any difference?

Based on: YVU9

EDIT: [EDIT: Wonkey says this is in error]
And this:- https://www.linuxtv.org/downloads/v4l-dvb-apis-old/re31.html

V4L2_PIX_FMT_YVU410, V4L2_PIX_FMT_YUV410 —
Planar formats with ¼ horizontal and vertical chroma resolution, also known as YUV 4:1:0

[wonkey_monkey]

Found this

EDIT:
And this:- https://www.linuxtv.org/downloads/v4l-dvb-apis-old/re31.html

That page is in error. 4:1:0 is 1/4 horizontal, 1/2 vertical.

You can't express 1/4 horizontal 1/4 vertical in that format.

[StainlessS]

Thanks Wonkey,
I guess that the main thing to remember is that you cant use em [ie edit directly in] Avs, although that may not apply to VapourSynth.
(Myrsloik has limited tolerance, so expect not editable in VS either)

[manolito]

After doing more than 30 test conversions using these two AVS+ versions I think I am entitled to a conclusion:

I only use AVS+ 32-bit, my CPU is an Intel Core i5, I have 8 GB of RAM, and I am under Win7-64bit. I use Prefetch=4 in my AVS scripts.

The difference between these two versions is that AviSynth+ 3.6.2-test2b crashes in more than 3 out of 4 instances while AviSynth+ 3.6.2-test1 runs stable in all instances.

I did not find any hints in the version history about changes in the MT modes. The older AviSynth+ 3.6.2-test1 version runs a tiny bit slower than the newer version (about 0.6 FPS slower), but this is no justification for the crashes.


So if this is a regression (maybe only for 32-bit), please fix it before issuing a stable new build...


Cheers
manolito

[real.finder]

I am using test 2 (but with x64) and didn't find any crash, can you post a simple script using ColorBars that show the crash?

[pinterf]

New build, Avisynth+ 3.6.2 test3
https://drive.google.com/uc?export=download&id=1Ow2g4H16B3PvXWiD9VGXkJSQk7mBaxb5

Fixing a missing rounding in 8-16 bits GeneralConvolution, now it matches with mt_convolution, and the (8 bit converted) 32bit float version of GeneralConvolution

[FranceBB]

Hi Ferenc.
First of all, thank you.
Second of all, I noticed a slightly different behavior in Avisynth between bit depths.

If I try to apply Dithering from an high bit depth source to a lower bit depth source like this, it works:


ColorBars(848, 480, pixel_type="YV12")
ConvertBits(10)
ConvertBits(bits=8, dither=1)


and that's fine.

If I try to apply dithering to 8bit starting from an 8bit source, it clearly outputs an error:


ColorBars(848, 480, pixel_type="YV12")
ConvertBits(bits=8, dither=1)


Floyd-S dithering is allowed only for 10-16 bits sources.
And that's fine, it makes sense as we're getting an 8bit input and we're getting an 8bit output so it doesn't make sense to apply dithering and Avisynth returns the proper error.

The behavior changes, though, if I'm trying to do the same thing, like 10bit sources to 10bit sources, 12bit sources to 12bit sources etc:


ColorBars(848, 480, pixel_type="YV12")
ConvertBits(10)
ConvertBits(bits=10, dither=1)


This one, although it's like the 8bit example, it doesn't output any error and just passes through the input.
After all, we're getting a 10bit input and we're trying to get a 10bit output, so it doesn't make sense to apply dithering.

Same goes for 12bit, 14bit, 16bit:

ColorBars(848, 480, pixel_type="YV12")
ConvertBits(12)
ConvertBits(bits=12, dither=1)

ColorBars(848, 480, pixel_type="YV12")
ConvertBits(14)
ConvertBits(bits=14, dither=1)

ColorBars(848, 480, pixel_type="YV12")
ConvertBits(16)
ConvertBits(bits=16, dither=1)


So, all in all, the actual stage of things is:

8bit to 8bit -> Error
10bit to 10bit -> passthrough
12bit to 12bit -> passthrough
14bit to 14bit -> passthrough
16bit to 16bit -> passthrough

16bit to 8bit -> Dithering
16bit to 10bit -> DIthering
16bit to 12bit -> Dithering
16bit to 14bit -> DIthering

14bit to 8bit -> Dithering
14bit to 10bit -> Dithering
14bit to 12bit -> Dithering

12bit to 8bit -> Dithering
12bit to 10bit -> Dithering

10bit to 8bit -> Dithering


So... shouldn't it be either passthrough or error for all the BitsPerComponent==Bits?


The thing can be worked around with something like:


Function ConvertBits_ErrorCatch(clip clp, int "mybits", int "mydither") {

Bpc = clp.BitsPerComponent

Bpc >= 10 ? ConvertBits(clp, bits=mybits, dither=mydither) : ConvertBits(clp, bits=mybits)

}


which passes through the 8bit source if 8bit dithering is specified, just like 10bit to 10bit, 12bit to 12bit, 14bit to 14bit and 16bit to 16bit.

Should we fix the default behavior in the Avisynth code itself and make it the same for all of them?

[wonkey_monkey]

I'd vote for getting rid of the 8-bit to 8-bit error. Specifying dithering will have no effect, but it's not intrinsically contradictory to the filter's purpose to request it.

which passes through the 8bit source if 8bit dithering is specified

Are you sure that's right? Looks to me like it passes through if the source is 8 bit, and converts to 8 bit, even if mybits = 10/12/14/16.

[FranceBB]

I'd vote for getting rid of the 8-bit to 8-bit error. Specifying dithering will have no effect, but it's not intrinsically contradictory to the filter's purpose to request it.

+1 for getting rid of the error.





Are you sure that's right? Looks to me like it passes through if the source is 8 bit, and converts to 8 bit, even if mybits = 10/12/14/16.

Oh crap, you're right, I only had going down in mind, but not going up. Fixed with bits=mybits. ;)

[rack04]

Hoping to get some help to understand why after "upgrading" to Windows 10 I'm getting the following error with Avisynth+ 3.6.1 release.

AVI Import Filter error: (Unknown) (80040154)

The script I'm using is version() and I'm using VirtualDub64.

Here is what I get using avsmeter:

C:\Program Files (x86)\AVSMeter>avsmeter -avsinfo -log

AVSMeter 3.0.6.0 (x86), (c) Groucho2004, 2012-2020

Error: AVSMeter (x86) cannot load a 64 Bit avisynth.dll.

Reverting back to AviSynthPlus-MT-r2728 everything works.

[videoh]

If you are wanting to run all 64-bit then you need Avsmeter64 and the 64-bit Avisynth.dll in system32.

If you are wanting to run all 32-bit then you need Avsmeter and the 32-bit Avisynth.dll in SysWOW64.

You may also have to check your registry paths (not sure how you installed things), as well as verify that no stray Avisynth.dll is laying about getting picked up.

I'm just a noob with this Avisynth stuff but the guys here are aces, and they will probably get you going in no time.

Avsmeter(64) is a great app. Avisynth+ too. Bravo!

[rack04]

If you are wanting to run all 64-bit then you need Avsmeter64 and the 64-bit Avisynth.dll in system32.

If you are wanting to run all 32-bit then you need Avsmeter and the 32-bit Avisynth.dll in SysWOW64.

You may also have to check your registry paths, as well as verify that no stray Avisynth.dll is laying about getting picked up.

I'm just a noob with this Avisynth stuff but the guys here are aces, and they will probably get you going in no time.

Avsmeter(64) is a great app. Avisynth+ too. Bravo!

Thanks for the tips. Apparently I switched the 32-bit and 64-bit versions between system32 and SysWOW64 folders. Seems to work now.

[videoh]

It's not unusual to make that mistake. Because the names make so much sense. Not!

Good to hear you are rocking and rolling.

[zambelli]

I'd vote for getting rid of the 8-bit to 8-bit error. Specifying dithering will have no effect, but it's not intrinsically contradictory to the filter's purpose to request it.

I second this request. Throwing an error on a no-op parameter seems to run counter to how all other internal AviSynth functions are implemented.

[pinterf]

I agree that it should work in a consistent way across all type of inputs. Probably I forgot to throw an error for N to N bits dithering which was my intention.
And yes, it makes sense to ignore the dithering parameter in this case.

[pinterf]

There is a huge delay in the online documentation I wanted to do before any further programming.

Can someone help in making sections in Avisynth wiki for all these new 3.6 related functions and features? Or just proposing where to put them?
Let's discuss it.

- Language extensions (such as "function objects")
- Runtime function parameter extensions (ConditionalSelect, ConditionalFilter, ScriptClip, etc.. to wotk with function objects and a new "local" parameter)
- New helper functions - only some of them documented such as http://avisynth.nl/index.php/Internal_functions#SetMaxCPU
e.g. SetCacheMode, Prefetch with new parameter, SetGraphAnalysis, DumpFilterGraph
- Syntax for string constants that allow "escaped" characters (e.g. e"Hello \n")
- the new Format function
- sections for frame properties (concept, helper script functions such as propSet and propGet family, propShow, etc.
- An extension to the SDK section with the new Avisynth+ and V8 interface.

All these are documented only the readme.txt or readme_history.txt (https://github.com/AviSynth/AviSynthPlus/blob/master/distrib/Readme/readme_history.txt) (latter is more detailed).
(Usually I put there in the files-only distribution you can find them there as well)

I would be happy even with empty placeholder sections to fill in later as we have time for that work.

[gispos]

AviSynth_3.6.2_20200831_test2 32bit
The DirectShowSource.dll from the archive cannot be loaded.

There are no problems with the Dll from Avisynth-3.6.2_20200624_test1

And with ImageSeq.dll there is also an error message
I haven't tested the other dll's.

VirtualDub2 says the VCRUNTIME140D.dll is missing. But VCRuntime 2015 is installed!
Have the default DLLs been changed in the 32bit version?

[StainlessS]

AviSynth_3.6.2_20200831_test2 32bit
The DirectShowSource.dll from the archive cannot be loaded.

There are no problems with the Dll from Avisynth-3.6.2_20200624_test1

And with ImageSeq.dll there is also an error message
I haven't tested the other dll's.

VirtualDub2 says the VCRUNTIME140D.dll is missing. But VCRuntime 2015 is installed!
Have the default DLLs been changed in the 32bit version?

VCRUNTIME140D.dll is debug version runtime.
Supplied dll's are mistakenly debug versions.

[gispos]

OK. But that doesn't change anything. DLL error.

Even worse. The 32bit version 'test2' is incompatible with my 32bit applications that use Avisynth.

There were never any problems with any of the previous versions.

Edit: Just swapped the avisynth.dll to 'test1' and everything works again

[manolito]

Version 3.6.2 test2 was a not so public debug version, don't use it. In the first post of this thread you can find links to test 2b or better yet to test3 where your issues are solved.