HELP: VHS to VCD

[IndyJones007]

Hi everyone,

New to this forum. Have a question that must have been asked numerous times before, but here goes. As the subject pretty much says, I would like to know how to convert VHS home videos to VCD that could be played on standalone DVD players which will play CD-R or CD-RW VCDs(NTSC).

I have a Matrox Marvel G-200 TV that's installed in my computer and works great for the little things that I do. Now I'd like to unleash it's power in the home video editing arena. I hope there are users out there who own a Marvel G-200 or at least have upgraded from the G-200 to a higher end one and have experimented with creating VCDs. Could you rocket scientists guide this newbie into the process of creating VCDs with Marvel G-200. What's the right software for this card to capture in MJPEG and then to convert to VCD. There is no editing required, no adding of special effects or cutting frames, just plain capture and convert to mpg and then burn to CD to make the VCD. I have checked out some helpful websites suchs as vcdhelp.com and others but it seems there's a special better way to do this for the Marvel G200, using special software to overcome the 2Gig limit and other codecs for capturing in MJPEG and then using a 3rd party software for converting to MPEG. Most of this is new territory to me, so I hope you guys can help me. What settings would I have to input for capturing in MJPEG for making a NTSC VCD, and with what software, which is better than the PC-VCR provided by Matrox. What versions of the s/w would be better for the Marvel G-200. I am hearing a lot about Tmpeg for conversion to MPG, which version and what settings would be appropriate for VCD.

I would very much appreciate it if you guys can help me out on this, especially on going about using the Marvel G200 card for this purpose and all the specific s/w to go with it.

Thanks

IndyJones

[AndrewDV]

Hello,

If you need a primer on VCD or similar formats, check out:

http://www.vcdhelp.com

Just a warning, but no one is ever 100% satisified with their work. Users spend lots of money burning time and CDs whenever the slightest improvement can be made to existing footage. My take on it is:
-Set some basic goals
-Do the research
-Experiment to a point where you still enjoy it.
-Be happy with something that 'looks' good to the viewer.
-When you learn something new, use it in the next project. Keep moving forward.

This is just my take on it

Here is my process:
-Capture using PIC MJPEG (704x480 or 640x480)
-Edit in Ulead MSP6 (set deinterlacing and field flags)
-Master AVI to Hi8 (people still use tape)
-Convert AVI->MPEG using TMPGEnc
-Burn to VCD/SVCD using Ulead DVD Plugin
-Move on to the next project

[Brian Ellis]

Indy

You don't say where you are. As most VHS tape is of relatively poor quality, you may find that you cannot see the difference between capturing at 704 x NTSC 480 (or PAL 576) or 352 x 240 or 288. Try it and see. If so, you will be able to capture four times the length and stay within your allotted space, be it 2 Mb or 4 Mb.

You also don't mention your hardware. You may find it much better to capture in the Marvel native hardware MJPEG, which is remarkably good, at least to start with, if your resources are on the low side. You are less likely to experience dropped frames and other maladies. If that goes well, then you may wish to start experimenting.

However, you have a relatively poor signal source, do not expect miracles, such as broadcast quality VCDs.

Finally, many household DVDs do not accept home-burnt VCDs or SVCDs. I recently tried a VCD in 7 different DVD players at a shop: only one accepted it.

[AndrewDV]

Hello,

You can see VCD/SVCD (as well as X(S)VCD and miniDVD) player compatibility on the vcdhelp website. In general, standard VCDs burned on CDRWs give the most flexibility.

[funsoul]

Hi!

Yup...none of us who get anal-retentive about the subject are ever truly satisfied that I can tell.

My process for vcds is as follows:
- capture from vhs (using svhs connector) at 704x480 (yuy2), fullframes(uncompressed), 29.97fps, 44.1KHz [virtualdub]

- apply minimal filtering from virtualdub (mostly unsharp mask and precise bi-cubic resizing to 352x240 uncompressed or huffyuv)

- import all clips into premiere for editing

- frameserve to tsunami for conversion to mpeg

Some of my observations have been:
- final vcd always looks sharper when raw captures are larger framesize
- deinterlacing is not always necessary (is this just me?)
- improving encodings adds a long, long time

Yup....the quest for quality appears to be more or less a neverending battle.

Good luck!
-funsoul

[AndrewDV]

Hi funsoul,

For VCDs, I've also noticed deinterlacing is not necessary. On the SVCD side, I've noticed the opposite (that being deinterlacing is important). I think this is because the res of a VCD is 240 (1 field) instead of SVCD's 480 (2 fields).

Someday, someone might standardize the perfect.... wait... scratch that... There will always be a rule of thumb for video

[Karlson]

What I found out by try, email, and postings is similar. I tried MPEG encoding with Tsunami and used deinterlace odd to build a SVCD. Result looked well on the first sight.

When I started encoding with CCE SP, I noticed some ghost effects and used VirtualDub for deinterlacing. In the beginning I thought the image would be captured as a full frame and just split in even and odd lines. But in fact its captured with double the frame rate, so there might be slight movement between the two fields.

I learned then, that Tsunami deinterlace does nothing else but copying the odd lines and loosing the even. If you produce a VCD, you have no need to deinterlace, you simply use every other field. For SVCD, this is different. Using Tsunami deinterlace (odd) would mean here to double the odd fields and trow away every even field, resulting in some more steps at diagonal borders.

But in the end you see this only when you stop the movie. When it's moving, the effect is that the image is not as sharp as it used to be.

On the other hand, if you produce a SVCD with some deinterlacer (like smart-deinterlace in VirtualDub), it will try to combine the fields (when there is movement in the image) and thus not reduce the vertical resolution. You will notice a sharper, crispier image.

But again, you realy see this only when pausing the movie on your computer screen. On a typical "Aunt Annie" movie watched on your TV from 5 meters (excuse me: yards) distance, there is no big difference.

So for me the encoding process would be something like (PAL):

- capture 704x576 either HuffYuv or PIC Video 19, 18 also ok
- VirtualDub smart-deinterlace
- VirtualDub billinear resize 480x576
- no smoothing (is done through resize)
- eventually sharpen
- store compressed with HuffYuv or PIC Video 19, 18 also ok
- feed to Tsunami or CCE SP
- burn with Nero.