Thanks Scott,

I'm assuming DVDA3 can' t tell me where the I frames are....

So what I have to do is to keep moving the out point right a frame at a time (doing test burns) until I hit the next I frame, hoping that there is one before the end of the fade? Ouch.

I found the following on Doom9:

"MPEG2 compliant max. GOP length: PAL 15, NTSC 18 (frame based)"

Does that mean that if I set my outpoint 18 frames (just over 1/2 sec) to the right on the timeline, the jump will happen on the I frame after the one it's jumping on now (though that might take me into the title)?

I don't think DVDA3 tells you where the I-Frames are (I could be wrong), but it's a little more complicated, it has to be the first I-Frame (or the first GOP) in a VOBU.

Your idea of doing test burns is probably the only way to do it, unless you can extend the amount of time in black. If you 2 seconds of black, there's a good chance that by putting your in/out point at the center, you'll hit what you need.

Not sure what you found on Doom9, but it appears to be talking about the number of frames per GOP - the mainconcept encoder with Vegas usually does 15 frames per GOP. However, the encoder doesn't have to always put 15 frames per GOP, sometimes it will insert short GOPs.

One last thing, Scott. I had this problem (with the same project), and though I couldn't fix it because the fade was so tight with the title (I had overlapped the fade and the title (on the track above) on the timeline, so the title came in immediately after the screen had faded to an apparent black).

I rendered the project leaving a little space on the timeline (probably a handful of frames) between the end of the first clip and the start of the title generated media so I'd have a little more black space to work with. I wonder if an I frame is automatically inserted at the end of the first clip when Vegas 6 renders?

Scott is absolutely correct. This was one of the most aggravating things about DVDA until version 4 finally dealt with it. Not only does Vegas 7 let you encode so that "I" frames are created wherever you place markers (if you set the preference in V7 to do this), but if you put MPEG-2 files into DVDA4 that were encoded with an earlier version of Vegas or with another program altogether, as you move the marker or in/out points in DVDA, it will show you (via a change in the marker) whenever that marker is directly over an "I" frame.

As for why this happens in the first place, a DVD player can only start at one of these boundaries, so that's not the "fault" of Vegas or DVDA.

I don't have a workaround for your present situation, other than to encode longer sections of black between scenes (at least 0.5 seconds at every transition). This is what I did for the years before version DVDA4/V7 came out.

Thanks for replies, everyone.

Since I'm already in the soup, I wonder if the DVDA4 trial will let me create a disk? That would be the quickest fix.

Actually, in a pinch, I could open the project in DVDA4, find out where the Iframe is on the timeline, and the set the out point there in DVDA3 (it looks like when you zoom in to the max on the timeline, the smallest subdivision mark is one frame).

You have to have both parts - Vegas 7 and DVDA 4 - it takes Vegas 7 to tell the mainconcept encoder where to put I-frames and it takes DVDA 4 to correctly interpret things when building the DVD.

I don't remember off-hand how crippled the trial versions are; maybe someone else remembers.

--Scott

You have to have both parts - Vegas 7 and DVDA 4 ...

Well, yes and no. You definitely have to have V7 if you want the encode to be correct. However, with just DVDA4, you can at least see exactly where the DVD chapter stop is actually going to be (just move the chapter stop marker on the timeline until the exclamation mark disappears). That eliminates the surprises and often lets you create a usable disk.

In DVDA4, you can also zoom the timeline until you see increments of about "half second", then when you grab the chapter mark, you will see blue triangles on top of the timeline (look to the left and right of the current position) -- those blue triangles are I-Frame reference markers.

Usually authoring DVDs you'll create separate titles -- each segment in your sequences -- then in your case assemble the titles in playlists corresponding to your buttons. It doesn't have to be done that way, but it eliminates the problems you're having now.

In general Vegas & DVDA have a habit of not being really precise at the end of a title video -- usually want to have a full second or more of black at the end of a video included in your render, then it's fairly easy to adjust the out point *if or as* needed. If you're not doing a fade to black [i.e. with your sequences, & also for intros & transitions] render an extra frame or 2 of video only, then trim the endpoint. Doing this I've never had a problem.

You mention burning with DVDA4... With v 3 or 4 create a layout on hdd that can be tested, then when satisfied burn with Nero, Roxio, or whatever.

At any rate *for me* this is easiest, most foolproof [knock wood] method in DVDA and a few other apps. If you don't want to re-render your sequences, slice them in cuttermaran or similar, though I'd re-render the ac3.

[tip: with video made not using vegas dvda templates, or after demuxing etc. basically any time you have m2v, for older versions of DVDA run the mpg2 through a muxer like TMPGEnc Mpgtools without an audio track & all versions of DVDA will normally accept it.]

MPM - thanks!

I originally rendered the whole thing as one file because I was having the title come up right as the fade went to black. Now that I have to have a little full black for accommodation, there's no reason not to render separate segments - the jump won't be noticeable during playback of the full program....

there's no reason not to render separate segments

Many (most?) DVD players will not navigate backwards through titlesets (i.e., multiple, separate MPEG files) when you press the chapter back button on the remote. Thus, even if you set end actions and titleset order correctly (the later can only easily be set in DVDA4), your customer may not have the same navigation ease that they would with a single titleset.

Fortunately, some (if not all) of this downside can be overcome if you put your various assets into a "Music Compilation." The only downside of this approach is that you cannot add additional chapter stop points within a given asset -- only between assets.

>>Many (most?) DVD players will not navigate backwards through titlesets <<

Oh hell, now I'm back to fiddling with the black space between segments.

Oh hell, now I'm back to fiddling with the black space between segments.

Well at least try a test DVD with what you've got now before you go back and rearrange everything. While what I said is absolutely correct, it may or may not make any difference in your particular project and for your particular client. I sure wasn't trying to get you to spend unnecessary time on the project.

Thanks, John. This is a youth orchestra disk, so it's going to be playing on a bunch of different parents' players. I can live with the sloppy jump (the footage is all parent-shot, so it's not like the bad jump is the only glitch on the disk).

I really appreciate all the help you've given me on this.

Not disagreeing with anyone, thought it might be useful to post some of the *basics* concerning title-sets… Hopefully it’ll make it easier for folks to decide their DVD design, as well as being an [admittedly simplified] future reference.

I’ll start with 2 big disclaimers: There is no official instruction manual because the specs are Pay-Per-View, complete with NDA. And, any DVD authoring program can both limit your input, and have quirks in how it writes the underlying DVD structure. Because of this you’ll see what seem conflicting rules, advice, and/or instructions, plus your work is subject to app-specific limitations and work-a-rounds.

You can spot different DVD title-sets by looking at the rendered files: you’ll see an .IFO file, followed by one or more similarly named .VOB files [usually no larger than 1 gig each for player compatibility] for each title. Each title can be further broken up however, including having otherwise separate video programs or files combined.

If you lock yourself away, hide where the Knowledge Police won’t find you, retail DVDs can be taken apart – dissected to get an idea of what the average player expects & can handle. [I’ve read that] you’ll see some video that’s run together, like maybe all the individual cut scenes in one video file, despite their being treated like separate titles when you play the DVD. You also might see a lot of individual titles, and there are times one title is split across multiple title-sets. You’ll also see some *dummy* structures (menus & PGCs) usually directing program flow. [Of course while you *can* look at the innards of a retail DVD, you really, really shouldn’t because of some fat cats with armies of lawyers.]

So apparently regarding title-sets there really is no one *right* way to do it… In DVDA mpg2 video clips can be imported separately, treated as individual titles. They can also be run together, rendered as one longer mpg2 clip, with menu buttons pointing to sections of that larger file. Playlists are a special case, might even say a trick, where the player is told what to play, in what order, in the same way the player’s directed to play a single video file. [Before playlists a separate clip was rendered to include all the individual clips]. Video compilations & angles work in much the same way.

When a DVD layout is rendered, all the video is divided into cells. These cells do not have to be breaks in the video, like splitting a clip in Vegas – they just reference points or times – they are what chapter markers point to. Inserting a chapter marker means DVDA will create a cell at that point. During playback, a DVD player follows directions on which cell to play next – it can be the next cell in the same video file, or a cell in another mpg2 file or title. Within certain rules [which are beyond the scope of this post] you can program almost any order of play, & through scripting make the order variable, have it depend on user input for example.

The next question, the one I’ve been leading up to is: “How should I create my DVD”? Should I use separate titles or run everything together. If there’s a difference in encoding settings, audio channels, number of audio &/or sub tracks, BOV, angles etc. you want to go with separate titles – Otherwise it’s up to you. There are practical limits to each. At some point having everything in one file just becomes too unwieldy, too much to handle, and there’s extra work involved, especially should you ever have to edit your video. If everything’s done separately, the number of files you have to handle in a complicated project can grow by leaps and bounds, and it might impair playback, although you also might have more options with menus and programming. Vegas & DVDA have limitations which I’ll get to later.

You might encounter differences in how the DVD plays in many players – jumping from title to title can be erratic, though that might or might not be a concern... Personally I’ve never seen a lot of viewers jump titles, don’t think it’s a big deal, but that’s just my humble opinion. I have seen viewers get confused when any remote button operation is erratic, so I try to limit & control any possible viewer behaviors, including forcing them to use menus to select a title.

I’ve created & always have on hand a number of extremely short black mpg2 clips in both aspects. Using the example of an episodic DVD because it’s something everyone is familiar with, I import *in order* 1 episode file, 2 black clips, 1 episode file and so on… The end action of the episode video can be whatever you want, menu, next title, whatever. For its end action the first black clip jumps to the last [not end] chapter of the first episode. The end action for the 2nd black clip can either be a menu or the start of the second episode or whatever. Both black clips have their endpoints set to something like 12 - 15 frames duration, & all remote operations are turned off. None of the black clips are referenced or called by any menu or end action themselves.

This way I control what happens if someone repeatedly presses the Next button during the first episode – they don’t find themselves watching the second episode, which can be confusing, plus it’s a real pain if the Previous button on *their* player doesn’t do anything. Watching the second episode, I control where they go if repeatedly pressing the Previous button. This can of course be tailored to many situations, customized etc, & is limited by player functionality. At any rate, while the viewer can say they wished it worked some other way, I personally think that’s better then, much less serious than confusing the average person.

As for Vegas & DVDA and how they relate to titles: they have the habit of inserting a black frame (or what looks like a partial frame on the time line) at the end of an mpg2 clip, and you have to adjust the end point in DVDA if you don’t want to see it. They can also have trouble with the end of clips that fade to black – I put a second of black at the end in Vegas, but this needs to be muted video, not generated, for it to work every time. Adding a second or two of black at the end of a clip before render to mpg2, it’s easy to set the end point of that clip in DVDA. While that’s extra work, in practice it’s not always that easy getting frame accurate response using longer but partitioned clips in DVDA – maybe it should be, but…