 -
11 years 8 months ago
- Posts: 441
|
...to DVD (dang subject line ran out)
Okay, I've had a number of people asking me lately how I get such lovely quality for my homebrew DVDs, and some just flat-out refusing to believe that it's possible to make an old VHS recording look better than it is (and wondering why I use 3-4 discs for a 13 episode series instead of cramming it all on to one disc). Well, I can assure you it IS possible (you just have to know what you're doing), and the reason for the number of discs is - QUALITY (which once again, you'd know if you know what you're doing).
So here's a rather lengthy blurb about just HOW I do it, and what techniques I use to get my old recordings looking and sounding their best for my DVD sets. Enjoy!
It depends on the recording, but I'll give you the basics:
The Wonderful World of TBC
The most important thing I do is use time-base correction when transferring videos. A time-base corrector (TBC for short) will lock down and stabilize the video signal, effectively reducing or eliminating generation loss. Even a first generation recording will lose its signal strength over time, since the magnetic particles on a tape drift over time. With a TBC, you can re-lock the signal, removing any "waviness" or "jumpiness" in the picture. This is especially important when transferring to digital, as digital wil take any errors in your original recording and often make them look even worse. But if you're capturing a properly TBC'ed signal, it can reduce or eliminate such problems. The end result is a clean, stable video signal ideal for capturing. This is why a lot of my DVDs actually look BETTER than the original recordings they came from. I cannot possibly stress enough how important this is. It makes SUCH a difference.
I use my Digital8 camcorder as a pass-through TBC, and it works beautifully (and is MUCH cheaper than getting a standalone TBC unit, whihc can be several grand for a good one). Transferring from VHS, I find the camcorder's built-in TBC does the job quite well. Most of today's digital camcorders have this feature, and you can usually find older models on ebay that also have that feature (and cost considerably less than a new model!)
DV and why you should use it
For capturing and editing, I work STRICTLY with DV format. It takes up a ton of disc space, but it is by FAR the most reliable format for video editing. Editing in MPEG can be a pain, since the computer has to RE-compress everything every time you make a change. This can actually REDUCE the overall quality as you go. DV doesn't compress anything, it's just a straight video stream. Which means you can edit and re-render all you want in DV and your umpteenth copy will look identical to your original master capture. Yes, DV is true lossless editing folks! MPEG ain't.
Additionally, don't be surprised if you run into some major audio sync issues if editing in MPEG. DV rarely, if ever, has these problems. MPEG should only ever be used as the ifnal delivery format, and never as an editing format (unless you're working with HD, which is an entirely different beast that I won't go into here...). SO - Once I have all my editing and restoration done, I then (and ONLY then) convert the final video to MPEG-2 for burning to DVD.
MPEG - why bigger is better
A good rule for DVD authoring is - if you're using single layer discs (which most people do because they're so affordable), don't put more than about 90-100 minutes of video per disc. Anything higher and you start to lose quality because of the higher compression required. This is why my sets never contain more than 3-4 episodes of a half hour show or 2 episodes of an hour show per disc.
The minimum bitrate I use is 6000kbps, although I often go as high as 8000 for disc with 3 episodes. My commercials compilation are all done at 8000-9800kbps. And I always use 720x480 resolution. Some people think it's okay to use 352x240 mode for VHS recordings. They're wrong. Yes, VHS has 240 lines of resolution, but that's 240 lines of ANALOG resolution. Analog and Digital are not the same. For the most accurate capture of the video, you MUST use 720x480, period.
Simple rule of thumb: Low bitrate = low quality. High bitrate = high quality.
Most consumer DVDs are encoded at around 6000kbps. High end titles (such as Criterion Collection or SuperBit titles) often go as high as 9500kbps. Basically, KNOW the rules of the format you're working with. Yes, with compression algorithms such as DIVX and XVID you can easily cram 13 half hour shows on a DVD, but despite what anyone tells you, neither of those formats are up to DVD standard for quality. You CAN cram that much on a DVD in mpeg, but only if you use 352x240 MPEG-1 (or VCD quality, in other words...gag!) Nobody who is truly serious about maintaining quality will EVER work with that format. If you're working with DVD, you're working with MPEG-2. And with MPEG-2, bigger is ALWAYS better.
For Pete's sake, stop clicking that DE-INTERLACE button!!!
NEVER de-interlace video unless you're only ever planning to watch it on a computer. All television (with the exception of a handful of HDTV broadcasts) is interlaced. If you're capturing from VHS, Laserdisc, DVD, or making recordings of your favorite video game, it's interlaced. Simple rule here: MATCH YOUR SOURCE. If your source is interlaced, encode interlaced.
Now that we have the technical stuff out of the way...
Here's what I do with certain types of recordings:
If it's a relatively clean, glitch-free hi-fi SP recording, then I'll just do a straight transfer - just a straight pass from the VCR, through the camcorder, then out to the computer via firewire. If it's a series, then I'll do some audio checking just to equalize the levels of all the episodes in the set. But for the most part, if it's a clean recording there isn't much work to be done.
If the original tape has some glitches, there are a few things you can do. The TBC will help to clean these up a bit (at the very least it will reduce or eliminate the jumping in the picture that often occurs with glitches), but with large glitches there's not much you can do.
If the image where the glitch occurs doesn't have a lot of motion, you can "cheat" by taking it into a decent editing program and replacing the bad frames with good frames in the same sequence (meaning you take a "clean" section from before or after the glitch and manually patch it in - you need to have the audio as a seperate audio track to do this properly, however, since you're replacing sections of the video). This doesn't work as well if the sequence contains much motion, though, as the replacement sections will be really obvious. I usually only go to the trouble of doing this if it's a major issue with a clip, otherwise I'll just let the glitch go.
EP/SLP vs SP...6 hours of garbage, or 2 hours of gold?
EP (aka SLP on some machines) recordings are MUCH worse for glitches than SP recordings. SP runs at 3 times the speed as EP, and thus can blow past the average tape glitch fairly quickly. Which also means an EP recording will take 3 times longer to pass by the glitch...which can be excruciatingly painful to watch. This is one of the many reasons very few of my tapes are EP (the other being the superior picture quality that SP gives you).
Plus, SOUND is a big issue with different speed recordings. It's not too bad with hi-fi audio, as it sounds pretty much the same for EP and SP. But if it's a lo-fi recording, then the difference in audio quality between SP and EP can be quite dramatic.
Audio - and why you should pay attention to it
Which brings me to the next thing I do - audio. If you really want to get heavily into the whole restoration thing, you almost ALWAYS want to do your audio and video seperately. Sometimes a recording may have great picture, but leave something to be desired in the audio department. That's when it's really useful to have a standalone audio editing program (I use Cool Edit Pro).
By working with the audio seperately, you can do things like boost the volume if it's too quiet, remove excessive hiss, and get into even fancier stuff like parametric equalization - which, if you know what you're doing, can take a muddy lo-fi EP audio track and make it sound comparable to hi-fi (not QUITE, of course, but pretty darn close). The trick when doing this is not to cut any section OUT of the audio, as this will screw up the sync when you put it back together with the video. The best way to seperate the audio in the first place is just to rip it straight out of your captured video. That way the timing is the same right off the bat. Then you do your cleaning/processing/equalizing without changing the timing, and you can just slap them back together without any sync problems.
My Cadillacs and Dinosaurs set is a perfect example. I had a beautiful recording of one of the old retail tapes, which had 3 episodes on it. Picture quality was as close to perfect as I could want. But - the audio was all mono. It's a stereo show. Fortunately, I had the same three episodes on tape from television in stereo. So I used the video from the retail tape, and used the audio from my other recordings. It meant having to manually sync up each act of the show to match, but the end result was unmistakeably superior to both of the original sources. In short, I took the best quality ELEMENTS from two different sources and put them together to provide the best quality end result.
Once you've got your audio sorted out, and your video is cleaned up as much as you can, then you put the two back together in your editing program (Speaking of which, I use Ulead Video Studio 10 for all my capturing and editing).
Now the short version!
And that's pretty much it in a nutshell! Every recording is different, so I don't always have to do everything described above, but it's nice to be able to do all that when required because the end results definately make it worth the effort.
So, to sum up, here's the short version:
1. Use some form of TBC when transferring video
2. Always capture and edit in DV
3. Always use 720x480 resolution (or 720x576 for PAL)
4. Never de-interlace your video. Just make sure you're using the correct field order (it's usually bottom field first for NTSC, upper field first for PAL)
5. Restore the video and audio seperately
6. Equalize the volume for all the episodes in a set so some aren't louder/quieter than others.
7. Never try to fit more than 90-100 minutes per DVD
8. Have fun and take pride in your work!
That last one is probably the most important. I put a lot of effort into my DVDs, and take great pride in my work. And I think anyone who encodes should do the same! 
- Sean
EDIT: If you're wondering about all the edits, I kept seeing typos as I was reading this over...so I fixed them.  |