Fitting 2 hours on DVD??

cgoose23 wrote on 2/13/2007, 10:47 AM
I am creating my Movie in Vegas Movie Maker, and have 1 hr 31min so far edited, When I went to render it and burn off a copy to watch my progress it wont go, because it is up to 5.7 instead of the alotted 4.7 on standard DVD.

On previous editing systems I have used it would decrease the quality to fit a full 2 hours on the DVD..just needing a little help on how to do this?

Thanks.

Comments

johnmeyer wrote on 2/13/2007, 11:02 AM
Use a Bitrate Calculator to determine the average bitrate to use for the encoding process. The lower the bitrate, the more you can fit on the DVD, but the lower the quality.

You can get full, no-compromise quality on a single layer DVD of up to about 75 minutes of material. Beyond that point, you have to decrease the bitrate from the maximum (there is a maximum allowed by the DVD spec). Until you get to about 90 minutes (at least that's my estimate), you won't see much degradation. Beyond that (i.e., as you use lower bitrates to fit more on a single disk), you may start to see some degradation from your original, although this is highly dependant on the nature and quality of your original. I have actually been able to fit close to three hours of low-quality VHS material on one single layer disk without much degradation. Also, I recently received a set of three disks that contains twenty-one episodes of a one hour show. That's seven one hour (actually 48 minutes, without commercials) on a disk. That's about 5.5 hours. It was encoded at 320x480, and it is from 24p original material (which helps a LOT). The quality is pretty bad on the fast motion scenes, but remarkably good on most of the material. Quite an achievement.
teaktart wrote on 2/13/2007, 11:18 AM
" It was encoded at 320x480, and it is from 24p original material (which helps a LOT)"

I've had the same dilemma about fitting longer projects onto one disc.
With the encoding you mention above do you view it in full screen when played on a set top dvd player?
Also, if the original footage is 30fps does it make sense to render to 24fps in .avi format and then final render to 24p dvd to save data space?

Thanks,
Teaktart
johnmeyer wrote on 2/13/2007, 11:34 AM
With the encoding you mention above do you view it in full screen when played on a set top dvd player?

All my DVD projects are designed to be viewed on a standard, interlaced TV set (i.e., the NTSC TV we've come to know and love). In my opinion, you MUST view a project on a TV set in order to be able to correctly judge what you've done (color, temporal problems, etc.)

Also, if the original footage is 30fps does it make sense to render to 24fps in .avi format and then final render to 24p dvd to save data space?

Definitely NO. There is a large misconception that Progressive is somehow "better" than Interlaced. That is not true at all. Also there is some misconception that 24p is better than 30 fps. Also not true. If your original footage is shot at 29.97 fps interlaced, and you plan to show it on a TV set, then unless you want to achieve what amounts to a "special effect," do not alter the framerate and do not attempt to deinterlace the footage. Both operations degrade the footage (you throw out information and have to interpolate to create the progressive fields).

Even if you have a modern "HD" TV that is capable of displaying progressive footage, you still will make the video look worse if you deinterlace and change framerate. Now, if you want the specific "look" of 24p, then by all means go that route, but if you want the video to look its best, don't deinterlace and don't change frame rate.

As for encoding, the reason you can go to lower bitrates is that 24p has only 40% of the temporal information contained in 30i (i.e., 60 fields per second compared with 24 frames per second). The tradeoffs between the lower spatial content of a single interlaced field compared with a full progressive frame, when it comes to encoding, favor the progressive version considerably, so you will get a better encode (less "mosquito" noise and other low-bitrate artifacts) from material that is truly progressive and which is only 24 fps.
mikkie wrote on 2/13/2007, 11:41 AM
Squeezing video, check the videohelp, doom9.org & similar sites for a lot of tips and info...

You've got the various Shrink type programs, which atlower compression levels at least, discard data from the b frames -- think of a ball rolling across the screen... this is the frame data that fills in where the ball was, and so is often not so critical.

The correct way is to add more compression, drop both the average & peak bit rates, but as John mentioned, reduces quality -- eventually it just gets too low. Smaller fame size is the logical next step -- 704, 320 & SVCD's 480 can work for you. See:
http://www.thedigitalbits.com/officialfaq.html

480 is interesting because it was derived from measuring how much of the signal actually reached the average TV after losses in cabling. As such it works Very well, *IF* your player will handle it on DVD. If not there are various tricks that *might* work you can research, but it's VERY unofficial & iffy.

All it takes to fill a TV screen is 320 x 240 if NTSC. It will not look as good as 720 or 480, but in a pinch will usually work, and some folks swear by it.

Reducing fps does work, but is a bit controversial sometimes. On original digitized film the fps is 24, & in NTSC land pulldown is added to achieve 29.976 [Vegas docs have decent explanation]. Problem is removing the extra frames however they got there. ;?P

Ripped mpg2 with pulldown can be processed without [pulldown] using DGIndex. Captured vid can get to 23.976 fps using IVT, which throws out *roughly* 6 frames a second. You can also *sometimes* use IVT routines to throw out 6 fps on non telecined video, video which hasn't had 6 fps added to the original 24. IF it works the benefit is the remaining frames might be cleaner than an averaged frame created in an NLE. The downside: motion will not be as smooth, and it's never going to be as good as 24p shot originally.

DivX on CD is also an increasingly viable option for the players that support it, but no menus or anything normally allowed.

[Edit] Reading John's post done whilst writing this, perhaps another (I think complementary) take on fps.

When you've got 24p film, you need to add ~6 fps to get the NTSC legal signal your TV wants. Telecining adds frames/fields by interpolating existing data, & generally looks bad to the purist. DVDs use pulldown, which is a routine of showing fields more than once (again see the docs for diagrams). In the case of 24 fps source then, why add actual frames? Much of what you'll read about 24 fps comes from people removing the pulldown to get back to the original 24, so that they can add pulldown again for DVD, or leave it at 24 for mp4 and similar. Obviously the need for this sort of thing is more evident when you don't hold the (c)...

Reducing fps is both popular and sometimes necessary for web formats -- there may not be any alternative to allowing the bandwidth a frame needs to appear as something more than junk, & flv requires 15 I believe. For TV it's a bad thing, but if someone's really, really desperate, and won't shell out the $ for DL, it's an alternative to even higher compression, but smoothness of motion will suffer -- then again I've seen video with almost no motion so...
teaktart wrote on 2/13/2007, 11:47 AM
Thanks John,
Great explanation....

I think I've had my process completely upside-down...!

I thought the way to get rid of, or at least reduce, jagged edges on thin lines was to "de-interlace" my 29.97 footage.
In the case of my HDV footage which I capture using ConnectHD I've always checked (yes) for the "deinterlace" option assuming this would smooth out those jaggy edges....wrong?

MY confusion is that I seem to be "deinterlacing" at capture, then again with all the clips on the timeline (using Vegas script), and then there is the option to "reduce interlace flicker" when rendering for DVDA.....
I usually did the first two then hesitated when rendering for DVDA as I saw that it would re-render my mpegs once again if that was checked. So I stopped checking off that last option.

Am I correct in saying that you would not recommend any of these "de-interlace" operations? Only on still images but not for video clips?


Seriously confused here...!

Thanks again,
Eileen
mikkie wrote on 2/13/2007, 12:11 PM
If it helps at all...

"...reduce, jagged edges on thin lines..."

A lot IMHO depends on the source of the jaggies... Is it the video format, the camera or source, the basic pixel nature of digital vid?

With interlace: What happens is 1/2 the lines are recorded, then the 2nd, and motion occurs constantly as the fields are set down. Put the fields back together, and neither reflects the same info. On an interlaced monitor this is cool, because only 1 field is refreshed at a time -- you're seeing ~60 fields NTSC of motion. On a progressive monitor it looks off when you focus on a single frame, but players &/or graphics cards know to deinterlace on playback.

Now deinterlacing usually does some sort of blend between fields, which may not do more than slightly blur the differences, at worst adding a ghost image during motion, and might make jaggies actually worse. If your purpose is to eliminate them, view a field at a time to see if the problem is indeed interlacing. If it is, then experiment with various methods of deinterlace (Vegas has a couple) to see if you get an improvement -- bearing in mind all you may be doing is making an individual frame look pretty -- not necessarily improving the final video.

Generally only want to deinterlace once (if at all), and not all deinterlacing routines are equal.
cgoose23 wrote on 2/13/2007, 12:16 PM
Thanks guys. Now when I change the bit rate is that while in my project? in properties or something?

Thanks
BrianStanding wrote on 2/13/2007, 12:19 PM
Another option, if you have a recent burner, is simply to burn to a DVD+R dual-layer disk. Most DVD players read these just fine. You can go up to 8GB on these.

Media's pricey, though.
mikkie wrote on 2/13/2007, 12:20 PM
In the custom dialog for the mpg2 codec or renderer.
johnmeyer wrote on 2/13/2007, 12:32 PM
Here are links to a few posts that are related to the various things we are talking about here:

Fit Video to DVD

6 Hours on a SL DVD - why stop at 6?

Note that in the second post I mis-state that there is a bug in Vegas because it cannot encode to DVD at 320x240. At that time I was under the mistaken impression that this is a legal DVD format. It is, but not for MPEG-2 content. The lowest you can go, and still be legal is 320x480, something I have done quite a number of times, with great success. AFIK, all players should be able to play this, since it is part of the DVD spec.

As far as DVD Shrink goes, it is a wonderful tool to get you out of a jam when you've just done a two-day encode only to find that the resulting files are just barely over the limit. However, I sure wouldn't recommend that you use it as a standard part of your workflow, something that has not been suggested here, but has been suggested in other posts.

If you are curious as to how DVD Shrink works, here is an explanation, direct from the author himself:

DVD Shrink Explained


Maverick wrote on 2/13/2007, 2:30 PM
Thnaks John, some excellent reading there which has helped my understanding a little more.

Something that puzzles me, though:

With the MainConcept MPEG-2 set for rendering in V5 I could not see any difference in settings in custom between DVD PAL, DVD PAL Separate streams and DVD Pal video stream except if audio is included or not.

Also, whether I select Best or Draft I could not find any further differences in the settings. So, what excatly changes to make better MPEG2 files with BEST set rather than DRAFT?

Cheers
johnmeyer wrote on 2/13/2007, 3:01 PM
could not see any difference in settings in custom between DVD PAL, DVD PAL Separate streams and DVD Pal video stream

Many of these templates are similar except, as you noted, how the audio is encoded.

So, what exactly changes to make better MPEG2 files with BEST set rather than DRAFT?

If you are talking about the "Video Rendering Quality" setting in the Custom Project tab, then "Best" is an unfortunate, and rather poor choice to describe what is going on. In many cases, Good and Best will produce virtually the same quality. The only instance in which you need to use best (which takes 2-3 times longer to encode) is when you are downsampling, such as when you use high res still photos, or when you are taking HDV video and creating an SD DVD.

teaktart wrote on 2/13/2007, 4:35 PM
"...reduce, jagged edges on thin lines..."

On an SD project I have a band playing and there are some white wires in the background (against dark green foliage) that are all jagged on the dvd. I have the same jagged edges on one fellows' very red shirt which stands out next to the dark green foliage in a very obvious way.

What would be a good way to smooth those out, if possible, without 'softening' the rest of the image?

This video is about 1 hr 37min so I was also trying to jam the whole performance onto one dvd and I managed to squeeze it in just at the 4.7 GB limit. So there really wasn't any wiggle room to tweek the bit rate/quality settings....

What would you suggest in such a case?

And, Thanks again for you help.
GeorgeW wrote on 2/13/2007, 6:23 PM
<<<
Note that in the second post I mis-state that there is a bug in Vegas because it cannot encode to DVD at 320x240. At that time I was under the mistaken impression that this is a legal DVD format. It is, but not for MPEG-2 content. The lowest you can go, and still be legal is 320x480, something I have done quite a number of times, with great success. AFIK, all players should be able to play this, since it is part of the DVD spec.
>>>

I think instead of 320, you mean 352 -- as in 352x240 or 352x480.

352x240 mpeg-1 can look "decent" (depending on the source quality, subject matter, and the encoding method -- I've seen some VCD's that look amazing -- you'd be able to fit ~6-7 hours of VCD on a DVD5 disc -- after upsampling the audio to 48khz). The problem is DVDA does not allow dvd-compliant mpeg-1 video (i.e. it always wants to re-render a dvd-compliant mpeg-1 video stream).
rs170a wrote on 2/13/2007, 6:32 PM
GeorgeW, you're correct in that 352x240 or 352x480 are valid sizes for DVDA.
And, as you also mentioned, it does have to be MPEG-2 format to avoid recompression.
This info is covered in the "Getting Started" chapter of the DVDA manual.

Mike
johnmeyer wrote on 2/13/2007, 6:34 PM
352x240 or 352x480 are valid sizes for DVDA.

You are both correct. Sorry about that.
mikkie wrote on 2/13/2007, 7:17 PM
RE: mpg2 from Vegas... The biggest difference between the DVDA templates & the DVD streams is that the templates produce roughly the same file as if you took regular m2v & muxed it without audio -- in fact that's how you get m2v into older DVDA.

I'm sure somebody will suggest better, but for the jaggies against green background I think I'd try playing with masking, colors, and *maybe* some localized blur - the foliage has gotta make it rough I'd think. The masking to prevent altering the majority of the frame, the colors to maybe reduce saturation & anything else I could come up with ;?P

Might also try one of the logo filters on the lines... Might try some of the edge filters with masking... Might try a png with transparency, with a feathered area sort of tight around the lines filled with lightened foliage, and blend set to darken or something like that so it only hit the lines. Might blend a copy of the vid to get rid of the jaggies and try to composite just the bad edges on the original...? Trying to stimulate ideas anyway -- sorry nothing specific -- hopefully someone that does less flying by seat of pants will have a real answer. I just take a short section and start playing till it goes away. ;?}

With reduced frames on DVD one thing to please remember... A LOT of times a player will work with VCD & SVCD sizes based on it's internal capability with VCDs and SVCDs -- but a lot of them won't look for or recognize it on DVD, some will only do VCD, some won't do either. And apologies to George -- it's been so long since I've done VCD or otherwise played with 1/4 size, I got unforgivably sloppy... Very correct it's 352 width.... I hang my head in shame. <;?{

Thankfully I haven't run into any problems rendering 320 x 240 mpg2 in Vegas 4 & 7, the 2 versions I use and have installed. Usually it's the DVD authoring app that has problems, and it can be necessary to change the size flags before and after. For those looking for size info (am sure it used to be on the FAQ I linked earlier): http://www.videohelp.com/dvd.

For 1/4 size mpg1 & DVDA, I suspect the only way would be to insert the mpg1 into the rendered layout, but it's been a looong time since I kept up on the various tricks for VCD & SVCD on DVD disc. Probably use VobBlanker -- pretty nice tool for replacing streams.

craftech wrote on 2/14/2007, 4:36 AM
Every time we get into this discussion I always recommend the same thing. Two DVDs

I never could see the obsession with trying to cram what should be on a dual layer DVD on a single layer DVD just to have one disc instead of two. It would be different if most clients cared one way or the other, but most don't.

In terms of the artwork (if any) I simply leave the artwork off the two discs in favor of a very nice jacket insert.

In terms of the cost and compatibility, two single layer discs are cheaper than one dual layer disc and infinitely more compatible.

John
mikkie wrote on 2/14/2007, 8:50 AM
"...I always recommend the same thing. Two DVDs"

True enuf ;?)

Or... as we're editors... make it not 2 hours long. ;?P
Snip here, tug there, squeeze a bit... know it sounds like vanity insisting we can still fit into old jeans - sorry 'bout that - but it's generally pretty easy to cut a few minutes, especially as good as Vegas is at time compression [applied of course here and there judicially], & with mpg2, that can mean all the difference.
johnmeyer wrote on 2/14/2007, 9:29 AM
"...I always recommend the same thing. Two DVDs"

I've also recommended that in the past, but let's not lose sight of why it is sometimes desirable to cram a little extra onto a disc -- albeit with a quality compromise:

1. Production time and costs. If you have to make 100 copies, the cost goes up with more discs. Yes, I know that discs only cost $0.50 each, but unless you don't package them or add artwork, and unless you don't pay anyone to feed the discs into the machine, into the printer, assemble into the case, print the insert, etc., there are other costs, many of them proportional to the number of discs. Also, mailing costs go up (I spent about $300 in mailing costs for a big project last fall where I was mailing 4-disc sets to 80 people).

2. User perception. The laserdisc didn't catch on for many reasons, but one was the need to turn the disc over, or to have multiple discs to view one movie. People don't like to interrupt their viewing experience.

3. Branching. For training discs, if you want to have lots of branches, and some of those branches lead to other discs, the whole experience rapidly breaks down and is useless.

4. Storage. I now have a pretty massive collection of DVDs, both those I've made and those I've purchased. I find myself a little more reluctant, each time I consider another boxed set of 30 DVDs, to make the purchase and fill up another corner of my closet. I suspect I am not alone with this problem and the ambivalence towards multiple-disc sets that it creates.

So, these are some of the reasons, both for me and for my clients, as to why I sometimes (but not always) work a little harder at trying to squeeze everything on one disc.

Stuart Robinson wrote on 2/14/2007, 9:39 AM
The simplest solution was posted by BrianStanding earlier in the thread; burn to DVD+R DL.
johnmeyer wrote on 2/14/2007, 9:43 AM
The simplest solution was posted by BrianStanding earlier in the thread; burn to DVD+R DL.

Except that player compatibility isn't as good as with single layer. I've yet to have a return (thousands of different customers) using DVD-R single layer.
craftech wrote on 2/14/2007, 9:59 AM
In terms of time, burning twice as many single layer discs as dual layer discs doesn't take that much time if you use a home built DVD duplicator. Mine does eight at a time.

What takes a long time is printing designs on the discs!
It's incredibly time consuming and expensive. By not printing artwork on the back of the two discs in favor of printing an attractive jacket insert instead the time savings is incredible. A simple neatly hand printed numeral 1 and 2 inside the hubs or something similar on the discs and an attractive jacket and the whole thing costs less, has great quality, and is done in a relative jiffy.

John
Paul_Holmes wrote on 2/14/2007, 10:01 AM
Now that the price of dual layer has come down (about 2.50 per disk), I love the fact I don't have to compromise anymore. I did some research and found that Verbatim was the only dual-layer disc recommended. Nero was not recommended for burning since, although you can set the book type to DVD-ROM (an important step for compatibility with older players) it doesn't allow you to decide where the layer break will be. Instead I use ImgBurn. The dual layer works perfectly with my newer player (of course), but also with my brother's 4 year old player and a friend's 3 year old player. I will be testing it on my parents 4 year old player this afternoon. So far it's looking good.

I tried putting 2 hours on a single layer and all I can say was that the quality was barely acceptable. In static scenes it still looked pretty good but once there was camera movement there were a lot of distracting artifacts. The dual-layer looks pristine.