Removing wrapper

UncleBenji1977

Well-known member
I'm sorry, I tried to search for this but found nothing. I've decided to leave 24p unchecked as it seems to be far more stable. But now I have the headache of removing the 60i wrapper... Is there a cheaper solution than neoscene? I have After effects cs3... I just need options and more importantly options that won't degrade my image. Thanks.
 
Just hack to native 24p using the Blackout-Powell patch or the Max Latitude patch, they're quite stabile.
 
I've tried that... Haven't had great results with those... Crashes all over the place... Unless I did something wrong when hacking it?
 
Added the blackout/powell hack to my gh13 folder. loaded firmware into ptool, and clicked the B button where the settings were saved. inserted the sd card and installed firmware.... I just reloaded that hack, and ran some quick tests, and it's a little more stable than what I'm used to.... On a high detailed scene (grass/trees) I can get about 30 seconds before it crashes with no recovery of the file. It's fine with movement, and steady low detail shots seem to be stable... I can probably live with this until something more stable comes out seeing as there may be no other option outside of neoscene.
 
Added the blackout/powell hack to my gh13 folder. loaded firmware into ptool, and clicked the B button where the settings were saved. inserted the sd card and installed firmware.... I just reloaded that hack, and ran some quick tests, and it's a little more stable than what I'm used to.... On a high detailed scene (grass/trees) I can get about 30 seconds before it crashes with no recovery of the file. It's fine with movement, and steady low detail shots seem to be stable... I can probably live with this until something more stable comes out seeing as there may be no other option outside of neoscene.

There's some sort of AviSynth unwrapping solution but I have no idea how it works. After Effects can do it, I'm just not sure. You'd have to use trial and error to find out what the GH1's pulldown phase is, and then you'd just interpret all the footage and go from there. From what I've read and tested, the Blackout and Max Latitude are similarly stabile in FHD. Max Latitude does not have 720p60, however, only 720p30. And in my experience, the Blackout patch's 60 fps is pretty unstabile, but my experience with FHD has been stellar.
 
Thanks guys... I'll look at that Avisynth deal if I go back to the wrapper.

I have never gotten stability with 24p checked. I have never had a crash with my settings (without 24p checked) except once a long long time ago on a complicated chart...
 
There we go:

Original MTS

PAL 1080i footage

Assume TFF path

Code:
directshowsource("00000.mts").AssumeTFF().SeparateFields()
x0 = selectevery(2, 0,1).Weave().ConvertToYUY2()
return(x0)
Assume FrameBased path

Code:
directshowsource("00000.mts").ConvertToYUY2().AssumeFrameBased().ComplementParity()
VirtualDub UT conversion:

Code:
VirtualDub.video.SetInputFormat(0)
VirtualDub.video.SetOutputFormat(0)
VirtualDub.video.SetCompression(0x32796c75,0,10000,0)
VirtualDub.video.SetCompData(4,"AQEAAA==")
NLE Import dialogue (dropdowns show only the 1080i option):

Import_ScreenShot.jpg
 
I don't understand what you are trying to do? The 25p MTS original file that you sent me is seen as 25p in AE or virtual dub if brought in through avisynth (if you change the frame rate to 25 in virtualdub).

Are you trying to render an intermediate file?
 
Your script can be simplified to

DirectShowSource("00000.mts")

The rest is not necessary.




It's very likely an issue with your NLE software. Are there any options to interpret the file ? What software are you using ?

The other possibility is your directshow filters are acting up (directshow isn't reliable and can produce different results on differently configured systems) . How do you have it configured ? What splitter ? What audio & video decoder ? If you don't know , render the native file in graphstudio and report back the results

Also, when using UT video codec, did you uncheckmark the assume interlaced video box in the configuration

(Note: Technically this isn't "removing wrapper". The file is encoded as interlaced, but the content is progressive. The elementary stream has a header that signals it has been encoded as interlaced - thats why all NLE software" see" it as interlaced, and upsample the chroma as interlaced. It has nothing to do with the transport stream "wrapper" or "container". When you encode through avisynth, the exported intermediate doesn't have that flag, so to the other software will look progressive. It's not as simple as "stripping a flag", it's not a soft pulldown/telecine flag , like a progressive DVD. It's encoded as interlaced, so the analogy in DVD terms would be hard telecined - it's been encoded that way.)
 
Your script can be simplified
Do you mean that Directshow inputs to avisynth an already deinterlaced stream??

My understanding is that the stream is recorded as progressive and in-camera splitted and put in an interlaced "wrapper" (I'm sorry if it's the wrong word, I don't know how to otherwise express this technical point) ie. one frame top lines, the following one bottom lines. So I though I had to put them back together.

It's very likely an issue with your NLE software. Are there any options to interpret the file ? What software are you using ? How do you have it configured ?
Lightworks (I like it quite much, but it's actually a beta); as 25p project.

What splitter ? What audio & video decoder?
Haali Media Splitter & FFDShow.

when using UT video codec, did you uncheckmark the assume interlaced video box in the configuration
I did. Tried both options.

The elementary stream has a header that signals it has been encoded as interlaced - thats why all NLE software" see" it as interlaced, and upsample the chroma as interlaced
Actually VirtualDub doesn't allow me to transcode to UT if I don't convert to YUY before, in avisynth pre-processing.

When you encode through avisynth, the exported intermediate doesn't have that flag, so to the other software will look progressive.
I can't get it seen as progressive, but always as interlaced.

I can get it seen as progressive just when I rewrap the stream in a MOV container (w/ ffmpeg or ffmbc) although not every time. :-/
 
Last edited:
Do you mean that Directshow inputs to avisynth an already deinterlaced stream??

My understanding is that the stream is recorded as progressive and in-camera splitted and put in an interlaced "wrapper" (I'm sorry if it's the wrong word, I don't know how to otherwise express this technical point) ie. one frame top lines, the following one bottom lines. So I though I had to put them back together.

No, it's encoded (recorded) interlaced, but the content is progressive. In this scenario, the "interlaced" refers to the method of encoding, not the content. The content is what is important. "Progressive content" means consecutive field pairs are the same moment in time. "Interlaced content" means consecutive field pairs contain information from different moments in time. That's why you see "combing" or those horizontal lines with interlaced content

When you process it through avisynth into an intermediate, both the encoding, and content is progressive. There should be no way for software to misintpret the file




Lightworks (I like it quite much, but it's actually a beta); as 25p project.

Very buggy software. I would wait a while until it gets developed a bit farther and becomes stable.

Even blender's VSE is a lot better


Actually VirtualDub doesn't allow me to transcode to UT if I don't convert to YUY before, in avisynth pre-processing.

If you are using UT 422, then add ConvertToYUY2() in the script

I can't get it seen as progressive, but always as interlaced.
Very likely Lightworks problem . No other software has problems with this.

Most software will allow you to interpret the file, can you do that in Lightworks ? (I unstalled it and reformatted my disk. It's very unstable software. I don't intend to test it again until it gets a lot more stable)

Another option is to use Matrox VFW MPEG2 I-frame as an intermediate. It seems to run in Lightworks ok.
 
Very likely Lightworks problem . No other software has problems with this.

I'm not quite sure:
http://www.dvxuser.com/V6/showthread.php?252435-Edius-conversion

There is no chance that Edius recognizes video clips made by VirtualDub as progressive (even when I created video from picture sequence and then encoded it with Canopus codecs), except if encoded to XViD. I've tried with UT as well.
I was 100% sure that the problems were with NTSC clips, but it seems that PAL is not immune as well.
 
Back
Top