Tool to change the Display aspect-ratio?

[cumi]

Can anyone suggest a free tool to change the DAR "Flag" of a an AVI captured from DV Camcorder? I allready tried the mpeg4modifier, but the source is not mpeg4.

I want to do this, because youtube and the Media Player doesn't recognises, that the captured movie has 16:9 and not 4:3 AR. I don't want to convert the whole video in the first step, just modify the AR, if it is possible.

Thanks

[guns1inger]

WMP won't notice if you do change it - it is stupid that way. I can't speak for Youtube, but I suspect it would also be more comfortable with video that had a 1:1 PAR.

That said, try the Enosoft DV processor.

[manono]

You'd better reencode it as they won't recognize it even if you do change the flag. They expect 1:1 videos.

[cumi]

So you say, I should simply re-process the whole video and convert it from 720x576 to 1024x576?

[manono]

Or some other 1.78:1 ratio. 624x352, for example. They're still going to reencode it themselves, whatever you send them - as you discovered.

[Froddy1]

Hello,

what a coincidence, I have the same question:

I have a wrong flagged "DVC/DV Video" file, which should be native 16:9 but is flagged as 4:3. So the picture is streched. Mpeg4modifier does not accept this file.

How can I change the aspect ratio?

Am I also able to change the aspect ratio of a "Microsoft AVI" (i.e. BI_RGB Raw Bitmap avi)?

Please be so kind to help me.
Thx, Froddy!

[Baldrick]

Froddy: Tried Enosoft DV Processor ? It claims to support aspect ratio change.

[Froddy1]

Froddy: Tried Enosoft DV Processor ? It claims to support aspect ratio change.

Hello Baldrick!

No, I can't use Enosoft DV Processor. Although I meet the system requirements I'm not able to load this programm. Well, my computer is pretty old...

Any other suggestion of how to change the aspect ratio of an DV-Video file from a Camcorder?

(The video-file is 720x576 and flagged as 4:3 but should be in native 16:9 aspect ratio). Ulead Video Studio loads it as 4:3 or adds letterbox in a 16:9-project.)

I also use VirtualDUB and I could be able to resize this file correctly. But in which target resolution? And will be a great loss of quality when reencoding it back to DVC/DV DigitalVideo AVI?

Thx.
Froddy!

[guns1inger]

DV AVI PAL only has one resolution - 720 x 576. The difference between 16:9 and 4:3 is the way the pixels are displayed, which is determined by a flag.

What happens when you try to install the enosoft DV processor ?

[JohnnyMalaria]

Froddy1,

Sometimes installations fail (in general, not just our software) if a previous installation of something else didn't finish either because of unhandled error or because a reboot is required to complete the installation. Can you describe what happens? Looking at your PC profile, it should be fine.

All Windows installation programs can create a log file that can help pinpoint problems. Whether a log is created for our program depends on how your Windows is configured. If it is configured to create a log, it will save it to:

C:\Documents and Settings\YOURPROFILE\Local Settings\Temp

with a thoroughly unhelpful name such as MSIxxxxx.log. You need to look for the most recent one after attempting to install (it will be a few 100K in size). You can attach it to a reply (don't copy-and-paste the contents since it has 1000's of lines!) If it doesn't create a log, let me know and I'll let you know to force it.

Our software will change the aspect ratio flag in the DV stream itself, not the AVI header. I'm not sure if your encoding software looks that deeply into the file or not. Changing the aspect ratio in the DV stream requires creating a new file and each frame has to have its aspect ratio changed. If your encoding software simply looks at the AVI header then a much simpler and faster is possible using other software that can edit the AVI header. Such software is likely to find Type-1 DV AVI files confusing but Type-2 should be okay.

John.

[Froddy1]

Hello everybody! Thank you for the fast help!

DV AVI PAL only has one resolution - 720 x 576. The difference between 16:9 and 4:3 is the way the pixels are displayed, which is determined by a flag.

Yes, you're right. There is only 720x576 on PAL DV AVI. To edit this file in Ulead Video Studio with the correct aspect ratio I thought of resizing it to 1024x576 with VirtualDUB which will than have an aspect ratio of 1,78 (16:9).

Would that be a good idea?

After cutting and editing in Ulead Video Studio I plan to save the new video as PAL-DV-AVI 720x576 with 16:9. So I prevent not to have a loss in vertical resolution (I mean 576 lines).

What happens when you try to install the enosoft DV processor ?

It seems to be my bad that Enosoft DV Processor doesn't work. Although I have Windows XP SP3 I still work with an old PIII-S Tualatin 1,4 Ghz CPU which does not support the required SSE2 commands. So therefore the setup abruptly stops and rolls back installation.

Our software will change the aspect ratio flag in the DV stream itself, not the AVI header. I'm not sure if your encoding software looks that deeply into the file or not. Changing the aspect ratio in the DV stream requires creating a new file and each frame has to have its aspect ratio changed. If your encoding software simply looks at the AVI header then a much simpler and faster is possible using other software that can edit the AVI header. Such software is likely to find Type-1 DV AVI files confusing but Type-2 should be okay.
John.

Thank you for the information. I am not so familiar with the DV-Digital Video codec: Will this process you mentioned above provide a loss of quality?

Sorry for that question: I you reencode for example a mpeg2-file you will have a slight quality worsening. Does this also happen to DV-Digital Video?

I'm thinking of resizing the file to 1024x576 and than make some edits in my video editing software and save it again as DV. (I could not still find any solution directly in Video Studio.)

Then I want to use my preferred MainConcept Encoder as mpeg2-Encoder to create a DVD.

I fear that I will obtain a loss of picture quality if I reencode this file too much. I hope you understand my concern.

But thanks for the useful hint with the DV-processor. I'll keep that in mind.

Froddy.

[poisondeathray]

Yes, DV is lossy, so all your conversion steps listed will result in compounding quality loss at each stage.

No it doesn't make sense to resize it to 1024x576, then resize it back to DV @ 720x576, you just incur more quality loss

If your final format goal is a DVD, just edit the native DV footage, then encode straight to DVD. This is the shortage route, with least amount of quality loss.

Don't do any resizing; this causes more quality loss. (for DVD, there is an aspect ratio flag for 16:9, the frame size is still 720x576 in PAL land but it displays as 16:9)

If you just wanted to change the aspect ratio for viewing on a PC, without quality loss, one option is to place the DV-AVI into an .mkv container with mkvmergegui and change the aspect flag to 16:9. Most popular software will obey the flag (e.g. VLC, SMPlayer etc....)

[jagabo]

You can force VirtualDub to display the before/after panes with whatever aspect ratio you want. Just right click on a pane and select the AR. Most media players have this too. And most MPEG conversion and DVD authoring software allows you to override the DAR of the the source files. So even if a program thinks the the DV AVI is 4:3 DAR you can force it to 16:9. In short, there's not much reason to worry about the DAR flag in the DV AVI.

[Froddy1]

Hello,

Yes, DV is lossy, so all your conversion steps listed will result in compounding quality loss at each stage.

No it doesn't make sense to resize it to 1024x576, then resize it back to DV @ 720x576, you just incur more quality loss

Okay, thx, good to know. But DV is less problematic than mpeg2, right? GSpot states me that DV-video files have about 28000 kbps which I consider as huge.

If your final format goal is a DVD, just edit the native DV footage, then encode straight to DVD. This is the shortage route, with least amount of quality loss.

You gave me a bright idea: In MainConcept Encoder you can select if the file (720x576) should be in 4:3 or 16:9. With Ulead Video Studio you will only get a letterboxed picture,even in a "16:9 project". Beside that I don't think that the built-in encoder of Ulead is as strong as Mainconcept or Canopus Procoder.

Don't do any resizing; this causes more quality loss. (for DVD, there is an aspect ratio flag for 16:9, the frame size is still 720x576 in PAL land but it displays as 16:9)

Okay, I will obey your advice - no resizing.

So I will do that: Loading that wrong flagged DV-video to Ulead, editing it as usual and then saving it again as DV with the same parameters as the source file has. Then I use my external Mpeg-Encoder to create a correctly flagged DVD video.

If I wish to encode it as Divx I'll use Mpeg4modifier to change the DAR after encoding.

So far so good?

My last question refering to lossy compression:

What do you think about extracting the whole DV-video to uncompressed (Raw Bitmap) AVI, editing this file in Video Studio and then converting this new uncompressed material as mpeg2 in my external encoder? Better or useless? I know that the file then is getting very big... I intend not to cut the video solely but also to add some coloring and overlay effects so that the original video will be heavily modified.

BTW: A very nice forum here :-)

Froddy!

[jagabo]

What do you think about extracting the whole DV-video to uncompressed (Raw Bitmap) AVI, editing this file in Video Studio and then converting this new uncompressed material as mpeg2 in my external encoder? Better or useless?

Useless. Worse than useless depending on what software you're using. DV uses the YUV colorspace internally. Converting to RGB will result in a small loss of accuracy and can result in loss of detail in very dark and very light areas of the video.

In general video editors/encoders work like this:

1) Decompress source video to YUV or RGB (frame by frame, as needed).
2) Apply filters to decompressed YUV/RGB frames.
3) Encode filtered YUV/RGB frames to output format.

By decompressing your DV and saving as RGB all your doing is performing the first step and creating a huge intermediate file as a separate step. And if the editor works in YUV you risk losing light/dark detail. Even worse, the editor may convert the RGB back to YUV for filtering, losing even more accuracy!

[Froddy1]

What do you think about extracting the whole DV-video to uncompressed (Raw Bitmap) AVI, editing this file in Video Studio and then converting this new uncompressed material as mpeg2 in my external encoder? Better or useless?

Useless. Worse than useless depending on what software you're using. DV uses the YUV colorspace internally. Converting to RGB will result in a small loss of accuracy and can result in loss of detail in very dark and very light areas of the video.

By decompressing your DV and saving as RGB all your doing is performing the first step and creating a huge intermediate file as a separate step. And if the editor works in YUV you risk losing light/dark detail. Even worse, the editor may convert the RGB back to YUV for filtering, losing even more accuracy!

Okay, thanks. You're right: "Raw Bitmap" is RGB.

So should I better remain with the DV codec and not uncompress the video?


But I'm a little bit confused now. Because you also write:

In general video editors/encoders work like this:

1) Decompress source video to YUV or RGB (frame by frame, as needed).
2) Apply filters to decompressed YUV/RGB frames.
3) Encode filtered YUV/RGB frames to output format.

I don't understand your words linguistical: Am I meant to execute these three steps or is this the automatic procedure of the Video Editor?

I also have the following choices to decompress a video: RGB24, RGB32, RGB555, RGB565, YUY2, YV12.

Should I keep DV as a working codec or decompress the video file before editing? And if I should decompress video for editing purposes which uncompressed codec should I use?

(I feel sorry about my questions but video editing is absolutely new for me.)

Thank you!
Froddy.

[jagabo]

So should I better remain with the DV codec and not uncompress the video?

Yes, you should leave your video as DV until your final output -- MPEG for DVD, Divx, whatever. As noted earlier, the exception to this would be if your filtering in one program and then need to filter more in another program. In that case you will benefit by saving to an uncompressed, or losslessly compressed (HuffYUV, Lagarith), format in the first program.

In general video editors/encoders work like this:

1) Decompress source video to YUV or RGB (frame by frame, as needed).
2) Apply filters to decompressed YUV/RGB frames.
3) Encode filtered YUV/RGB frames to output format.

I don't understand your words linguistical: Am I meant to execute these three steps or is this the automatic procedure of the Video Editor?

I meant the software does all that automatically in the process of editing/filtering/converting. I was just pointing out that, at best, decompressing your video would simply be turning the first step into a separate process. A waste of time and space.

I also have the following choices to decompress a video: RGB24, RGB32, RGB555, RGB565, YUY2, YV12.

Should I keep DV as a working codec or decompress the video file before editing? And if I should decompress video for editing purposes which uncompressed codec should I use?

YV12 is the closest to PAL DV. YUY2 is the closest to NTSC DV. But again, you generally don't want save an uncompressed intermediate file.

Another thing to note is that smart editing software will not decompress and recompress DV frames if not necessary. For example, if you are simply cut/paste editing and adding transition effects, only the frames involved in the transition effects need to be decomrpessed, processed, then recompressed. The rest of the frames will pass through untouched (and thereby not losing any quality). Less intelligent editors will decompress and compress every frame. If you are applying filters (noise reduction, color correction, etc) to the entire video all frames have to be decompressed, filtered, and recompressed.

[Froddy1]

Hello! I'm back. I hope you are still monitoring this thread.

So should I better remain with the DV codec and not uncompress the video?

Yes, you should leave your video as DV until your final output -- MPEG for DVD, Divx, whatever. As noted earlier, the exception to this would be if your filtering in one program and then need to filter more in another program. In that case you will benefit by saving to an uncompressed, or losslessly compressed (HuffYUV, Lagarith), format in the first program.

Okay, so I will stay with the DV-codec. But below is my final question, I would need your suggestion...

Another thing to note is that smart editing software will not decompress and recompress DV frames if not necessary. For example, if you are simply cut/paste editing and adding transition effects, only the frames involved in the transition effects need to be decomrpessed, processed, then recompressed. The rest of the frames will pass through untouched (and thereby not losing any quality). Less intelligent editors will decompress and compress every frame. If you are applying filters (noise reduction, color correction, etc) to the entire video all frames have to be decompressed, filtered, and recompressed.

That is the point: I only use Ulead Video Studio for color correction/grading and for the intro & credits. (For simple cuts I use VirtualDub with SmartRendering function.)

So my video mostly has to be rendered completely, not only parts of it.

So should I still stay with the DV codec or better uncompress or transform to a more lossyless format?

My procedure now would be: Loading the file in Videostudio as DV-PAL, editing it (i.e. color correcton), saving it again as DV and then loading the new DV videofile to my external Mpeg2 or Divx-Encoder. Okay so?

Thanks, Froddy.

[guns1inger]

Dv is lossy, but it is also far more resiliant that, say, mpeg-2 or mpeg-4. That is, you can go one or two generations, with good software, without make noticeable reductions in quality. With very good software, it may be possible to go even more. So long as you can do all your filtering in a single go in videostudio, and the filtering itself doesn't introduce artifacts, you should be OK. I suspect that VideoStudio probably uses the mainconcept mpeg-2 encoder, so you should investigate outputting directly from videostudio to mpeg-2 if you need that as an output.

[jagabo]

VideoStudio -> intermediate file -> MPEG encoder

For the best results from this scenario you should use a use an uncompressed or losslessly compressed format like HuffYUV or Lagarith. Using DV shouldn't be too bad though.

As guns1inger pointed out, you could just do your MPEG or Divx encoding right from VideoStudio to avoid using the intermediate file.

[Froddy1]

Dv is lossy, but it is also far more resiliant that, say, mpeg-2 or mpeg-4. That is, you can go one or two generations, with good software, without make noticeable reductions in quality.

Okay, thank to you both for the helpful information.

I suspect that VideoStudio probably uses the Mainconcept mpeg-2 encoder, so you should investigate outputting directly from videostudio to mpeg-2 if you need that as an output.

I don't know but you might have right. It seems that the built-in Ulead Video.Now.Encoder is a licensed product of MainConcept. I definitely know that Adobe Premiere uses Mainconcept Mpeg2 encoder, respectively, there is a plug-in available for Abobe Premiere. But that doesn't matter.

VideoStudio -> intermediate file -> MPEG encoder

For the best results from this scenario you should use a use an uncompressed or losslessly compressed format like HuffYUV or Lagarith. Using DV shouldn't be too bad though.

As guns1inger pointed out, you could just do your MPEG or Divx encoding right from VideoStudio to avoid using the intermediate file.

I took a quick look at the HuffYUV codec: I have ffdshow installed and there I can select between YV12 and YUY2 colorspace. Which one is the correct setting?

And btw, just in case:

Would be HuffYUV also a good lossyless format for other video formats than DV? Because I sometimes work with HD-Videos (ie. WMV-HD, QT, 1280x720) and before editing with VideoStudio I have to resize it to 720x576 PAL because my CPU is too weak to handle such big files.

So, in that -new- case it would be more beneficial to decode and resize these HD-files to HuffYUV, editing it and than saving again as HuffYUV or encoding directly to mpeg2/Divx. Right? (Till today I've transformed HD-Video files also to DV-PAL before loading them to my Video Editor.)

Thank you
Froddy1

[jagabo]

I took a quick look at the HuffYUV codec: I have ffdshow installed and there I can select between YV12 and YUY2 colorspace. Which one is the correct setting?

In general I'd say use YUY2. It has twice as much color resolution and is less likely to lead to chroma sub-sampling errors in interlaced video.

[Froddy1]

I took a quick look at the HuffYUV codec: I have ffdshow installed and there I can select between YV12 and YUY2 colorspace. Which one is the correct setting?

In general I'd say use YUY2. It has twice as much color resolution and is less likely to lead to chroma sub-sampling errors in interlaced video.

Hello jagabo!

I discovered that you wrote this in a former post of this thread:

YV12 is the closest to PAL DV. YUY2 is the closest to NTSC DV. But again, you generally don't want save an uncompressed intermediate file.

This sentence is contradictory in my eyes.

I live in Europe and therefore I'm using PAL. So should I still use YUY2 colorspace?

(Once again, I sometimes use wmv/H264 HD video files for editing which are too big for my PC. So I want to resize these files to an intermediate format. As far as I understood, you suggest me using HuffYUY.)

Sorry for asking you again.

Thanks
Froddy!

[manono]

In my opinion you save them to the colorspace your MPEG-2 encoder prefers. Didn't you say you'll be encoding to MPEG-2 using MainConcept? I'm not entirely sure which it prefers, but I think it's YUY2. So I agree with jagabo, but for a different reason. Now, if you intend to encode using HCEnc then you'd convert to YV12.

This sentence is contradictory in my eyes.

I'd say it depends on the context in which he said it. Maybe the situation was different for the other poster. I don't know as you didn't give a link to the thread.

[jagabo]

Whether to use YUY2 or YV12 depends on several factors including what colorspace and subsampling the editor is works in, what the final encoder accepts (HcEnc for example only accepts YV12), what algorithms different programs use to convert colorspaces, etc. I don't know what colorspace VideoStudio works in but I think it's safest to save with 4:2:2 subsampling (YUY2). That will also avoid improper handling of interlaced YV12 -- a problem with many programs.

[Froddy1]

Hello!

In my opinion you save them to the colorspace your MPEG-2 encoder prefers. Didn't you say you'll be encoding to MPEG-2 using MainConcept? I'm not entirely sure which it prefers, but I think it's YUY2. So I agree with jagabo, but for a different reason. Now, if you intend to encode using HCEnc then you'd convert to YV12.

Yes, you're right. I normally intend creating mpeg2 with Mainconcept Encoder (Maybe I'll use the built-in encoder of VideoStudio in future.) But I absolutely don't know which colorspace either Mainconcept or Ulead Video Studio uses. I've never thought about it. Today it's the first time I'm deliberately dealing with this.

This sentence is contradictory in my eyes.

I'd say it depends on the context in which he said it. Maybe the situation was different for the other poster. I don't know as you didn't give a link to the thread.

It's not that important but here is the link (see the last lines of the post):
https://forum.videohelp.com/topic362385.html#1928219
(Subject was either to decompress DV-files or not and I have several choices how to decompress video (RGB24, RGB32, RGB555, RGB565, YUY2, YV12). So I was told to use "YV12" because it is the closest to PAL DV. Now, you both recommend me to use YUY2 if I want to transform my footage to HuffYUV as an intermediate format. That's why I was confused...)

Whether to use YUY2 or YV12 depends on several factors including what colorspace and subsampling the editor is works in, what the final encoder accepts (HcEnc for example only accepts YV12), what algorithms different programs use to convert colorspaces, etc. I don't know what colorspace VideoStudio works in but I think it's safest to save with 4:2:2 subsampling (YUY2). That will also avoid improper handling of interlaced YV12 -- a problem with many programs.

Okay, so if I have to convert to HuffYUV I will use YUY2.

My difficulty is that I also use parts of external video files for creating my personal video. So I have different sources like WMV-HD, DVD-Mpeg2, Divx in one project, and I don't know which colorspace is used in this codecs. (I hope it is not important to know.)

BTW: During research about different colorspaces I discovered this website: http://www.animemusicvideos.org/guides/avtech/colorspace.html, especially the last sentences about unintentional colospace conversation is quite interesting.

I quote:

"Many people once used YUY2 mode in HuffYUV all the time thinking it was lossless - so they would import into Premiere, edit, export to YUY2, import again, edit, export again and each time it went into Premiere it was being converted into RGB and each time it was compressed to HuffYUV it was being converted to YUY2. This has lead to many faded-looking videos in the past."

Well, this is excately what might happen to me: If necessary I intend to import all my source files as HuffYUV (YUY2), editing and than saving as HuffYUV again to export this file later to my Mpeg2-Encoder.

Thank You!
Froddy!

[poisondeathray]

WMV-HD, DVD-MPEG2, and DIVX are all YV12. As a general rule, you should keep the same colorspace, because conversions between colorspaces are lossy. I think there might be some confusion when you add interlaced DV into the mix - are you talking about the same project or different project? Does your project have mixed inputs?

[jagabo]

If you find VideoStudio works in RGB you can use HuffYUV or Lagarith in RGB mode.

[Froddy1]

Hello! (I'm back, I had the flu.)

WMV-HD, DVD-MPEG2, and DIVX are all YV12. As a general rule, you should keep the same colorspace, because conversions between colorspaces are lossy. I think there might be some confusion when you add interlaced DV into the mix - are you talking about the same project or different project? Does your project have mixed inputs?

Ok. Thanks for the information about colorspace.

Yes, you're right: My actual project consists of different inputs (wmv-HD, Divx and DVD-mpeg2).

Although you can directly import all of these files to Video Studio 9 the software or my PC are too weak to handle this files fluently especially when you add some color filters etc. Therefore I want to recode all source videos to a uniform video codec. First I wanted to convert to Uncompressed Raw Bitmap, but you all told me not to use it.

So I'will select "HuvYUFF YV12" which is included in the ffdshow filters. Right?

If you find VideoStudio works in RGB you can use HuffYUV or Lagarith in RGB mode.

Here I have two problems: I still couldn't figure out which colorspace Ulead Video Studio uses and I'm only able to convert HuvYUFF in YUY2 or YV12 colorspace.

Thank You!
Froddy

[jagabo]

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.