DirectX Lazarus.... comments, questions...

[Howard_Casto]

If anyone has been wondering why the releases of my fe have slowed to a crawl it's because I'm finally getting to upgrading the gui... slowly but surely.....

I had planed on releasing a new version of lazarus, with various overlay and graphical special effects right after the holidays.  Unfortunately I feel that I have pushed the built-in windows gdi to it's limits at this point....  I have decided to slowly, but surely start the process of porting the gui to directx.  I have already ported almost the entire sound engine to directshow and so far it's a great success in terms of freed resources.  

This is really just a "what do you think" post because porting to directx will mean several things....

#1 Due to the way dx8 draws it's text, font options for skins might be slightly crippled.  

#2 Many of the wild, new features I've added over the past version or two would be gone until I make dx versions of them.  

#3 Video previews would still be supported, but most likely only in mpeg or avi format... in other words if you get them from 3darcade they will have to be converted.  

#4 The same goes for cabinet models... they would be supported, but probably not in w3d format, so they would have to be converted.  This isn't a big issue as I make half of the models myself and am in contact with the other two guys who make the rest, but it would still take time and extra bandwidth.  

#5 No more pos machines.  I want to use dx8, so if your video card/os can't handle dx8 very well then you will be s.o.l.  I decided upon this because, quite frankly, if you machine sucks then you should be using gamelauncher or some other, "resource friendly" fe, not mine.  If I work in dx8 then the fe will be modern for years to come, dx7 is the past, unforunately.

#6 It will most likely be slower than lazarus is now for a long while until I can get it optimized.  This shouldn't be a huge issue as lazarus is mad fast now, I down-clock it in several respects.  

#7  The next "transition version" of lazarus will be horrible at best.  I'm not saying it won't work, but considering the sound will be dx, the graphics will be gdi and I will still have all of these goofy effects indside the code, even if I don't make you a "skinner" for them, there may be conflict issues.  Obviously there are a lot of possible things to go wrong, and thus it will probably have a bug or two until I get everything in place.  

I want to do this is for two reasons....  

#1 In dx8 there isn't any difference between a 2d and 3d object.  That means text, graphics, backgrounds, ect can all be in 3d as easily as they can be 2d.  Imagine a select screen similar to the main menu of mortal kombat deadly alliance.  (2d text in front of a 3d hallway that you travel down as different options are choosen. )

#2 No more flickering/layering problems.  This means an avi or animated gif could be incorporated without any issues as well as more advanced animation effects.  


Ok now I need feedback..... I will post additional progress on this at fe.donkeyfly.com from now on but I really need the "average joe's" input on this as well.  

[CM]

I think this is a great step in the right direction.   Here are my average-joe comments:

- agree with machine speed and dx8.  Make it modern .. computers are cheap enough to buy/upgrade and if not there are other FEs out there.

- Release this work as "LazarusPro-alpha0.01" or something to let people know it is work in progress.  A Lazarus5 that is buggy/slower won't help the fanbase.  I would rather you squash bugs on a limited beta than release it to the world and have the discussion threads flooded with "Laz is broken/won't work/sucks" comments.  Else you might spend more of your time mad at people for not RTFM!!!     Just keep everyone updated on your progress to keep them interested.

- The 3dcabinet/vid file-format change is slightly annoying (but understandable ... especially if it does away with dependences for 3rd party apps like realplayer/shockwave).   Make sure you readily point to a batch-conversion utility or get 3daracde/other to host your new formats.  If you want users to change them manually one at a time on their own .. most average-joes will turn the feature off or move on to another FE.

-  The dx8 improvements and possibilities are worth growing pains.  
Keep up the good work!


Average Joe's motto:  customizable/fast/free/now.
Programmer's motto:  my-way/fast-enough/$$$/when-it's-done

[Howard_Casto]

Thanks for your reply, keep em coming people.... In regards to releasing the alpha versions perhaps you are right.  The question is will people wait several months for it?  

There is also something I didn't mention that I need to clarify.  The new Lazarus will NOT be like 3darcade.  )p( and myself work very hard to make sure our fe's compliment each other but do not end up becoming the same thing.  Quite frankly, he does the fps environment way better than I ever could, and thus there is no point in me trying.    So before anyone asks there will not be a fps mode like he uses.  3darcade does this perfectly already, so I wouldn't want to do that.  What will happen is the fact that "skins" will evolve into maps and lists will evolve into map positions. That is optional, of course... you very well could just stick to a 2d background image.  I just want everyone to understand that it will still "feel" like lazarus, it will just have interactive, 3d, elements.  

[)p(]

 Make sure you readily point to a batch-conversion utility or get 3daracde/other to host your new formats.  If you want users to change them manually one at a time on their own .. most average-joes will turn the feature off or move on to another FE.

We will most definately make the models available on the 3darcade website in the format lazarus will use! So there will be no problem.

Peter

[Spectrum]

Sounds good to me Howard.  I've been working on setting up Laz the past day or so (not the whole time just sporadically when schedule allowed) and so far I like it.  Would love to see the DX8 features included  As far as releasing it as an alpha, I think that's the way to go.  We already see enough post's about not being able to get it working now.  I may have a question or 2 about the configs after I go through them a time or two if I still can't get it to work.

Also, I would like to say thanks for putting this out for the masses even if it was primarily designed for your own use.  IMHO it's worth the people that whine and complain for those that appreciate your work to have it available.  I will definately do some testing on it as you release, and report the good/bad to help you sort things out and get it up and running as you intend.

Robert

[Lilwolf]

Howard, let me know what file format you plan to use.  I have been in contact with others for a project I'm working on.  

There are some 'issues' with different versions.  Mainly the public domain ones.

[Howard_Casto]

Yeah I've ran into similar difficulties....  One day when I get a chance i will try converting to every compatable format and see which one works best. The 3d model viewing is way down the road on my list of things to do at the moment though.... My first order of business is re-creating the original lazarus skin features (ie "2d" stuff.) It shouldn't be hard though....    

[)p(]

Howard, let me know what file format you plan to use.  I have been in contact with others for a project I'm working on.  

There are some 'issues' with different versions.  Mainly the public domain ones.

It would be mean less work in converting the models if Howard and you can find a way to use a commen format for both of your frontends. Maybe the x3d format you suggested?

Peter

Peter

[Lilwolf]

As long as it's an open standard, I will be fine.  I haven't been able to do any work at all since I'm working out of my home.  At night when I get time to work, it's on real stuff instead of fun stuff.... (I've also gotting into EQ again).

but my cabinet is outside and might not make it back inside until spring when I get a chance to work on a cocktail table.  I don't know how much effort I will put in until after I have a cabinet to work with... So you might be working on your 3d stuff before I am...  sigh...

But just talking about it makes me want to start working on it again..

[Dave Dribin]

I haven't looked much, but I think MPEG2 or MPEG4 are fairly standard.  They're probably as open as they come.  I've been thinking about adding support for MPEG2 in GL, and I found a couple libraries that could potentially work.  I've also been looking into MNG.  AdvanceMAME can save animated screen shots in MNG and this is obviously supported by AdvanceMENU.  Adding GL support for MNG would allow it to use these screen shots, too.  MNG is a cool format, and probably easier to add than MPEG2.  It's also lossless.  But it's also fairly new and not well supported outside of Advance*.

-Dave

[Howard_Casto]

I haven't looked much, but I think MPEG2 or MPEG4 are fairly standard.  They're probably as open as they come.  I've been thinking about adding support for MPEG2 in GL, and I found a couple libraries that could potentially work.  I've also been looking into MNG.  AdvanceMAME can save animated screen shots in MNG and this is obviously supported by AdvanceMENU.  Adding GL support for MNG would allow it to use these screen shots, too.  MNG is a cool format, and probably easier to add than MPEG2.  It's also lossless.  But it's also fairly new and not well supported outside of Advance*.

-Dave

Well from what I've learned by playing around, directshow for dx8 will play pretty much every format natively EXCEPT real files.  There also may be a filter graph available for real at this point, so it might be that a conversion won't be necessary.  However, if I don't convert it will most definately be a mpeg format.  Why?  Because every coding language I've seen can play mepgs without too much hassle and as the mpeg format isn't "copyright locked", anybody can convert their media files to mp3 or mpeg.  Mng seems very interesting, but as you said nobody uses it yet.

The thing is these preivew videos are time consuming to make, and It would be nice to allow people to make them for us with as little hassle as possible.  That means we are pretty much limited to capturing in avi or mpeg formats and converting from there.  If I have to go through the hassle of converting all of the media files, then it would be best to simply remove the seconds step of converting from avi/mpeg during the capture process and stay with those formats.  

Oh and btw.... if any potential capturers ae out there both )p( and myself have capturing tools available.  His is integrated into 3darcade, and I have a as of yet unreleased one that does exactlyt he smae thing but is stand-alone.  

And yes that's a VERY big hint.  

Thanks for the info Dave, I'll still look into mng a bit more and think about it before I make a final decision.  At this point the audio is all I have done.  I start on the video today.