338 replies to this topic

[The Loafer]

Why in the world don't you wait to hear back from them on your good comments first Tipoto? You never know, there could be a bug at work here. There is no hurry to fix, people just have to use an earlier build for now if they want to play your table.

Edited by The Loafer, 25 February 2013 - 09:39 AM.

[ClarkKent]

That's right! I also do not understand why the coders make this part of the code even more complex. I'm no programer myself but I think it would be wiser to fix these bugs once and forever instead of adding or changing some code to the patchwork...

 

I do not want to lose the incredible skills of Tipoto because of such things!

[Zablon]

n/m

Edited by Zablon, 25 February 2013 - 02:01 PM.

[toxie]

@Tipoto: In that case i would rather revert my changes i did in that revision, so unless you are already finished by now with the conversion..

That's right! I also do not understand why the coders make this part of the code even more complex. I'm no programer myself but I think it would be wiser to fix these bugs once and forever instead of adding or changing some code to the patchwork...

 

It's not that i do such things on purpose.

What we are dealing with in that special case is screwed up numerical precision of some of the floating point calculations within VP, not adding or removing "special code cases" that are more or less complex.

So i actually thought the new variant would be more robust, but as it turns out, it is not.

That's what the beta of VP915 is for: Testing and (hopefully) improving stuff..

[Aurich]

Not sure if this is off topic or relevant: Who's maintaing the ledcontrol.vbs file? An interesting tip in this thread:

 

http://www.vpforums....showtopic=23287

 

about adding a simple line of code (LedControl.FastCom=True) that apparently helps with some stutter issues. I haven't tested it myself, but if it's legit that's the kind of thing that everyone could benefit from, and I'm afraid will just get lost in a little thread.

 

In fact, that seems to be the biggest weakness of fixes right now, there isn't an easy way for 'normal' people who aren't obsessively reading the forum every day to have any clue about some of the great contributions the community makes, be it toxie or someone else.

[chepas]

There is no hurry to fix, people just have to use an earlier build for now if they want to play your table.

 

This is true, take your time, we can play 9.14 but with 445 it's a double fps increase on all tables for me. Probably the best VP build I've used so far.

[tipoto]

I fixed 33 textures and there are 169 in total, not all textures need a fix but most of them, and sometime it takes only 30 seconds for 1 texture, but  sometimes it takes 10 minutes or more.

I had to fix a few textures to try to understand what happened, but once I was in the table, I decided to continue because I was in the mood. It's hard to re-enter in the table after months, so once I'm on it, I take advantage of it.

But I can wait a little bit, to see if we can do something about it, I agree with the fact that it would be much better to definitively fix this bug for the future.

[JimofPinball]

I'm here and I am hoping that Visual Pinball 915 will run on e-machines without lag. The problem is when there are so many lights flashing that the frames skip, losing the fluid motion of the game and thereby ruining the pacing.

[Joe]

I'm here and I am hoping that Visual Pinball 915 will run on e-machines without lag. The problem is when there are so many lights flashing that the frames skip, losing the fluid motion of the game and thereby ruining the pacing.

what hardware do you have?

[Pinhead22]

T2 Chrome is one of the best tables offered....period! I think you guys are getting ahead of each other. Maybe some better communication between the development team and the table authors is warranted?

[Aurich]

With all due respect to tipoto, T2 is hardly a test case for how most people build tables. I'm all for him not having to redo all those textures of course though.

[Zablon]

Unless there's some pressing reason that T2 needs to run on the newer small increments, I say just leave it as is and people can run older versions with it. That's in Hyperpin now so it's not even an issue. There's no real reason for Tipoto to update it with every nightly

[tipoto]

No worries, I should have been less rough in my post I guess.

I know that with the EMreels there is no much visibility for the developers, because even when you think that it works, suddenly in a very specific case, their behavior will change.

 

I can wait before doing more work on the T2 and see if you figure the problem out. Hopefully you have enough handle with your tools to try some fixes.

Anyway, I'm available to test any version to fix together this problem which is painful even in the current official version.

[ClarkKent]

That's right! I also do not understand why the coders make this part of the code even more complex. I'm no programer myself but I think it would be wiser to fix these bugs once and forever instead of adding or changing some code to the patchwork...

 

It's not that i do such things on purpose.

What we are dealing with in that special case is screwed up numerical precision of some of the floating point calculations within VP, not adding or removing "special code cases" that are more or less complex.

So i actually thought the new variant would be more robust, but as it turns out, it is not.

That's what the beta of VP915 is for: Testing and (hopefully) improving stuff..

It was not my intention to offend somebody. I only thought if something is too complex and there is always something added to fix the misbehavior my thought was only to remove the whole thing and make it new...

 

By the way - will the transition to DX9 result in a working antialiasing with Windows7 x64?

[Pinball999]

Unless there's some pressing reason that T2 needs to run on the newer small increments, I say just leave it as is and people can run older versions with it. That's in Hyperpin now so it's not even an issue. There's no real reason for Tipoto to update it with every nightly

 

Talking about this. Blur mentioned a way to use Hyperpin with different versions of VP. Anyone here managed to have it working?

This would solve the problem we have with T2 Chrome and other tables.

Here is the method to use:

--------------------------------------------------------------------------------------------------------------------------

you need to create file vpexetables.txt inside settings directory inside hyperpin and in it you put table names without ,vpt you want to run with different exe, and in next line exe name

 

like this:

 

BlackHole0.30

vpinball908.exe

Amazing Spiderman bob 2008

vpinball904.exe

 

--------------------------------------------------------------------------------------------------------------------------

 

Cheers

[tipoto]

I don't know... I'm not sure it will fix the problem, because it means that anybody will need to know they have to use vp914 for this table, then only people with multiple VP setup will be able to launch the table with the appropriate VP. Also the table is already very restrictive with the fact you have to play it in 1920 x 1080 / 32 bit + antialias on the ball... It would add another constraint on top of that.

[skeleton60]

antialias on the ball correct some bugs?  

[Zablon]

Unless there's some pressing reason that T2 needs to run on the newer small increments, I say just leave it as is and people can run older versions with it. That's in Hyperpin now so it's not even an issue. There's no real reason for Tipoto to update it with every nightly

 

Talking about this. Blur mentioned a way to use Hyperpin with different versions of VP. Anyone here managed to have it working?

This would solve the problem we have with T2 Chrome and other tables.

Here is the method to use:

--------------------------------------------------------------------------------------------------------------------------

you need to create file vpexetables.txt inside settings directory inside hyperpin and in it you put table names without ,vpt you want to run with different exe, and in next line exe name

 

like this:

 

BlackHole0.30

vpinball908.exe

Amazing Spiderman bob 2008

vpinball904.exe

 

--------------------------------------------------------------------------------------------------------------------------

 

Cheers

 

Yes this works.

In your Hyperpin\Settings folder you create text files called vpexetables.txt and fpexetables.txt
Then for any table you want to use a different version of VP for you enter the table name (as in your xml) and on the line under it, the exe it should use.

 

It's that simple. You do not need to do every table, just the ones that won't be using your default Vpinball.exe.

I don't know... I'm not sure it will fix the problem, because it means that anybody will need to know they have to use vp914 for this table, then only people with multiple VP setup will be able to launch the table with the appropriate VP. Also the table is already very restrictive with the fact you have to play it in 1920 x 1080 / 32 bit + antialias on the ball... It would add another constraint on top of that.

 

The idea isn't to fix, it's to keep you from redoing the table with every release. It's a workaround until whatever version it actually gets fixed if it does. The problem with program updates is they do tend to break older stuff We have an easy workaround.

[fuzzel]

Hi,

some comments to the DX9 port: yes I'm working on that but it will take some time until you will all see a good running version, be patient

Why porting to DX9? Well porting from DX7 to DX9 is not that hard. Porting it to DX10+ would be a pain because the API has changed a lot from DX9 to 10+.

Another reason for DX9 is that a lot of virtual pins are running under WinXP (mine too ) and DX9 is the last DirectX version running on WinXP.

One day if we have removed all legacy code we can use pixel/vertex shaders adding new effects to tables, e.g. doing some nice environment mapping on each

ball, bump mapping and so on All that isn't possible with DX7. I assume that it will speed up the whole rendering because all graphic card drivers today

have better DX9 support.

 

...to be continued

fuzzel

[Zablon]

How realistic would it be to add the ability to modify properties (such as height position or color) for anything selected regardless of the type of object?