Capture VHS without deinterlacing

[MagicSparky]

Hi all,

in the capture options of my Terratec Grabster 250 device driver I can only choose one of the following deinterlace methods:

-HIGH Motion
-LOW Motion
-WEAVE
-Bob

I want to transfer some old VHS PAL tapes to DVD and want to keep the best quality possible.
I read in some post that deinterlacing is not good, if planned to play the authored DVDs on TV.
I don't want to throw away video data through deinterlacing.

My question(s):

- Is the method Weave the right one to choose. Does it throw away video data?
- Can I then encode to interlaced MPEG2 without having the mouseteeth effects which I get with the Weave option?

..or is it maybe better to buy an Canopus ADVC-110 (or so) device to transfer DV to PC, and then encode to MPEG2?


Thank you so much.

------------
MagicSparky

[themaster1]

I don't understand, you are asking which deinterlace method is the best yet you don't wanna deinterlace .....i'm kinda lost

For the record, the best method would be a weave+a bob at the same time, but that's impossible to achieve.

If you wanna make a dvd, indeed you must do nothing (not deinterlace) for a top notch quality and use a lossless codec such as huffyuv

The software i use for this task is virtual vcr (subject to A/V desync with my device unfortunately) but by far the best ...well to me.

Considering the Dv format take less space within a HDD and that some Dv camcorders have a passthrough ability and convert an analog signal into digital directly i guess that's the very safe option here.

[MagicSparky]

Thanks, but I got lost somewhere...

I read the following at http://www.100fps.com/ :
"Weave: do nothing. Show both fields per frame. This basically doesn't do anything to the frame, thus it leaves you with mice teeth but with the full resolution, which is good when deinterlacing is NOT needed."

Because of that I thought that this was the right option to capture with.

Is it true, if I use a DV-Device (like Canopus ADVC-110) and transfer DV via Firewire from the the VHS player, that the video file is then interlaced?!

Sorry, I must read some more about it...

[jagabo]

Use the weave setting. ADVC-100 will weave.

[Richard_G]

Is it true, if I use a DV-Device (like Canopus ADVC-110) and transfer DV via Firewire from the the VHS player, that the video file is then interlaced?!

Yes

I'm not familiar with the device you have but can you select none of the options? If you don't choose a deinterlace method, hopefully it won't deinterlace.

[MagicSparky]

@Richard_G:

No, I must select one of the options. Default is LOW LOTION.
I've tried Low- and High-motion and there are no interlaced artifacts then, but the picture is somewhat smooth with less details.

@jagabo

You said ADVC 100 will weave. If I understand it correct the ADVC will deinterlace the material too (because of the weave) and then send DV via Firewire to PC?!

It's very confusing...

[jagabo]

You said ADVC 100 will weave. If I understand it correct the ADVC will deinterlace the material too (because of the weave) and then send DV via Firewire to PC?!

It's very confusing...

Actually, it's very simple:

Broadcast video is sent as a series of half pictures, called fields. One field consists of the topmost scanline and every other scanline down from there (ie, scanlines 0, 2, 4, 6... 574). It's called a top field because it contains the topmost scanline. The next field consists of all the scanlines that were missed in the first field (scanlines 1, 3, 5, 7... 575). It's called a bottom field because it contains the bottommost scanline (or simply because it's not a top field). When you watch this on an interlaced TV you see one field at a time. By the time a field is being drawn the previous field has faded away. You never see an entire frame (both fields) on an interlaced TV screen. You see an alternating sequence of top and bottom fields. In PAL video you see 50 fields per second.

When computers capture interlaced video they normally combine two sequential fields together to make complete frames. One method is to simply weave the two fields togther. Since pairs of fields contain complimentary scanlines the frames are completely filled. Since each frame contains two fields, PAL's 50 fields per second is now 25 frames per second. But you will see comb artifacts whenever something moves during the 1/50th of a second that elapsed between the two fields. I don't think of weaving as a form of deinterlacing because the frame still contains two intact fields. When properly played back on an interlaced TV you will see exactly the same thing as the original broadcast - one field at a time, 50 fields per second.

Other forms of deinterlacing attempt to create progressive images with no comb artifacts. BOB deinterlacing fills in the missing scanlines with data interpolated from other scanlines of the same field. 50 field per second video is thus converted to 50 frames per second. In theory, this could be played back on an interlaced TV and deliver exactly the same image as in the original broadcast (the frames contain the original fields plus the interplated data) but in practice DVD doesn't support 50 frame per second encoding.

I don't know exactly what your software is doing during low/high motion deinterlacing. Presumably remove comb artifacts with some smart algorithms that examine the prior and/or subsequent fields/frames, using different methods for low and high motion video. Smart deinterlacing like this will compare at least two frames the previous one and the current one. Parts of the frame that don't change (no motion) are simply woven together in order to retain as much detail as possible. Parts that do change get some other treatment ranging from simply interpolating the missing scanlines to something more complex that tries to take into account the motion.

The disadvanage of smart deinterlacing is that it reduces temporal resolution from 50 pictures per second to 25 pictures per second. Motions will be noticably less smooth. And smart deinteracers are not perfect. Some comb artifacts get through and some areas are needlessly deinterlaced resulting in a blurrier picture.

[VD]

The Canopus ADVC 100 does NOT deinterlace. It will simply pass the DV to the computer via the firewire port. A DV "capture" program will copy the DV 1:1 maintaining the interlace. All capturing takes place in the Canopus box, the bits are just copied to the computer in DV format and wrapped in an AVI container.

Some programs will allow to transcode the DV to another codec on the fly. In this case you can select a deinterlace option if you want. But note, the program deinterlaces, not the Canopus ADVC.

[jagabo]

The Canopus ADVC 100 does NOT deinterlace.

It depends on your definition of "deinterlace". Many programs consider weaving to be a form of deinterlacing because the individual fields are not stored seperately but rather are woven together into frames. By this definition Canopus ADVC devices "deinterlace".

Personally, I don't consider weaving to be deinterlacing because the process loses nothing and is reversible.

[MagicSparky]

Wow, jagabo,

thank you so much for this detailed explanation. And thank you, VD. I think now I have understand it.

I ordered a Canopus ADVC-110 to check this out. I can return it with money back warranty.
I'm quite interested how the Terratec performs against the Canopus.

My plan is to transfer (or capture) 60 VHS Tapes to PC and then encode to DVD.
The software I plan to use is WinDV, Virtualdub, Avisynth, some Plugins and HCenc.
I think I have a lot to do with testing and (understanding various things).

Thanks again.

-----
MagicSparky

[themaster1]

i just found a nice little piece of software which work with my card "terratec cynergy 250 pci" it's called fly 2000 tv

it's a mix bewteen virtual vcr,win dvr, dscaler in my opinion plus "so far" it's stable unlike chris tv

you can record in mpg or avi

I thought i should inform you

[MagicSparky]

Cool,

thank you.
I will try this out. My ADVC-110 still not arrived. I have time to test it.

[Vegito72]

I actually a question similar to this topic, so I'll ask it here.

Would it be a good idea to de-interlace an interlaced video if you plan to play it back on a dvd-player that is hooked up to an HDTV? Will your DVD-player have to support progressive scan for that to work?

[jagabo]

Your HDTV will deinterlace (assuming it's progressive) as the DVD is played.

You could deinterlace before you encode, and you might even do a better job than your current HDTV. But the changes you make will be permanent. In the future, if better deinterlacers are developed for HDTVs, you won't be able to take advantage of them.