Avisynth open failure:AVIS...couldn't locate a decompressor for fourcc HFYU

[CZbwoi]

Here I am trying to finally delve into the world of Avisynth and start to dabble with it, started by following this video, but did a step it didn't mention and created the .avs file first off with avisource("test 1b.avi") inside of it in the folder with the video. Followed the rest of the steps, got AvsP and...this error came up. Googling it is only taking me to foreign forums I have to translate and try to read, one of them pointed to install the huffyuv codec but I'm pretty sure I already did that because that's how I captured this .avi, with VirtualDub and huffyuv.



Any insight onto what's going on?

[manono]

What happens if you add:

.ConvertToYV12()

right after (on the same line) as the AVISource line?

Video=AviSource("k:\test 1b.avi").ConvertToYV12()
return video

By the way, you wrote .avs (I boldfaced it for you) instead of .avi for your AviSource line. That will keep it from opening as well. If you really meant to open another AVS file, you use Import and not AviSource.

[CZbwoi]

What happens if you add:

.ConvertToYV12()

right after (on the same line) as the AVISource line?

Video=AviSource("k:\test 1b.avi").ConvertToYV12()
return video

The same thing happens:

By the way, you wrote .avs (I boldfaced it for you) instead of .avi for your AviSource line. That will keep it from opening as well. If you really meant to open another AVS file, you use Import and not AviSource.

Hm, I was just following what the guy was doing in the video I linked to in the first post (and what's seen in the background of the ^picture) and it seemed to work for him the entire time. In the comments he even addressed that and that he also does it with the created .vdr file:

"Why would the frameserver be running and .avi file but your AvsP is using a .avs source file?"

Don't be afraid of that! But its true its a little bit too less explained. You can also use the path to the video.vdr file which you save when you create the frameserver with VirtualDub ( https://youtu.be/WXhLmH_ul94?t=64 ).
The video.avs file I use in the video just includes the video.vdr file as an avisource like this:
#filename:video.avs
AVISource("D:\Video-Aufnahmen\video.vdr")

So in your avisynth script you can directly include the video.vdr file.
video = AviSource("D:\Video-Aufnahmen\video.vdr")
video = ConvertToYV12(video)
video = AssumeTFF(video)
video = ...
...
Hope that helped

[manono]

That script in the video is referencing a VDub filter which, as far as I know, isn't true in your case. That guy's whole thing is way too convoluted, in my opinion.

HFYU is the Huffman codec. Do you have it installed? Can you see it in the list of installed codecs in VDub (Video->Compression)?

What happens with just a simple:

AviSource("k:\test 1b.avi")#.ConvertToYV12() ###Shouldn't really need to convert the colorspace just to open the video.

[CZbwoi]

That script in the video is referencing a VDub filter which, as far as I know, isn't true in your case. That guy's whole thing is way too convoluted, in my opinion.

I would prefer to see a live preview of what is happening in the actual video as I'm adding codes and filters since I'm new to this, and his method with AvsP seems to do that...

HFYU is the Huffman codec. Do you have it installed? Can you see it in the list of installed codecs in VDub (Video->Compression)?

I do not, I thought that was referring to the Huffyuv codec I installed for capture (?).

What happens with just a simple:

AviSource("k:\test 1b.avi")#.ConvertToYV12() ###Shouldn't really need to convert the colorspace just to open the video.

Inserted the whole thing into the first line once, same error happened. Erased everything from the ### and on, same error.

edit: So huffyuv isn't even in video > compression when I'm in capture mode anymore, that's where it always was for me. I'm checking in SysWOW64 and both files are still in there, huffyuv.dll and huffyuv.inf. I literally just captured with it enabled a day ago, what happened?

[manono]

I have no objection to AvsPMod. Lots of people swear by it. The 'convoluted' I was referring to is his scripts - making things unnecessarily complicated, in my opinion. So, reinstall your HuffYUV codec and try again.

[CZbwoi]

I have no objection to AvsPMod. Lots of people swear by it. The 'convoluted' I was referring to is his scripts - making things unnecessarily complicated, in my opinion. So, reinstall your HuffYUV codec and try again.

I seriously must be doing something wrong here. I followed the text file instruction in the download and right clicked > install on huffyuv.inf. That did nothing, it still won't come back into VirtualDub. Then I moved both files and replaced them again in SysWOW64, still nothing. Everything worked a couple days ago when I first installed it by moving both files there.

I have no objection to AvsPMod. Lots of people swear by it. The 'convoluted' I was referring to is his scripts - making things unnecessarily complicated, in my opinion.

How would you do it to make it less convoluted with AvsP? (it's not AvsPmod I'm using, it's just AvsP like he is, unless you think mod would be better or easier)

[jagabo]

Huffyuv was written back in the day when nobody had 64 bit computers. It needs a special installation procedure to work in 64 bit windows. This post is about Lame but the procedure is the same for Huffyuv:

https://forum.videohelp.com/threads/271712-Lame-MP3-Encoder-on-Vista-x64#1723154

[CZbwoi]

What if I downloaded the 64bit version of Huffyuv and installed that the 64bit way in SysWOW64? Because now a 64bit *does* exist, I mean it's right there on the download page, and I followed this video and it worked. Until I tried to use Avisynth and uninstalled it and reinstalled, I think that's when the codec stopped appearing in Vdub.

Now you're telling me to install the (direct link) version into system32? Or the 64bit version into system32 since I run 64bit? And what do I make of the fact that huffuv is already in SysWOW64, do I erase them..?

edit: I seem to may have answered my own problem, it all changed when I was advised to switch to the 32bit version of Vdub because that one has more filters and such available apparently, which is odd because I never thought there'd be a discrepancy that large between a 32bit and 64bit version of something. Therefore the codecs I installed for the 64bit one would have no place working with the original 32bit version. I'll report back if I do the new huffyuv install correctly.

[CZbwoi]

I ran the command line your thread pointed to, computer made a noise like it was doing something for a second or too. Went inside Vdub to check and it's still not there.

Opened up System32 and there's a HUFFYUV.DLL file and the date modified says a few minutes ago, so I'm guessing that's what the command line did, but it still doesn't work.


[Attachment 36615 - Click to enlarge]

Edit: thank you Jesus, I got it. Thanks to edt (no pun intended) and javaman for pointing it out.

Now if we can get back to how to make it less complicated than the guy that taught me in the video for AvsP or AvsPMod, whichever would be better, I'd appreciate it.

[poisondeathray]

Jagabo is referring to 32bit codec on a 64bit OS

If you plan on using 64bit version of huffyuv (or any codec), the entire application chain must be 64bit. This means 64bit vdub, 64bit avisynth , 64bit whatever, or there must be an application "bridge" of some sort to pipe 32bit<=>64bit

Most avisynth filters are still 32bit (although there are more 64bit versions these days), and most avisynth/vdub workflows are still 32bit

[CZbwoi]

Most avisynth filters are still 32bit (although there are more 64bit versions these days)

Do you recommend I go back up to 64 or stay at 32 if I'm just restoring old tapes?

[poisondeathray]

Do you recommend I go back up to 64 or stay at 32 if I'm just restoring old tapes?

Depends on what filters and what you're doing specifically , but 32bit is still the more stable workflow for avisynth

[manono]

Now if we can get back to how to make it less complicated than the guy that taught me in the video for AvsP or AvsPMod, whichever would be better, I'd appreciate it.

Just open your video like so:

AviSource("k:\test 1b.avi")

You won't even need the 'k:\' if the video is in the same place as the script:

AviSource("test 1b.avi")

You'll almost certainly have to convert to YV12 (or YUY2) to use most filters. Since your final colorspace will be YV12 anyway, no problem, unless, perhaps, you plan on using any VDub-specific filters. But they almost always have their AviSynth equivalents.

[CZbwoi]

Now if we can get back to how to make it less complicated than the guy that taught me in the video for AvsP or AvsPMod, whichever would be better, I'd appreciate it.

Just open your video like so:

AviSource("k:\test 1b.avi")

You won't even need the 'k:\' if the video is in the same place as the script:

AviSource("test 1b.avi")

You'll almost certainly have to convert to YV12 (or YUY2) to use most filters. Since your final colorspace will be YV12 anyway, no problem, unless, perhaps, you plan on using any VDub-specific filters. But they almost always have their AviSynth equivalents.

So just to be clear: I have to manually create a text file, change it to .avs, insert the avisource code into it and save it, and rename it exactly as the avi file each time I want to restore a video? OR can AvsP create one for me each time?

You'll almost certainly have to convert to YV12 (or YUY2) to use most filters. Since your final colorspace will be YV12 anyway

How do I determine if that is the case, that the final colorspace should be YV12 and not the other, is that depending on how it's captured, what kind of tape the source is from, etc?

And is it a good idea to always start off my first line like so to get doing that out of the way?

video=AviSource("k:\test 1b.avi")#.ConvertToYV12()

[poisondeathray]

Most final distributions formats will be YV12 (4:2:0) . You don't necessarily always want to convert it to YV12 right away - it depends on what you are doing.

If you've captured from a tape with huffyuv, likely it will be interlaced YUY2 (4:2:2) , then if you are converting to YV12 for filtering, or for the destination format, you should use ConvertToYV12(interlaced=true), otherwise you will get chroma artifacts, very similar to the ones posted in your PP thread

[CZbwoi]

You don't necessarily always want to convert it to YV12 right away - it depends on what you are doing.

So say if you still want to see things like the source colors for manipulating and filtering you would hold it out until the end?

If you've captured from a tape with huffyuv, likely it will be interlaced YUY2 (4:2:2)

The Elgato I got only gets UYVY, but the same thing applies for that I assume? For what it's worth, a USBLive-2 is coming in the mail tomorrow so that might capture at YUY2.

then if you are converting to YV12 for filtering, or for the destination format, you should use ConvertToYV12(interlaced=true)

But that shouldn't be the final line right? After I do that do I end it off with the doube rate deinterlacing QTGMC? So keep it as interlaced while converting it to YV12 but then in the end you deinterlace it with QTGMC as the final step?

[poisondeathray]

So say if you still want to see things like the source colors for manipulating and filtering you would hold it out until the end?

When you "see" something on a monitor for colors it's actually converted to RGB. You're "seeing" the RGB representation of that YUV video

The point I'm trying to make is don't make a habit out of converting right away "just because". Or try to think of what you are doing and why. It really depends on what you are doing specifically. manono is correct - most filters will require YV12, most destination formats will be YV12, but some filters work in YUY2. YUY2 to YV12 is a "downgrade", because you're cutting the chroma resolution in half. Sure, from a tape source it might not have that high chroma resolution to begin with, but it's bad practice to throw away information unless you have to, or hold it off as long as possible. Theoretically, filtering with more information will get you better results than filtering with less information. Also , YV12 can be problematic for interlaced footage in some programs like NLE's, because each chroma sample is split between 2 lines, there are different ways to handle that and there can be many problems. You avoid various interlaced chroma interpretation issues with 4:2:2.

If you've captured from a tape with huffyuv, likely it will be interlaced YUY2 (4:2:2)

The Elgato I got only gets UYVY, but the same thing applies for that I assume? For what it's worth, a USBLive-2 is coming in the mail tomorrow so that might capture at YUY2.

Functionally UYVY and YUY2 are the same in avisynth. They are both YUV 4:2:2 formats. The difference is only in the way they are stored, packed vs. planar and the byte order.

then if you are converting to YV12 for filtering, or for the destination format, you should use ConvertToYV12(interlaced=true)

But that shouldn't be the final line right? After I do that do I end it off with the doube rate deinterlacing QTGMC? So keep it as interlaced while converting it to YV12 but then in the end you deinterlace it with QTGMC as the final step?

There is a QTGMC variant that can use YUY2 input. If you're using the standard QTGMC, then you need to add it before, because it only accepts YV12 input and it's still interlaced before QTGMC, but becomes progressive after. Once it's progressive, you obviously don't need to use interlaced=true switch for any filter

[manono]

And is it a good idea to always start off my first line like so to get doing that out of the way?

video=AviSource("k:\test 1b.avi")#.ConvertToYV12()

I already showed you how I like to do it. The simpler the better, as far as I'm concerned. Especially if you're just starting out. Remember that YouTube link where the first comment was some guy questioning the video creator about his script? That would have been me, and I still don't buy his explanation. He's trying to teach people something and needlessly making it sound like rocket science.

Thanks, pdr, for reminding us both about having to add the "Interlaced=True" for interlaced sources. I had forgotten about that. But then, I rarely work with color sources.

So just to be clear: I have to manually create a text file, change it to .avs, insert the avisource code into it and save it, and rename it exactly as the avi file each time I want to restore a video? OR can AvsP create one for me each time?

I don't use AvsP (Mod) so I can't help there. I have a bunch of template scripts I use and just change the name of the video, mostly, together with parameters of filters, adding some filters and removing others. That is, my scripts are already mostly made. If you're serious about your hobby and intend on becoming a video improvement enthusiast, I expect you'll do the same thing at some point.

I don't know about AvsP (having never used it) but other programs such as MeGUI or XviD4PSP can create scripts for you to use or edit to your liking. XviD4PSP isn't for only XviDs, but all kinds of audio and video.

And I almost always convert to YV12 immediately after capture (if I've capped with HuffYUV) because the first thing I have to do most of the time is to unblend and YV12 is required by the QTGMC I use.