Home Video record questions

[emann]

I have a G400TV and want to record home videos from 8mm tape to CD. Currently, I just want to get all of the footage onto the CDs as archive. What is the best way to do this?

Should I capture in MJPEG then convert to MPEG-2 or some other way. I want Quality to be as high as possible, keeping in mind that the source is std. 8mm.

Also, I am looking for something that can automatically split the recording into seperate avi files when the scene changes. (MGI VideoWave I am using can find the scene changes, but does not split into seperate AVI's.)

On that note, I have both MGI Videowaave & Ulead Visual Studio. Which is better for this application & for future editting.


Thanks for any help, I am just getting started in this.

[AndrewDV]

Hello,

If you are daring, you can utilize Huffyuv to capture the video. Then, use TMPGEnc to create VCD/SVCD files which can be put onto CD using something like Nero. That is a lot to take in, but do a search on this forum for more info.

Also, I just took a look at the Dazzle DVCII. Well, this is not a great product to do a lot of edits with. However, for copying video to SVCD/VCD/miniDVD with minor edits, it is hard to beat for the price.

[emann]

Thanks for the response.

Is it worth recording in Huffyuv if the source is 8mm analog?

Also, is there any value to saving as VCD if I don't have a DVD palyer, or should I just save the MPEG2 files on the CDS.

PS - Where can I get Nero?

[emann]

One more question:

What's different between the Dazzle you recommend and using the G400TV?

Thanks again.

[AndrewDV]

The Dazzle DVCII captures MPEG2 directly while the Marvel captures either YUY2 (with the registry hack), RGB, or MJPEG which has to be transcoded into MPEG2. Where that comes into play is the time/steps it takes to create VCD,SVCD, or miniDVD compliant MPEG2 files. Since the DVCII does the capture in MPEG2 realtime, it is very quick to get a CD copy of existing material.

The key item to remember is that you are limited to editors with the DVCII (MovieStar only). With the standard editor, you can only do basic clipping, sound, and transitions. Also, the package comes with DVDit LE which allows the creation of miniDVDs (PC playback only).

As far as Huffyuv, it does a very good job preparing AVIs to be sent to TMPGEnc. It is a bit of pain to work with some times.

As far as just copying videos to CDs, remember that a disc drive can only spin at certain speeds. Files just can not be copied over to a disc and be expected to maintain playback. That's where the VCD/SVCD/miniDVD "standards" come into play.

You can find Nero at:
http://www.ahead.de

[Dr Mordrid]

Sorry...wrong.

First off YUY2 is a flavor of YUV, the standard video mode for broadcast and analog video (read TV tuners, non-DV camcorders etc.). HuffYUV just does a lossless compression of this YUV source.

Lossless=no degredation.

ANY video signal has to be decoded to YUV before an MPEG encoder, hardware or software, can utilize it.

MPEG encoders work by encoding YUV and if fed RGB or some other format will extract YUV from it before starting the process.

This is why feeding an MPEG encoder a species of YUV (HuffYUV, for example) delivers such good results: you're feeding it a pure source of the signal it wants to see.

Also:

There is no inexpensive realtime MPEG encoder, software or hardware, that can meet or exceed the quality of a good offline encoder like TMPGEnc, CinemaCraft or even Panasonic (given it does MPEG-1 only) working with a good YUV source.

That extra encoding time is used wisely and goes towards noise reduction, motion search accuracy and many other very important functions of MPEG encoding.

Dr. Mordrid



[This message has been edited by Dr Mordrid (edited 28 March 2001).]

[This_Idiot]

Hi Emann,

Seems to me that you are asking several questions in one. I'll try and split down into what I think are the important components.

1) Archive. You mention that your prime concern is currently to archive your 8mm footage. It's very important to determine WHY you are archiving, as the format you select will dictate what you can use the archived footage for in the future. For instance, if you are archiving "finished" projects, MPG2 will probably be ideal both for playing back to TV or VCR, or for burning a variety of readable CD formats. If you want to edit "raw" footage then you'll want to keep it either in the same format in which it was captured, or in another un-recompressed format that can be edited (this probably precludes main MPG-1 or MPG-2 formats). On the other hand, beware of archiving in a proprietary format that may not be available to you in a year or two when you've changed your hardware. I'm thinking here of the RR-S footage that I captured 3 years ago. Will I still be able to edit it ?

2) Format.8mm isn't the highest quality source available, so while it's necessary to maintain the maximum available quality, it's a waste of time trying to capture at too high a frame size (since the resolution would nominally greater than the source material). On the other hand, dropping the compression ratio to maintain as high as possible quality per frame size would be desirable. (for instance, you'd be far better off with 1/4 frame at 2.2:1 than full-frame at 7:1)

3) Scene splitting. Like you, I've only come across this within VideoStudio, and to be honest it's a feature that I really don't like. I was damned glad when I found out how to switch it off. I can see that it would be a useful tool if you like to capture long amounts. I think that a tool to do this has been mentioned on the forum in the last week or so, possibly called scenealyser or similar ?

Don't get freaked out by all the tech-talk, and definitely keep thinking about your aims and refining your requirements, and posting here. The more stuff you try and the more specific your questions, the more you're going to get from the members

[Dr Mordrid]

I haven't tried using VS5's scene detection yet, but with some capture utilities of this type you can set a mark-in and mark-out time and then capture within those limits. If you're starting from the beginning of the tape just a mark-out will do.

As for encoding for decent quality and archiving 8mm video on CD I use TMPEGnc set up thusly;

MPEG-1, 2-pass VBR, max bitrate: ~1800, avg bitrate: ~1550, min bitrate: ~600 @352x240.

Dr. Mordrid

[AndrewDV]

Sorry if what I posted was confusing. I was trying to keep it on the simple side.

The bit of experimentation I had with the DVCII worked instantly to the 'naked' eye. It is nice that someone can just plug in the source and record directly into SVCD (or other) format.

[Dr Mordrid]

With the latest patch VideoStudio5 can do that now as well. Just set it up to capture MPEG-2 on a YUY2 enabled card, set it up for 480x480, your bitrate of choice etc. and you're off to the races.

Edit & feed the result to an authoring proggie (Nero 5 (no menus), Nero 5.5 (menus, whenever it comes out) or the Ulead GoDVD plugin when it's released etc.) and you're in SVCD land.

Ulead has announced GoDVD (at the least) to be coming for MSPro as well.

Dr. Mordrid


[This message has been edited by Dr Mordrid (edited 29 March 2001).]

[AndrewDV]

Sorry to go off topic here.

Is there any way in VS5 to avoid re-encoding after doing some trims on the beginning/ending of the capture? What MovieStar does somehow (the ONLY thing it does that I like) is allow the user to do simple edits and save the work without encoding. That would be a really help. The only other item is that I would not mind having high bitrates (>6Mbs).

Thank you

[DAR]

Hey Doc,

Regarding your setup for archiving standard 8mm video - how would you rate the quality of archived material? Is there any noticeable loss when you compare the aechived video to the source material?

[Dr Mordrid]

Whenever you go to quarter frame vs. full frame you lose something, but the tradeoff is that you can store about 50 minutes of video on one 80 minute CD. This reduces the playback quality on a large TV, but not much on a smaller one.

For the frame size the VBR MPGE's I produce with those settings look pretty darned good. MUCH better than VCD formatted MPEG's.

As far as "re-recording", more properly known as recompression, on trimmed clips goes this can be avoided by using an editor capable of "smart rendering".

Smart Rendering is where the editor only recompresses that portion of the video that has an effect, filter, title or such applied. It does not recompress video or portions of a project that has had a simple cut done to it.

VideoStudio and MediaStudio Pro both have SmartRender features.

Dr. Mordrid


[This message has been edited by Dr Mordrid (edited 29 March 2001).]

[emann]

Thanks very much for ALL responses.

I played with both AVI_IO and Virtual Dub last night, much more impressive for capture than Videowave or Visual Studio. I prefered Virtual Dub, just for its easier interface, but will keep playing with both.

As for what I WANTED to do, I wanted to import 8mm to archive to CD for future use in an editor. I would import the whole tape, and break it into scenes so it would fit on CDs. So it sounds like I would want to save to CD with as little compression as possible, which of course means I'll be burning a lot of CDs.

I was capturing in Matrox MJPeg at Full screen, 29.97, Least compression just to start, now I need to experiment to get the file size down. It sounds as if I'm better off going 1/2 screen or 1/4 screen instead of compressing more right?

How much worse off am I if I save to MPEG2 to get the size down. When I go to edit later, will the quality be real bad? Currently these are just home movies of kids, etc. Not any real high-quality shooting.

Also, I was worried about archiving to MJPeg since I would need my card to be alive in the future to play them back. Would the same be true for HuffYUV? I haven't played with that yet because I need to re-install Video Tools 1.52 (I have 1.54 now). Does HuffYUV have options(fps, compression, screen size? Probably not since its lossless?)

Again, thanks for all responses.

[Dr Mordrid]

As long as you aren't going to get too heavy into adding graphics half screen (352x480) is just fine. This is known as "natural" video. This is because the color information is spread over 2 horizontal pixels, so you only have 352 dintinct color values in a full size frame anyhow. It's the grayscale data that you lose.

When played back through the Marvel using the DualHead/DVDMax feature this 352x480 format will be resized on the fly to full screen. Most folks can't tell the difference unless detailed graphis are in the image.

Typically I capture natural video at 352x480 using the YUY2/HuffYUV technique you may have read about in this forum. This way we get uncompressed very high quality source files that can later be encoded into VERY high quality MPEG-2 video using TMPGEnc.

I use both this and the MPEG-1 technique, depending on what I plan on using them for. If they are destined for VideoCD then they get stored as MPEG-1. If for SuperVCD or DVD/miniDVD then they get archived as MPEG-2.

As for which capture proggie to use: I prefer AVI_IO....heaviliy. Reasons;

1. if a frame drop occurs it fills in the gap with the previous frame. This keeps the frame count correct and thereby maintains audio/video synch regardless if there have been a few drops.

2. it captures segmented files = multiple 2 gig files for a long capture. It can span these captures over more than one drive for up to 100 total files. These files are named with a numeric tag thusly:

capture.00.avi .... up to capture.99.avi.

These segmented files can then be "frameserved" into TMPGEnc from the disk using a free scripting utility called AVISynth. Great program.

Once they are being frameserved into TMPGEnc the whole bunch can be encoded into a single MPEG of whatever format you desire.

VirtualDUB can also do segmented files, but I've never been able to get it working as well as AVI_IO for this. AVI_IO is less likely to drop a frame at the seam between two files.

3. it's smaller than VirtualDUB and thus has a smaller memory footprint.

As for HuffYUV it's a Windows codec that runs under Win9x, ME and Win2K and will work in any editor capable of seeing such codecs (most all). HuffYUV files are more for use as a lossless capture & editing format than storage or playback. I encode them straight to MPEG most of the time.

HuffYUV only requires that the frame size be a multiple of 4. Otherwise it takes whatever depth, fps & such you feed it. It does have quality settings, but most moderately fast systems can use its highest quality setting without a problem.

Dr. Mordrid


[This message has been edited by Dr Mordrid (edited 29 March 2001).]