Possible to deinterlace MTS video with QTGMC?

[audiohack]

Hi,
Is it possible to deinterlace AVCHD video with the .mts file extension using QTGMC? I'm a complete noob when it comes to deinterlacing. I recently deinterlaced SD MiniDV footage (.avi extension) using an Avisynth+ script with QTGMC and VirtualDub2. I'm very pleased with the results - better than deinterlacing with Adobe Media Encoder. But when I tried doing the same with AVCHD video with the .mts file extension, the results are not good. The deinterlaced video jerks backwards and forwards.

The script is below. I tried changing the Assume field designation to BFF and the results are the same. Adobe Media Encoder is able to deinterlace it, but I was thinking that QTGMC might provide a better (and much cheaper!) result ... if I could get it to work. Any help is greatly appreciated! Thanks.

SetFilterMTMode ("QTGMC", 3)
FFMPEGSource2("soccer.mts", atrack=1)
AssumeTFF()
QTGMC(preset="Slower", EdiThreads=3)
Prefetch(10)

Here is some info from MediaInfo about the video file:

Width : 1920 pixels
Height : 1080 pixels
Display aspect ratio : 16:9
Frame rate : 29.970 (30000/1001) FPS
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Interlaced
Scan type, store method : Separated fields
Scan order : Top Field First

[WorBry]

The deinterlaced video jerks backwards and forwards.


It's difficult to understand this result, based on the information you have provided. If this is a native interlaced (59.94i) AVCHD.mts clip your script should work fine. If it was recorded in Progressive Segmented Frame format (29.97 PsF) i.e. progressive frame encoded as interlaced fields, you would simply get duplicate interpolated frames.

[poisondeathray]

Could be a source filter issue; try lsmash instead of ffms2. ffms2 is less reliable with transport streams

https://github.com/HolyWu/L-SMASH-Works/releases

LWLibavVideoSource("soccer.mts")

[audiohack]

Could be a source filter issue; try lsmash instead of ffms2. ffms2 is less reliable with transport streams

https://github.com/HolyWu/L-SMASH-Works/releases

LWLibavVideoSource("soccer.mts")

Thanks! That did the trick for the video. But I lost the audio track in the rendered file that I was getting with the original FFMPEGSource2("soccer.mts", atrack=1) command. I've been experimenting with LSMASHAudioSource but no success yet.

[audiohack]

It's difficult to understand this result, based on the information you have provided. If this is a native interlaced (59.94i) AVCHD.mts clip your script should work fine. If it was recorded in Progressive Segmented Frame format (29.97 PsF) i.e. progressive frame encoded as interlaced fields, you would simply get duplicate interpolated frames.

I appreciate your reply. Sorry if I didn't provide the right information. I'm new to this. I would post an example of the rendered video, but it's of my young nephews and I don't feel comfortable uploading footage of them to an internet forum. I'll look into the PSF format. I'm not sure if my footage is in that format or not.

[poisondeathray]

Thanks! That did the trick for the video. But I lost the audio track in the rendered file that I was getting with the original FFMPEGSource2("soccer.mts", atrack=1) command. I've been experimenting with LSMASHAudioSource but no success yet.


Aud=LWLibavAudioSource("soccer.mts")
Vid=LWLibavVideoSource("soccer.mts")
AudioDub(Vid,Aud)

[audiohack]

Aud=LWLibavAudioSource("soccer.mts")
Vid=LWLibavVideoSource("soccer.mts")
AudioDub(Vid,Aud)


Thank you!! That did it!

[WorBry]

..... I would post an example of the rendered video, but it's of my young nephews and I don't feel comfortable uploading footage of them to an internet forum. I'll look into the PSF format. I'm not sure if my footage is in that format or not.

No bother. If changing to lsmash did the trick, that's great. I always use DGDecNV myself.

[audiohack]

No bother. If changing to lsmash did the trick, that's great. I always use DGDecNV myself.

Is there an advantage to using DGDecNV over L-Smash?

[DJATOM]

Speed.

[poisondeathray]

Is there an advantage to using DGDecNV over L-Smash?

It's also more reliable for supported formats, I'd say the most reliable

Disadvantages - requires a compatible Nvidia card, small fee

[audiohack]

Thanks for everyone's contributions. At the moment I'm working with a laptop, which I see are not supported by DGDecNV's developer because of different laptop maker's implementations of OPTIMUS switching and other OEM driver related requirements. It does have an NVIDIA GTX 1070 in it, so I may roll the dice on a $15 license to see if it works. Either way, at least I can start deinterlacing my .mts footage with L-smash. Thanks again!

[Emulgator]

DGDecNV works well on most nVidia Laptops, I use it frequently, and only on laptops.
It might be just a disclaimer that a few configurations might be excluded.
Yours with GTX1070 should be fine.

[audiohack]

DGDecNV works well on most nVidia Laptops, I use it frequently, and only on laptops.
It might be just a disclaimer that a few configurations might be excluded.
Yours with GTX1070 should be fine.

Thanks. I'm going to give it a go.