Tool to change the Display aspect-ratio?

[Froddy1]

I'm only able to convert HuvYUFF in YUY2 or YV12 colorspace.

That's because you're using ffdshow. The HuffYUV and Lagarith codecs include RGB modes.

Okay, good to know.

BTW, I'm just curious: Which colorspace does PAL-DV use? RGB or YUV?

After you've read my difficulties about importing different videos codecs in the same project: How would you handle this matter? Would you convert all source videos to HuffYUV and use colorspace YV12?

Thanks
Froddy!

[JohnnyMalaria]

DV uses a 4:2:0 sampling (YV12)

[jagabo]

Almost all lossy video codecs uses YUV internally, and most often 4:2:0 subsampling. PAL DV uses 4:2:0 subsampling internally, very similar to YV12. But what you really want to know is what the DV decoder outputs. Panasonic DV Codec always outputs RGB. Cedocida can be configured to output YUY2, YV12, RGB24 and/or RGB32. If you enable YUY2 and YV12 it will negotiate which to use with the editor calling it. You will usually get YUY2 from NTSC, YV12 from PAL.

I would either pick YUY2 or YV12 and use that for all your losslessy compressed intermediates. My preference is for YUY2 because some programs don't handle YV12 properly (they don't differentiate between interlaced and non-interlaced YV12).

For example, set Cedocida to output only YUY2 (VirtualDub may not display this correctly, but it doesn't matter). Start VirtualDub and make sure its internal DV decoder is disabled (Options -> Preferences -> Avi -> (disable) Prefer internal video decoders... Open your DV AVI. Set VirtualDub to Video -> Fast Recompress. Set the video encoder to HuffYUV or Lagarith. Set the codec's compression mode to YUY2. Save as AVI.

[Froddy1]

Hello! I wanna smoothly close this thread, just a final question:

Almost all lossy video codecs uses YUV internally, and most often 4:2:0 subsampling. PAL DV uses 4:2:0 subsampling internally, very similar to YV12. But what you really want to know is what the DV decoder outputs. Panasonic DV Codec always outputs RGB. Cedocida can be configured to output YUY2, YV12, RGB24 and/or RGB32. If you enable YUY2 and YV12 it will negotiate which to use with the editor calling it. You will usually get YUY2 from NTSC, YV12 from PAL.

Very interesting. I recently installed the Panasonic DV Codec. So it's better to throw it away and install Cedocida instead? Good to know.

I would either pick YUY2 or YV12 and use that for all your losslessy compressed intermediates. My preference is for YUY2 because some programs don't handle YV12 properly (they don't differentiate between interlaced and non-interlaced YV12).

Okay, so I will also set colorspace to YUY2.

For example, set Cedocida to output only YUY2 (VirtualDub may not display this correctly, but it doesn't matter). Start VirtualDub and make sure its internal DV decoder is disabled (Options -> Preferences -> Avi -> (disable) Prefer internal video decoders... Open your DV AVI. Set VirtualDub to Video -> Fast Recompress. Set the video encoder to HuffYUV or Lagarith. Set the codec's compression mode to YUY2. Save as AVI.

Thanks for the guide. I already printed it out to remember for further projects.

Just two tiny questions:

1) Do you know a tool similar to GSpot where you can read out in which colorspace a video is actually saved? (It's only for understanding proposes.)

2) As I use VirtualDUB frequently: What is the difference between "Fast Recompress" and "Full Processing" Mode? For creating a Divx from a mpeg2 for example I only useed "Full Processing". Therefore my question.

Thank you cordially for your help and for your patience!
Froddy.

[jagabo]

1) Do you know a tool similar to GSpot where you can read out in which colorspace a video is actually saved? (It's only for understanding proposes.)

I don't know of any program that tells you what colorspace and subsampling is used internally by the codec. Again, what you care most about is what the decoder outputs. For some files VirtualDub's File -> File Information option will tell you the colorspace from the decoder. AviSynth's Info() function will do that for any source.

2) As I use VirtualDUB frequently: What is the difference between "Fast Recompress" and "Full Processing" Mode? For creating a Divx from a mpeg2 for example I only useed "Full Processing". Therefore my question.

Full Processing mode converts the video to RGB and enables filtering. RGB data is passed to the output codec. Normal Recompress converts to RGB and passes RGB to the output codec but does not enable filtering. Fast Recompress leaves the video in the colorspace it receives from the decoder and passes it to the encoder (no filtering). Direct Stream Copy copies complete compressed frames from the source to the output file (no decompression, no RGB conversion, no filtering, no quality loss, only good for simply cut/paste editing or remuxing). Actually, you should check the Video -> Color Depth options. They can override this behavior.

For simply creating a HuffYUV AVI from MPEG sources you should use Fast Recompress. You can still use simple cut/paste editing.

[JohnnyMalaria]

Just to add to the snooping the internal colorspace thing, it is up to the decoder to tell the software using it what colorspaces it can provide. What the decoder offers up doesn't necessarily tally with how the video is encoded. For example, a DV decoder may offer only RGB24 for output even though the video is encoded as a type of YV12. Furthermore, neither Video for Windows nor DirectShow really care about the encoded colorspace. Both technologies simply give the host program the right decoder to do the job. The only way to really dig into the video would be to know the detailed specifications for the format. e.g., you need to first of all determine the overall format (e.g., DV) and then, knowing how to decode a DV stream, find the specific information that identifies the colorspace. Quite simple for DV (it's 4:2:0 or 4:1:1). Other formats support a wider range of colorspaces.

[Froddy1]

Hello!

Thanks again to you for answering my last questions.

I would like to close the topic about colorspace as it's getting very detailed and also very hard to understand. (For me especially.)

I think for the first it will be good just to obey the provided steps by jagabo.

1) Use Cedocida instead of Panasonic DV Codec in future.
2) Change colorspace to YUY2.
3) Saving all my source videos to HuffYUV with VirtualDub, using "Fast recompress".
4) Importing this new files to Ulead VideoStudio and editing it.
5) Saving again as HuffYUV or directly encoding as mpeg2 or Divx.

Froddy!

@jagabo: I've now read the whole thread again. Did I understand you well:

When I convert ie. wmv-HD or mpeg2 or Divx video files (which all use YUV-Colorspace) to DV with the Panasonic DV codec, import the new files to Ulead Video Studio and than save my project as mpeg2, I'll get a so called double-colorspace-conversation? YUV to RGB and than back to YUV? Right? (A simple Yes or No will be satisfying.) I ask you because I remember that my last projects where made with this procedure.

[jagabo]

When I convert ie. wmv-HD or mpeg2 or Divx video files (which all use YUV-Colorspace) to DV with the Panasonic DV codec, import the new files to Ulead Video Studio and than save my project as mpeg2, I'll get a so called double-colorspace-conversation? YUV to RGB and than back to YUV? Right? (A simple Yes or No will be satisfying.)

Yes, that is correct.

But the other case where you convert everything to YUY2 then import into VideoStudio might be doing the same thing. I don't know how VideoStudio handles incoming YUY2 sources. It may convert to RGB before encoding in which case you have the same problem. I know that TMPEGEnc Plus does this -- or at least, the MPEG output from it shows all the symptoms (loss of blacks below luma 16, loss of brights above luma 235, and imprecise levels and colors in between) of having converted YUV to RGB and back.

I always use AviSynth to open files (AviSource(), Mpeg2Source(), DirectShowSource()) so I can check the colorspace as it comes out of the decompresor (Info()). Also keep in mind that DirectShow and VFW use different codecs. Just because DirectShowSource() gets video as YUY2 doesn't mean that VirtualDub opening the file directly (VFW) gets the same thing.

[Froddy1]

But the other case where you convert everything to YUY2 then import into VideoStudio might be doing the same thing. I don't know how VideoStudio handles incoming YUY2 sources. It may convert to RGB before encoding in which case you have the same problem.

Yeah, you're right. I'm actually trying to figure out how Ulead does handle imported video files, RGB or YUV or both. If Ulead only works in RGB than I'll have a double colorspace conversation anyway.

But you all wrote that DV works in YV12/YUY2. And Ulead VideoStudio 9 is a consumer software most notably designed for DV material from a DV-Camcorder. So I presume that Ulead *must* also work in YUV colorspace.

On the other hand if you would like to colorize the video, you can choose only from a palette of "RGB"-colors. (I'm not sure if this is has something to do with the colorspace.)

I always use AviSynth to open files (AviSource(), Mpeg2Source(), DirectShowSource()) so I can check the colorspace as it comes out of the decompresor (Info()). Also keep in mind that DirectShow and VFW use different codecs. Just because DirectShowSource() gets video as YUY2 doesn't mean that VirtualDub opening the file directly (VFW) gets the same thing.

As I also work with AviSynth (only for PAL-Speedup and Lanczos resizing) may I politely ask you how and where do you check the colorspace of the scripted video? And how is the exact "Info" command? I'm just curious.

Thanks
Froddy.

[jagabo]

Info()'s usage is simple:

WhateverSource("filename.ext")
Info()

Open the script with VirtualDub or a media player. Each frame will show stats about the video including colorspace at the point Info() was added. For example if you used:

WhateverSource("filename.ext")
ConvertToRGB()
Info()

You would see the colorspace is RGB -- because you converted to RGB before the Info() command.

I'll address the possibility of testing whether UVS converts to RGB internally tomorrow...

[Froddy1]

Info()'s usage is simple:

WhateverSource("filename.ext")
Info()

Open the script with VirtualDub or a media player. Each frame will show stats about the video including colorspace at the point Info() was added. For example if you used:

Thanks. I'll try it out.

You wrote a line which I always wanted to know:

What is the difference between AviSource, Mpeg2Source and DirectShowSource? I only use the command "DirectShowSouce" because I was told that it makes no difference. (DirectShowSource loads Avi and mpeg2 either).

I'll address the possibility of testing whether UVS converts to RGB internally tomorrow...

Thanks for doing it.

But after a minute's silence I don't think it's important to know weather Ulead converts interally to RGB or not.

If Ulead VideoStudio 9 only supports RGB than I'll have a colorspace conversation anyway. So it doesn't matter when my source file HuffYuv is encoded as YUY2 colorspace.

If Ulead VideoStudio 9 also supports YUV than it's good.

So my conclusion: It is the safest way to convert all my different source files uniformly to HuffYUV with YUY2 colorspace. At all events it will be no disadvantage.

Froddy!

[jagabo]

What is the difference between AviSource, Mpeg2Source and DirectShowSource?

DirectShowSource() uses Windows' DirectShow subsystem to open the file and decompress the video and audio. It works with any video for which you have a DirectShow file splitter and A/V decoders. AviSource() uses code taken from VirtualDub to open an AVI file. This uses Windows' VFW library, in part. So you need VFW video and ACM audio decoders. Mpeg2Source() uses DgMpgDec to open MPEG2 files.

[Froddy1]

Hello!

What is the difference between AviSource, Mpeg2Source and DirectShowSource?

DirectShowSource() uses Windows' DirectShow subsystem to open the file and decompress the video and audio. It works with any video for which you have a DirectShow file splitter and A/V decoders. AviSource() uses code taken from VirtualDub to open an AVI file. This uses Windows' VFW library, in part. So you need VFW video and ACM audio decoders. Mpeg2Source() uses DgMpgDec to open MPEG2 files.

And which command is now better for Avi and Mpeg2 files? DirectShowSource or AviSource/Mpeg2Source?

I have ffdshow installed (and I believe also ACM filter beside others), so both DirectShowSource and AviSource work well. But I use AviSynth mainly for PAL Speedup (23,976 to 25 fps) and importing to my Mpeg2-encoder because it's more comfortable as changing fps and audio playback rate manually.

Froddy.

[jagabo]

DirectShowSource() is more prone to problems -- a system configuration problem could keep it from working properly. Mpeg2Source() is more robust and works better with variable frame rate material. Use whichever works.

[Froddy1]

DirectShowSource() is more prone to problems -- a system configuration problem could keep it from working properly. Mpeg2Source() is more robust and works better with variable frame rate material. Use whichever works.

Okay.

Finally I want to thank for spending a lot of time to support me. This was very nice of you!

You wrote that you want to figure out which colorspace Ulead VideoStudio uses. It's not necessary to do it. I'll try to convert all my different souce files to HuvYuff (VY12). If Ulead supports YUV colorspace than it's good, if not it will convert to RGB anyway. So it doesn't make any sense to manually convert colorspace to RGB.

Best regards!
Froddy.