my brain hurts - converting SSA to something DVDAuthorGUI can understand?

[EddyH]

And I mean, properly. So that the results I got with using TextSub in AviSynth to make a DivX of my production - which worked just peachy and as-expected about five hours ago now - will be replicated on the DVD.

Didn't think it'd be that hard, but I'm now stuck here at work, quarter to 10 in at night, trying to get the damn thing to work. I've downloaded all manner of tools, half of which don't seem to work very well and haven't got anywhere very fast.

Sorry if this is answered elsewhere but as you may expect my mind is a bit fried (not the first late I've done pulling this proj together) and I have at least done a cursory search.

* First result I got, just dropping the SSA into DVDAuthorGUI --- it subbed the DVD, but ignored ALL of the formatting. All of it. Font was wrong, size was wrong, positioning was largely wrong, colours were wrong, and it rendered it in a very ugly way. Usable but NOT acceptable.
* Tried to use subtitle creator. It doesn't understand SSAs. (Turns out it doesn't understand the SUB + BMPs that MaestroSBT or whatever its name is spits out either.)
* Tried Maestro. Seemed very broken. Couldn't even load a file.
* Tried easySUP. Seemed broken. Abandoned it also.
* Tried SSA itself. Did a bit of tweaking. Resaved the file under a new name and had a peek to see if anything changed. Oddly, all the hex colours were now decimal. And wonders, it suddenly worked in Maestro. Seems TextSub tells you to use an incompatible colour speccing method. Joy.
* Used Maestro for an age, getting the things to work. Many broken things in that re: positioning and font size, but it seems to have barfed up something seemingly correct at the end of it (after about an hour repositioning stuff and scratching my head at the 1.625x / 0.6x / 2.0x sizing factors vs the actual pixels).
* But, DVD-A-G didn't like the taste of that, after best part of ten minutes chewing on it.
* Went back to easySUP. It ran thru the file, spat out a sub+idx instead.
DVD-A-G has JUST decided it doesn't like that, either. Augh.
(final message in log:
" INFO: Locale=C
INFO: Converting filenames to US-ASCII
INFO: Detected subtitle file format: subviewer 2.0
INFO: Opened iconv descriptor. *UTF-8* *ISO-8859-1*
INFO: Read 0 subtitles
INFO: Unicode font: 1418 glyphs.
ERR: Couldn't load file C:\decant recording\maestrosubbing\srt-45080922.eng.sub.

Process aborted due to above error
There was an unknown error. Please review the log for more information.")


So I'ma now letting it run thru my original SSA and see what it does. I suspect the same as above.
EDIT: Yup.

Maybe Maestro has a different output option somewhere?
EDIT: Trying Philips SUB + compressed TIFF... mmph.
ReEdit: No, that's broken as well.
Got a couple other programs to try but I'm having trouble getting the installers to run. May just leave it til tomorrow now cuz its like 10.15... hungry, thirsty, tired, p'd off, unable to understand why it just. won't. work.

Preferably i'd just like something that will take SSAs, or some intermediary format that preserves the formatting*, and spits out a DVD-A-G-friendly SUP at the end of it. Alternatively, some way of crowbaring them in after I make up a dummy disc image that has the option to turn them on and off but doesn't do it?
*(I have used: font, size, position, colour, time - that's it. There are some cases of overlap with two differently coloured subs onscreen at once so I'm prepared to lose antialiasing, but I do have six different colours. This can be reduced to three quite easily if needs be).

If this doesn't work, I think I'll have to just encode a whole extra copy of the MPG2 with the subs hard-coded onto it. Conveniently it clocks in at just over 2gb, so it'll fit...

Thaaaaaaaaaaaaaanks!

(so far I've conquered: a/ doing the filming (3 cameras, 2 extra mics, and some slides & titles), b/ finding out that the paid-for software we have here is a crock, c/ relearning all what i'd forgot in avisynth and a HELL of a lot more, d/ learning video cutting, audio mixing and syncing, e/ making menus and switching subs on and off, f/ making the actual sub file and applying it to a divx.... it's just this final g/ that's giving me a headache now!)

[manono]

When using DVDAuthorGUI I just use it to make the menus, with a short 'dummy' video divided into the final number of chapters I'm going to have. The real authoring goes on in Muxman using subs in SST format produced by MaestroSBT. They're 3 color (outline, alias, main color) and turn out fine. Then later on I replace the 'dummy' video with my Muxman authored one using the 'Replace' button in VobBlanker. As long as the number of chapters in both is the same, it works fine. If the original one doesn't have subs, they can later and easily be 'turned on' using PGCEdit. I also use PGCEdit (although DVDSubEdit can do the job also) to set the final colors (black, dark grey, light grey). I don't know if this way is convoluted or not, but I like the quality of the subs produced when I have full control. MaestroSBT can be tricky to learn, but once you do so it turns out excellent subs. The SSA subs I feed into it come out of SubStation Alpha.

[EddyH]

OK ... not sure I entirely follow but I might draw myself a flowchart of that and have a go
I'm happy to try and alter the workflow I've got myself stuck into, as it's still technically at the experimental stage (though I've already got the divx version ready for upload) - it's having to redo the whole subtitle file that I've spent several hours typing, timing, positioning and adjusting in what-i-thought-was-well-supported SSA format that's giving me horrors.

SSA is at least little more than a fancy CSV file, so I can try reprocessing it the same way as I originally made it - with the aid of Excel (and copy-paste to & from notepad to re-merge all the comma-happy lines of actual text). But, I'd still prefer to be able to just drop it into some tool - or set of tools - that can take what I've prepped in SSA format, and at the end of it provide a relatively accurate SUP compiled-bitmap file that DVDAG (or Muxman, or whatever) will accept and throw into the VOB, with suitable adjustments to the IFO palette.

Question 1: Are there any text formats that I can convert SSA to - or use instead - that will preserve the formatting I've used? IE font/size/colour/position by way of styles (even if it has to reprocess it into adding this data to each and every line), forced newlines, and a few horiz/vertical margin overrides.

Question 2: Does muxman have any way of doing multi colour subs? The DVD standard seems perfectly happy for me to choose between 16 different colours at any point in the video so long as each subpic only uses 4 of them at once. Sticking to 4 global colours pretty much forces me to either use a max of 2 (not really practical) or render it in a variety of ugly ways.

With a lot of people talking over each other at once I've found it very useful to differentiate between them by giving each "group" a different colour so the babble can be followed with less confusion (presenter 1, presenter 2, audience member who becomes pres.3, a fourth person they each have individual dialogues with, two guys from a particular department who jointly become the 5th, and then "everyone else" takes the sixth colour). This works out to 6x2 = 12 colours, plus black outline, plus transparent, plus two others for the menu buttons = 16. So it should work, shouldn't it? If I had access to a properly intelligent renderer then it should even be able to produce most of them antialiased, and just drop the smoothing for the few instances (about 10 max out of 550+ lines) where two different coloured subs overlap on the timeline.

I could mash it down to three different "voice" colours (pres 1, pres 2+3, everyone else) as there isn't actually any clash that way (just!), but it would then require an semitransparent background box to the subpic so the words would remain readable, and this might foul up the menu buttons unless I reduce them to simple underlines. Plus I have a moral aversion to doing it in such a primitive, worse-than-teletext way, when the ability is there to do it "right".

Had another go this morning with rendering out the Maestro-edited SSA to every major subpic filetype it supports, and still haven't really found anything that i've downloaded that can accurately load or convert them. Going to give the SUP+XML easySUP eventually deigned to make a go, if that fails, time to give your method a go. Then if that doesn't fly, revisit the tools list and do a text search for "SUP" and/or "SSA"...

Its that or take a couple weeks sabbatical to learn some kind of programming language in order to make my own small proggy to do it... which seems wierd. I can't be the first person in the 10+ years these two formats have both been hanging around to want to do this?

[oldfart13]

This is the same problem I've run into with Blu Ray players....no support for anything other than SRT subs. That means the nice fonts, colours and formatting of SUB, ASS, SSA, etc. is lost. No way around it. You have to convert all the other subtitling formats to SRT and that is that....

[EddyH]

Not sure about that Oldfart - I mean, easySUP *defaults* to Blu-ray SUP output...

However, no offense to anyone who's involved, but ... I've had a look at the "inserting subtitles with muxman(/etc)" guide, to get some of the detail behind the suspiciously simple looking "add subtitles with muxman, add video in vobblanker" instruction, and ... it's making my head spin again, in the middle of the day. I've spent a whole morning on this so far, and I've covered almost two A4 pages with tinytext flowchartage of how to carry out the actual steps involved. It would have been quicker just to remake the damn things from scratch.

Given that I already have all the source files, I have the raw M2Vs for the menus and videos, the AC3 audio streams, stuff that can easily become XML for menu authoring, and a full featured sub script file, I can't see why it has to be so damn complex. I'm not shoving a file downloaded off some website into an existing disc or anything. Seriously, are there not any simpler methods that don't involve pretty much making the disc without titles THEN hacking them into it? I can make single-colour, improperly positioned ones after all, so it's just that last step.

Like I say, no offence meant - if that's how it is, then so be it. But as YOU say - it seems convoluted. And even more so now I've gone through the steps. I'm supposed to make an instruction document on how to make these things for any colleagues who have to take over from me, and they're not people particularly given to reading long scripts (my normal writing is bad enough). If I'm on the verge of bafflement and frustrated exhaution then I think the whole idea's gonna fail

[Abas-Avara]

I opened recently a topic like this,
https://forum.videohelp.com/threads/318915-SSA-to-SUB

I converted it to SUB, but I'm not sure of DVDAuthorGUI can support it.

[manono]

But, I'd still prefer to be able to just drop it into some tool - or set of tools - that can take what I've prepped in SSA format, and at the end of it provide a relatively accurate SUP compiled-bitmap file that DVDAG (or Muxman, or whatever) will accept and throw into the VOB, with suitable adjustments to the IFO palette.

Fuggedaboutit. Although the readme for MaestroSBT seems to imply you can keep your styles, I haven't had any luck. But then I haven't really tried very hard either because I only want one style and set it up in MaestroSBT itself. Unless you're willing to go through the subs one-by-one to set a style for the different ones (like location on the screen), I think you should be satisfied with only one style. But if you want to edit a line, just double-click on it after making the preview. By the way, only 4 colors are allowed for DVD, so your idea of having 6 (or 12) different colors goes out the window.

Question 1: Are there any text formats that I can convert SSA to - or use instead - that will preserve the formatting I've used?

There are only 3 kinds of text subs which allow for formatting - SSA, ASS, and SRT with a styles sheet - and I know of no way to get them prepared for DVD at the same time keeping the styles you set up (unless you burn them or hardcode them into the DVD, a very bad idea, in my opinion). You wasted hours doing all this.

Question 2: Does muxman have any way of doing multi colour subs?

Muxman just authors whatever you send to it. The subs are prepared elsewhere (like in MaestroSBT). You have to remember that the BMPs prepared by a program like MaestroSBT aren't meant to be the final colors - the colors you see when playing the DVD. The 'real' colors are overlaid on top of the BMPs. I think MaestroSBT allows for the export of a pallete file that can be used later on in something like PGCEdit or IFOEdit to get your final colors, but I find it simpler to set the final colors in PGCEdit (although DVDSubEdit can do it as well).

With a lot of people talking over each other at once I've found it very useful to differentiate between them by giving each "group" a different colour so the babble can be followed with less confusion (presenter 1, presenter 2, audience member who becomes pres.3, a fourth person they each have individual dialogues with, two guys from a particular department who jointly become the 5th, and then "everyone else" takes the sixth colour)

It's a nice idea, and I've seen 2 different colors used in some Japanese films on DVD so I know it can be done, and it can be helpful when 2 people are speaking at once. But, again, I know of no easy way to do it. Even positioning the subs under the person speaking (as in some US DVDs I've seen) can be tricky. I have a feeling they're using expensive commercial programs and are doing these things line-by-line. About all I can think of to do would be to identify the people speaking

Mr. Smith: ....
Mr. Jones: ...
Mr. Green: ...

Although most DVDs don't have more than 2 lines of dialog on a screen at a time, I sometimes put 3. And again, with colors reserved for the outline and the transparent background, you only have 2 more colors with which to work. And that will then preclude any antialiasing, which for me is important for subtitle quality.

...but it would then require an semitransparent background box to the subpic so the words would remain readable...

I didn't quite understand the reasoning behind this one, but that transparent background color doesn't actually have to be transparent, but can have any degree of transparency you like and it can be just about any size you want. See here:

http://forum.doom9.org/showthread.php?p=1369921#post1369921

If you decide to try making SST subs for Muxman, use 4-bit BMPs and check the 'Don't clip bitmaps' box, both in the render section. You also want 29.97fps in and out with drop-frame timing in the timing section.

I've had a look at the "inserting subtitles with muxman(/etc)" guide, to get some of the detail behind the suspiciously simple looking "add subtitles with muxman, add video in vobblanker" instruction, and ... it's making my head spin again,

Muxman accepts subs in 2 different formats, SST and SUP. If you're creating them from scratch, SST is preferred, I think. SubtitleCreator can create SUP subs from SRTs, but I, for one, wouldn't do it that way. If you're interested, here's Baldrick's guide:

https://forum.videohelp.com/threads/277950-How-to-add-new-subtitles-to-an-existing-DVD

The method 2 also explains how to get the Muxman authored DVD into an already existing one with menus (yours created with DVDAuthorGUI using a small 'dummy video) using the 'Replace' button of VobBlanker. If you've never done any of these things before they might seem intimidating. Once you've done them you'll come to discover they're pretty simple to do after all.

[EddyH]

Right ... afraid I don't have time to read and digest all that right now as my mind is even more fried now (like an egg - could hardly get up this morning) and I'm going to be busy away from the PC for about the next 6 hours. Will certainly consume it with interest when I return though!

Just logged on to add a note of thanks to manono for the first reply. Admittedly I didn't follow much of it exactly, but the outline set me off on the right track, from it I ended up following the guide at https://forum.videohelp.com/threads/217500-How-to-add-your-own-subtitle-to-DVD-%28keep-...u-multi-pgc%29, and it's... sort of worked. As in I've got 99% of the subs I wanted, on the disc, displaying in the right colours, and at the right times (and roughly correct positions). There are some more niggles to iron out though - the menus don't work correctly any more for a start, which is what I'll be looking to fix next before tidying up the actual subs themselves. Wondering even if I could take my reauthored files and break the subs out as a .SUP to loop back into the original DVDAG project and make it "neat" that way.

The process looks quite bafflingly complex, but if you break it up into one-step stages and follow them through, it does actually start to make sense.

Cheers

[manono]

Wondering even if I could take my reauthored files and break the subs out as a .SUP to loop back into the original DVDAG project and make it "neat" that way.

I didn't really understand what you're asking, but if you're asking for a way to extract the SUP file, use PGCDemux.

[EddyH]

But, I'd still prefer to be able to just drop it into some tool that can take what I've prepped in SSA format, and provide a relatively accurate SUP

Fuggedaboutit. Although the readme for MaestroSBT seems to imply you can keep your styles, I haven't had any luck.

The annoying/infuriating thing is that it DOES preserve style information from the SSA... but in a completely crazy manner which gives output only indirectly related to the input. I mean, I've managed to figure out how to tweak the settings for the styles so that they more-or-less replicate what I saw via Avisynth (with output set to 720x576, as for input to HCEnc, mind), in a formulaic rather than merely trial-and-error way, but I haven't yet figured out how to compensate for this "in advance" - or automatically, or why it does this. Most particularly, the font size, margins and horizontal character width are on a totally different scale to the original simple pixel-position / 1:1 scheme, and it CHANGES when you alter the paragraph alignment - in fact, you can't force a left-align sub's right hand margin to be any further across than about 2/3rds of the screen. Any margin overrides only serve to make things worse - they operate on a DIFFERENT scale to BOTH Avisynth TextSub AND Maestro's own style spacing! (So, e.g., a centred sub specified to have a 200,30 start position, then exhibiting one forgotten line with a 200,30 override will appear to dance around, with neither position corresponding to the original intent).

This itself wouldn't be so bad if I hadn't already manually specified the style margins and a number of overrides by hand in the SSA, checking against the Avisynth output. The mathematically readjusted version works fine on an AVI, and the original looks OK if hardcoded on the DVD image, and looks fine in SSA itself also. Put it in Maestro... poof. All over the place. And have to do all that work a second time, working effectively "blind" (a small reprieve is that the settable "guidelines" DO correspond to pixel positions) - and will have to with any future dual-format versions because I can't apply the maestro subs to a divx.

ANYWAY. Moving on. It's a freeware app after all, and the only thing I could do to fix it is inform the author and wait. No sense in further gripes

Unless you're willing to go through the subs one-by-one to set a style for the different ones (like location on the screen), I think you should be satisfied with only one style.

Unfortunately I'm not / can't be. Let me set the scene - a recording of a two-speaker architectural presentation with interjections from the audience, a Q&A session, and some info-heavy PPT slides and images cut into the video, luckily with bold enough details/large enough fonts that they're readable after a dose of deflicker, anamorphic squish and MPG encoding. I need the styles to, at least, differentiate between the different speakers, of whom there can be 3 or 4 in rapid rotation, sometimes speaking over each other - this is what the text colours do.

Additionally, I need to shift the position around to 1/ avoid any onscreen text or diagrams that would otherwise become unreadable, and keeping the subs themselves readable by avoiding confusing background patterns; 2/ further emphasise who is talking - avoidance of sub collisions, putting them at least on the appropriate side of the screen and sometimes altering the height from baseline; 3/ keep them within the overscan safe limits!
(Actually what I originally did was keep them within the P&S safelimits for a 16:9 picture, but now know that's unneccessary... however that now complicates things further by my needing tw... three sets that are differently positioned to maintain avoidance regardless of screen shape and matting - we have DVD players stuck on all three modes (yes, some dork set a few of our main players to PanScan, and I've had to figure out how to make it work in HCEnc!*) in different places)

This seemed pretty simple, once I'd figured out that SSA format is basically CSV with a couple of knobs on. Let Excel do all the heavy lifting in terms of cell-filling, export it to CSV, then do a simple find/rep to remove all the surrounding quotemarks (whilst leaving the mid-line ones intact, of course). And it appeared to work well. With any luck I can find some way of using this discovery to easily change the styles and margin overrides to Maestro-friendly equivalents, rather than having to roll thru the whole goddamn thing one subpic at a time, looking for obvious hiccups...

*further awesomeness (not) - turns out PowerDVD's PS is broken. That, or the PS on the settop player I also used to test. When one goes left, the other goes right... Not particularly helpful when using it to keep a presenter who will not stand still within the PS frame, as they ended up only just within the widescreen underscan. I had thought HC was broken when all I tested with was PDVD, and had to insert a FlipHorizontal command into my script whilst setting up the positioning file. Luckily all this'll need to fix is a find-replace... all CRLFs => CRLF plus a random character, all instances of character then "-" replaced by null, then all instances of character replaced by "-"...

By the way, only 4 colors are allowed for DVD, so your idea of having 6 (or 12) different colors goes out the window.

Oh, but it doesn't! Each subpic is limited to 4 colours each (or on a practical basis, a Gameboy/NES-like "three plus transparent" unless you like large semi- or fully-opaque backing boxes) - but they pick those 4 from a master palette of 16 colours chosen on a PGC-by-PGC basis (so each movie has a separate palette, as does each menu). This one baffled me for a while before I saw how it works... the control line for each SP holds its position, size (ie creating a rectangular bounding box for the "background" colour), start/end times, plus four bytes (or eight nibbles) giving a PGC palette index and 16-level transparency for each of the 4 colours. Like most things I'm discovering in the spec, wierd and unneccessarily complicated, but presumably done on the basis of wringing the greatest utility out of bare-minimal hardware.

For example, with enough pratting about on my behalf or with suitably designed software, I can have my subs fade in and out (appearing to do so "smoothly" over the course of 1/2 second, at NTSC30), display a total of fifteen different "voices", with up to three simultaneous, by having them as non-outlined (or bright-on-dark AND dark-on-bright... or upto 225 by using ALL colour combis, with upto 9 onscreen this time) characters on a semitransparent box background - 16 (256, but again only 9 simultaneous) by switching between translucent white and translucent black, or... as is my eventual aim... have upto 7 antialiased, outlined solo speakers, with the ability to drop back to non-antialiased when two people are speaking at once or SFX need to be noted. All without changing the font, or font style/size.

What I HAVE acheived so far is to use all six of my intended colours in a non-antialiased fashion. Maestro AUTOMATICALLY whizzed up the files so that when dropped into the VOB via Muxman, it showed six different ways (once activated), and then all I had to do was figure out suitable YCrCb hex values (for black, white, cream, light cyan, light green, pink and "light brick") and poke them into the IFO/PGC palette table.
It's also made up a version with the antialiasing (using 13 colours, black plus full-bright and dim versions of the aforementioned colours), but of course when two people talk at once, it picks only one colour to render them in. However as they're all BMP files with an accompanying plaintext description file, it's a simple matter of elbow grease to find the crosstalk points and sub in the non-antialiased image, with any requisite futzing of the palette lookups and the BMP itself.

Kinda glad I come from the 16 bit generation (and a little bit of the 8) otherwise I'd probably have no hope of understanding all this crap ... whole computers with low bitplane paletted displays FTW, as far as teaching tools go.

Q1: any text formats that I can convert SSA to that will preserve formatting

There are only 3 kinds of text subs which allow for formatting - SSA, ASS, and SRT with a styles sheet - and I know of no way to get them prepared for DVD at the same time keeping the styles you set up (unless you burn them or hardcode them into the DVD, a very bad idea, in my opinion). You wasted hours doing all this.

Gahhh.
Way of the world I guess. Sooner or later I'm going to have to end up learning how to code in something other than QuickBasic, I keep running up against these situations where I want to do something that doesn't seem that specialised, but it still hasn't been catered for by any previous adventurer.
(Also including: finding suitable gearboxes to swap onto at least two old cars so far - I'm the one who ended up compiling and disseminating a summary sheet using multi source data though fair enough in the latest instance, you wouldn't believe how many barely-different variants Renault blow through)

Hmm, SRT + styles may yet be worth looking into. Pretty sure I've seen at least one SUP-generating utility that accepts it. However, being able to burp out Scenarist-compatible stuff from Maestro (and therefore integrate it via Muxman) will do for now. Seem to have lost a couple of subpics and gained a couple of timing errors somehow, but that's traceable and fixable.
I rediscovered PCGDemux independent of manono's mention and so can at least - using that rather longwinded method - come up with a reusable SUP file from an SSA input somehow (just having to change the IFO palette after). Now if I can just figure out WTBFH is going on with Maestro's margins and font sizes... alternatively cut losses and just use the SSA as a cribsheet for fixing up a simplified conversion in SubtitleCreator and the like in future.

I think MaestroSBT allows for the export of a pallete file that can be used later on in something like PGCEdit or IFOEdit to get your final colors, but I find it simpler to set the final colors in PGCEdit (although DVDSubEdit can do it as well).

Thanks... might see if that can be done somehow, though really it's not so much of a chore poking a few sets of three hex pairs into the IFO... compared to all the waiting around for the other stuff! (Even made myself an excel sheet to give approximate-but-good-enough YUV values for RGB input, including matching some of the menu select boxes to the institution's mascot colours). Have had a look through PGC and DVDsub and they're just plain baffling, at least at the moment, though I'm sure they can be used to fix my other problems somehow. IFO at least is reasonably well documented and simple enough to be almost understandable.

It's a nice idea, and I've seen 2 different colors used in some Japanese films on DVD so I know it can be done, and it can be helpful when 2 people are speaking at once. But, again, I know of no easy way to do it. Even positioning the subs under the person speaking (as in some US DVDs I've seen) can be tricky. I have a feeling they're using expensive commercial programs and are doing these things line-by-line.

Or not, as above... it's just not very widespread knowledge, it seems. Best analogue I can think of comparing it to is the palette-switching tricks on old computers used to put more colours on screen than officially specified - though of course you could still only have a certain number on each set segment of the screen... or in this case, in each successive subpicture - and unless you did something *truly* wild, you can't exceed the colours defined in the (virtual) machine's master palette. Might even be worth me making up my own guide once I manage to understand it myself... just as DVD is going out of fashion, hehe.

Although most DVDs don't have more than 2 lines of dialog on a screen at a time, I sometimes put 3.

Unfortunately I'm not dealing with movie actors and dialogue, or even well-experienced public speakers (I'm dreading the "next" one waiting in the queue, where the guy clearly understood NOTHING about projecting to an audience, and didn't even wait long enough for me to give him a radio mic before launching his spiel - dunno if it's going to be even audible enough to transcribe). They simply race through the presentation (and finished 20 minutes short of a 1-hour slot); transcribing it was a wierd few hours sat with the soundtrack playing in VLC at 33% speed.
A few of the SPs are one-line, most are two, and a few more are three - with a small number of fours, mainly when fitting around a diagram. Thankfully the slight pausing effect that comes from them interrupting each other means none of the crosstalk SPs are more than 4 lines high.
Wish they DID act more professionally... could use one colour and one line at a time then!

And again, with colors reserved for the outline and the transparent background, you only have 2 more colors with which to work. And that will then preclude any antialiasing, which for me is important for subtitle quality.

Yes... the non-smoothed ones do look a little rough, hence my wanting to try the hybrid version (if only it would happen automatically *sigh*). Still, they're not _that_ bad, even using something like 18-point Tahoma regular (Maestro won't even reliably render smoothed outlined text below that size!), so appearing that way for a net total of about 10-15 seconds out of the whole production shouldn't be too jarring.

...but it would then require an semitransparent background box to the subpic so the words would remain readable...

I didn't quite understand the reasoning behind this one, but that transparent background color doesn't actually have to be transparent, but can have any degree of transparency you like and it can be just about any size you want.

That kind of was the point
i.e. the fourth colour doesn't really count as one, unless you use this method, which is a bit of a last ditch variant. I've seen it used on several TV programs (who don't really have an excuse , including a currently-airing BBC show filmed on location in Lagos, where it makes the unoutlined text more readable (but still not brilliant), and in fully-opaque mode on teletext and digital cable (ugly, ugly, ugly).
Besides, if I DO actually use the Scenarist output from Maestro, it seems to make full-resolution (ie in this case 720x573) bitmaps.... which as Muxman doesn't understand RLE bitmaps (only been around since... oh, i dunno, windows 3.0?), end up chowing a few HUNDRED megs of space - but more importantly don't allow this clever use of trimming. Maybe there's an option to change this. Haven't played with it enough to know!

If you decide to try making SST subs for Muxman, use 4-bit BMPs and check the 'Don't clip bitmaps' box, both in the render section.

Then again, maybe not. Balls. That's what I get for deciding to go to Reply halfway through the post.

You also want 29.97fps in and out with drop-frame timing in the timing section.

Not for PAL I don't, but thanks anyway

Muxman accepts subs in 2 different formats, SST and SUP. If you're creating them from scratch, SST is preferred, I think.

Yes, it was bulky and not exactly fast, but it seemed to work well enough. SUP would require making them elsewhere anyway.

https://forum.videohelp.com/threads/277950-How-to-add-new-subtitles-to-an-existing-DVD

Think that's the one I already followed! Approaching dangerous levels of spodularity with all these, but I found that self-bulletpointing it and strictly following through in sequence made it easier.

The method 2 also explains how to get the Muxman authored DVD into an already existing one with menus (yours created with DVDAuthorGUI using a small 'dummy video) using the 'Replace' button of VobBlanker. If you've never done any of these things before they might seem intimidating. Once you've done them you'll come to discover they're pretty simple to do after all.

Yes, that's basically what I did - though as I had the disc space and time to spare, I just used my fullsize encode and skipped the extra bit of having to replace that in VOB Blanker. May use the short dummy next time... but, yeah... I have scribbled notes I now need to type up that codify this process in a bare-minimal manner.

Aaaaaaaaaaaaand I'll leave the current state of affairs for a second post so it's easier to reply to than this creeping horror

[EddyH]

So yes. Where I am at the moment:

Actual video, stillpic and audio parts conquered.
Making up the MPG parts of the menus conquered.
Transcribing, syncing and initial positioning of subs, with simple multicolourlarity.
I think I've now cracked turning the subs on & off reliably through the menu, even after poking the subtitles and palettes back in.
Possibly including Wide, PS and LB versions of the titles via Muxman (and then back thru DVDAuthorGUI)

^ ^ ^ ^ ^
both these last two still in testing... it was still reauthoring when I started writing all this nonsense


To be sorted:
FancySubs™
Feeding them through from initial SSA file to Maestro-mediated, fully correct Scenarist output in a way that doesn't cause a collosal pain in the nethers.
* Important * --- Including Wide and Pan-Scan versions of the menu buttons.
I might be able to crack the rest of it on my own, but ... with my current software complement (DVDAG and it's underlying XML description file and CLI-based programs, plus 85-ish megs of other freeware downloads including pretty much everything listed on this page so far and in Baldrick's guide plus random other muxing and subbing apps), is there any way to do this? I had a go with DVDSubEdit, and although I could export the subpic image to Paint and stretch it in a suitable fashion (when viewed on a screen showing PS, it appeared too narrow... can't remember if it was vertically stretched for LB), reimporting it just altered the SP for ALL versions - so Widescreen was now too wide, and I couldn't find any option to import as just one of the aspect options. Either way, some potential users end up with buttons in obviously incorrect positions.

Naturally, the GUI part of DVDAG only lets you firstly set whether the M2V still you use for the menu is 4:3 or 16:9, and then make a subpic tailored to this...
Not doing anything fancy with it - I have literally used only, and am happy with, the bogstandard, single-colour blunted-rectangles DVDAG spits out (at 2px thickness), and don't mind if they're slightly distorted by the resizing. They'll still be perfectly good enough for the purposes of selecting between text lines saying Start, Chapters, and Subs, and moving through and between a couple of pages with plaintext chapter titles on them. The styling is corporate and austere

[manono]

* Important * --- Including Wide and Pan-Scan versions of the menu buttons.

I don't mess with Pan and Scan menu buttons myself, but isn't P&S just a cropped version of the Wide? If so, you don't do anything different. Just make them for Wide (16:9), leaving a lot of nothing on the sides so they'll still show when cropped.

Most particularly, the font size, margins and horizontal character width are on a totally different scale to the original simple pixel-position / 1:1 scheme,

Yeah, there must be 2 different size standards in use here, and one's used by SSA and the other by MaestroSBT. I may (for example) make them sized 60 for SSA but then have to change them to 23 for MaestroSBT. I'm sure someone else will know more about it than I.

I couldn't see anything else in the long posts that needed my attention. Good luck.

[EddyH]

* Important * --- Including Wide and Pan-Scan versions of the menu buttons.

I don't mess with Pan and Scan menu buttons myself, but isn't P&S just a cropped version of the Wide? If so, you don't do anything different. Just make them for Wide (16:9), leaving a lot of nothing on the sides so they'll still show when cropped.

Unfortunately not! And that's the "last" thing I'm now having problems with.
The disc can actually store up to 3 versions of any subpic stream - plain widescreen, letterboxed, and panscan (if it's a 4:3 title, you only need to store one, because it's the TV that crops or adds borders not the player, so it presumably takes up the WS slot). It's because the player hardware standard is so simple that it can't (or at least, doesn't) resize subpics upon display - so they always show at 1:1 pixel size relative to what comes out of the player's physical video port. Presumably it's supposed to be some kind of genlocked overlay produced after the decoder/horizontal trimmer/line dropper.

So I take my disc with its menus on, somewhat dumbly made by DVDAG to only be of one type (maybe i need to change my author package - i've ended up with it sort of by default as it was the only one i could make to *work* last year)... it looks fine on a widescreen display... I put it in the player that's been annoyingly locked to pan-scan... and although the vertical spacing is fine, the horizontal is completely off; the buttons appear scrunched up and off-line. The authoring package, after all, DID pre-compensate for anamorphic stretching. Presumably if I was to take this and put it in a letterboxing player (didn't bother carrying on after that discovery), it would be a similar problem just with the axes reversed - horizontal fine, but appearing vertically stretched.

I also made that same assumption you did, when I was making the subtitles. Sure, they don't go off the side of a panscan screen, in fact they're never closer than about 1/4 in (as they were positioned to stay within a 4:3 box on a 16:9 screen - needlessly!), but they do end up overlaying things that they'd been spaced to avoid on the widescreen...

(maybe I could just throw in the towel and make my menus 4:3 and let the hardware sort it out? it's looking tempting - and the source bitmaps are already 1024x576 with everything kept within a centre ~680x500 safety window... bit of a cop out though!)
(* btw... yes... we have lots of 4:3 screens still - it's only very recently that affordable high-rez data projectors have started using 16:10 or 16:9 panels, and they spend most of their time slaved to an SVGA or XGA desktop)

Most particularly, the font size, margins and horizontal character width are on a totally different scale to the original simple pixel-position / 1:1 scheme,

Yeah, there must be 2 different size standards in use here, and one's used by SSA and the other by MaestroSBT. I may (for example) make them sized 60 for SSA but then have to change them to 23 for MaestroSBT. I'm sure someone else will know more about it than I.

Ugh. That may end up being me (again, like with the gearboxes ) --- I'll just have to experiment with it when I get some free time and see what numbers equate to what, and maybe even work out why. Or just email the author and see if they're friendlier to questions than the last freeware coder I talked to...

(3 different standards btw - the old flavour of MSBT must have used something about as different in the other direction; flick on "use old scale" in the compatibility menu, and your previously "too big!" text now becomes far too small. Happily I have found that making PS versions of WS subs is a reasonably easy though not idiotproof task - flick it from 16:9 to 4:3 compensation... then halve all the horizontal margin figures in all the styles - but DON'T touch the manual overrides AT ALL. Yeah, that program's got some bugs to be worked out...)

I couldn't see anything else in the long posts that needed my attention. Good luck.

Thanks ... and nope, it pretty much descended to chatting
I've cracked most of the problems now, partly down to you guys' help, so many thanks.

********
The one, last thing I need to know however - if I don't figure it out through the guides (doesn't look hopeful so far) - how can I extract and/or make, then insert, my own MENU subpics?
********
Because I think my last hope now is to make the skeleton with DVDAG, then fix up like 50% of what it produces using Maestro, Muxman, PGCEdit, etc... including manually editing the menu button images.

Making the subtitle ones is all well and dandy, but with the tools used so far, if there even IS an option to do it for the menu PGCs (rather than the video ones - not often), it merely results in an error... "stream too short" or various others when trying to extract it, and I don't even know where to begin for making my own from scratch (well, OK, making a new layer in paintshop, drawing boxes over my aspect-adjusted menu bitmap, then saving the layer off as its own 4-colour BMP... but AFTER that!)

Or in the case of DVDsubedit, only being able to get and insert ONE of them per menu (so no matter what I do, it'll always be wrong for two of the three settings!), and that a somewhat innaccurate-looking version the app pasted into MSPaint.

I mean, using PGCedit, I've even managed to figure how to change the button-window positions, how they link to each other and what actions they perform, and have it automatically adjust THAT for playback shape (so it's 99% there!).... but NOT the underlying subpic said buttons window onto. So now I just get chopped up fractions of the misplaced buttons!

Oh, such fun

[manono]

Of the thousands of retail DVDs with which I've worked, I've seen exactly one that had all three, WS, Letterbox, and P&S. I don't think that someone stupidly having set their DVD player as Pan and Scan (stupid, because there's never been a DVD released that makes use of that feature in the DVD standard). So, I don't think it affects the menus, but only the video, so it gets cropped and not letterboxed (and doesn't get panned and scanned). Making your menus for Wide + Letterbox or Wide + P&S should be enough. And for the person that insists on setting his for P&S, f*** him. Why should you 'dumb down' your DVD, and make extra work for yourself, for idiots?

I opened a DVD with 16:9 and Letterboxed menus and had no trouble saving out the BMPs. Admittedly, the first time I tried to save out the Letterboxed one it gave me some sort of an error message about having to select the SubPic, but afterwards no problems. If your problem persists, open the Menu, and then in the upper right of the DVDSubEdit screen choose one of the streams (WS, LB, P&S, whatever), it'll present only the stream you pick, and then you'll have no trouble saving the BMPs.

Once saved you'll edit in Paint, Photoshop, whatever. And then you can reinsert it

[EddyH]

Of the thousands of retail DVDs with which I've worked, I've seen exactly one that had all three, WS, Letterbox, and P&S. I don't think that someone stupidly having set their DVD player as Pan and Scan (stupid, because there's never been a DVD released that makes use of that feature in the DVD standard).

Really? Not a one? They all default to letterbox?
*goes home and samples entire DVD collection with DVD player set to "auto" then "PS"*

...why is it even an option any more then?
(it just works as an instruction to tell the player to crop to a certain 540 pixel wide chunk of the anamorphic image, btw)

So, I don't think it affects the menus, but only the video, so it gets cropped and not letterboxed (and doesn't get panned and scanned).

Unfortunately not, again. I didn't MAKE my video for PS originally - I took my happily all-widescreen production, put it in a player, and... only got the centre of the image for all the menus and the video. And mis-sized menu buttons, revealing it to be player-side, not something wrong with the projector.

Making your menus for Wide + Letterbox or Wide + P&S should be enough. And for the person that insists on setting his for P&S, f*** him. Why should you 'dumb down' your DVD, and make extra work for yourself, for idiots?

Because it's a player connected to the humongous projector in the lecture theatre at the college where I work? I'm not making it for "just some guy". It's my job.

And though I could probably hunt down the (naturally!) missing remote and set it to letterbox*, and though the disc would probably not get played on that exact system, it's alerted me to the fact that there may be multiple players out there which my proj might get spun in where this would be a problem.

Basically it's a stupid little niggle that I'd like to be able to fix now, whilst I don't need to, so I'm not stuck with it SHOULD I need to work around it in future.
Plus if I end up needing to adjust things to account for letterbox mode, it's not so much more to account for PS as well. Typically running a second round of "crop, resize, hit re-render" but in the other direction, and quick-scanning through the video track to see if there's any points where I need to shift the PS window left or right.

More troublingly than not being able to set up the three different subpic tracks - even in programs that specifically seem to cater for doing so - is the difficulty in changing the damn thing at all. I remade myself an example button pic in PSP, made sure it was set up properly, tried to import to DVDSubEdit ... no dice. It was just all blank and it wouldn't even let me change the opacity any more. Gnnrgh.

* I can't set it to 16:9, and the projector to anamorphic, and be done, because there's at least two other inputs to that projector on the composite line via an AV switch, and both of them are 4:3... So even once it's letterboxed, it may not show the menus correctly.

I opened a DVD with 16:9 and Letterboxed menus and had no trouble saving out the BMPs.

Yes, I've managed to do this with DVDSE now (was misunderstanding the program I think!), which is what I based my altered versions from (though I ended up starting from scratch using the make-a-layer & resize it idea). The trouble was trying to extract e.g. SUP files... and now re-inserting the damn things.

If your problem persists, open the Menu, and then in the upper right of the DVDSubEdit screen choose one of the streams (WS, LB, P&S, whatever), it'll present only the stream you pick, and then you'll have no trouble saving the BMPs.

Ah - now, ysee, my choice for that is either "all", or "WS+LB+PS". So, yeah...!
(What I did was choose from the one at the lower left of the preview screen ... not sure what the point of it is given that it seems to have little effect on the actual subpic that's displayed, but it allowed me to save something out)

For the record, I HAVE tried it both with the VOBs as presented - with only one apparent track present - and after jiggering it about in PGCEdit (and looking for clues with IFOedit - nothing) so that it looked for WS as x20, LB as x21 and PS as x22. DVDSubEdit then only presented me with a choice of "all" or "WS" and no way to make a new one. Ffffffffffff.......

Time to change or at least raise my game on the original authoring side I think..... maybe use DVDAuthor in its bare form, if there's nothing else more flexible available for gratis.
(Or as I said before - give up and do the menus as 4:3... it just seems so ungeekular to do that, letting the problem win when there's no good reason I shouldn't be able to make it work

[liquid217]

I haven't read this entire thread, but I have noticed that you are trying to produce a 16:9 menu using DVDAuthorGUI. Although DVDAuthorGUI does allow you to create and import 16:9 m2v's for your menu. The Pan/Scan, and letterbox modes will not show up correctly on the menu, do to the fact that 2 other sets of buttons must be generated. I have tried to work on getting this compliant for 16:9 menus, but have been unsuccessful. I have even had one user email me his project files (bmps, xml, scripts, etc) where he was able to get spumux to correctly mux the modified BMPs into the vob container. Although it worked in his example, I was not able to reproduce this systematically (inside my GUI, in other words). To be honest, there is very little in the dvdauthor man pages on the subject, and none of the other open source frontends seem to support either; so it will be next to impossible for me to integrate it into the GUI (at least at this time).

[manono]

Really? Not a one? They all default to letterbox?

That setting is for the video and not the menus. There's never been a DVD that uses Pan And Scan on a 16:9 widescreen video when output to a 4:3 TV set. All it'll do is crop equal amounts from the sides.

For automatic pan & scan mode, the anamorphic video is unsqueezed to 16:9 and the sides are cropped off so that a portion of the image is shown at full height on a 4:3 screen by following a center of interest offset encoded in the video stream according to the preferences of those who transferred the film to video. The pan & scan "window" is 75% of the full width, which reduces the horizontal pixels from 720 to 540. The pan & scan window can only travel laterally. This does not duplicate a true pan & scan process in which the window can also travel up and down and zoom in and out. Auto pan & scan has three strikes against it: 1) it doesn't provide the same artistic control as studio pan & scan, 2) there is a loss of detail when the picture is scaled up, and 3) equipment for recording picture shift information is not widely available. Therefore, no anamorphic movies have been released with auto pan & scan enabled, although some discs use the pan & scan feature in menus so that the same menu video can be used in both widescreen and 4:3 mode. In order to present a quality full-screen picture to the vast majority of TV viewers, yet still provide the best experience for widescreen owners, some DVD producers choose to put two versions on a single disc: 4:3 studio pan & scan and 16:9 anamorphic.

http://www.dvddemystified.com/dvdfaq.html#3.5

The bold facing is mine. Note that it also confirms what I said about a P&S menu being the same as a Wide menu, the background menu video anyway.

[EddyH]

Oh, now that's just perverse
However, I still hold reservation about the "none" statement Also putting the two versions on "for better quality" seems to be against all ideas of e.g. using the PS feature to allow you to present the two different versions without losing half of your potential average bitrate. And particularly rubbish if e.g. you're making a widescreen feature where an acceptable PS VHS version already exists, and the choice is either lose vertical rez (vs the VHS!) by letterboxing, blindly chop the left and right sides without using the re-centring feature, or wasting 4Gb on a second copy.

I do wonder about that "no equipment can do it" when I'm able to do it in a freeware MPG2 encoder. It's just part of the internal control stream/frame-by-frame tags or something.

The fact that they do it on the menus but not the video seems even stranger, when it's far easier to centre things on a menu.

Still, if that's the way it is, then that's the way it is... and it may even look now like I CAN force the video to show letterboxed regardless of player setting. Still doesn't help with having to make two sets of subtitles (actually MORE of a pain making the LB ones than the PS!), and the three-way menus though.