Recent Posts

Pages: 1 [2] 3 4 ... 10
11
Thanks, I should be working with 22 KHz ADPCM in the following days.
It's really impressive since 22 KHz ADPCM is the same file-size as 44.1 KHz mp3 files.
Sure, half as compressed as mp3 files, but the Saturn can handle it.
12
Awesome work ponut64! Hopefully this will be able to make more people use the SBL ADPCM library now.
13
/e: Nevermind, I seem to have fixed it!

Regardless for anyone who wants to use ADPCM, here is an updated LIBPCM.A and an updated jo engine makefile.

For anyone curious as to what the problem was, there's no SBL source code for ADPCM system. There is however LIBADP, which contains the object code for said functions, and they are in fact linked in the code its just there's no source.

So I downloaded SaturnOrbit and simply included the .o files in LIBPCM.A -- you shouldn't need LIBADP.A anymore, but it is there in any case.

You put LIBPCM.A in
\joengine-master\Compiler\SGL_302j\LIB_COFF
Keep your old LIBPCM! Who knows what I have broken!
You also should put the following line in your Jo engine makefile
JO_COMPILE_USING_SGL = 1
**IT SHOULD BE NOTED THIS IS THE SBL PCM LIBRARY, NOT SGL
14
General Jo Engine Help / Re: CD-DA Audio in Jo is broken...?
« Last post by SaturnTeam on June 21, 2018, 08:58:53 pm »
Thanks for the tip. Maybe I'll check out that emulator.
15
Share your code / Re: Model converter (.ZTP) -0.1 - WIP
« Last post by XL2 on June 21, 2018, 07:53:08 pm »
No problem.
There is a learning curve to use it as it's not very user friendly, but it works quite well.
Try also to use the jo_malloc function to allocate the player data to high work ram (for faster performances).
Keep us updated, it's always great to see these tools and functions being used by others!
16
Share your code / Re: Model converter (.ZTP) -0.1 - WIP
« Last post by ponut64 on June 21, 2018, 07:43:19 pm »
Yeah, it's the normals (at least that part of the code).
I definitely kept (and updated.. or is that a thing you are supposed to do? The generated .h file from converter has an ANORM section, I copy that into an ANORM.H) the anorm file.

If I load it into your demo (even using that demos ANORM file), that doesn't happen. The issue has to be the scale then.

Now I just wonder, how can I get the model visible without clipping into it and it still being the correct size! That is something I can figure out. Thanks for helping me again!

/e: Oh wait. Your demos ANORM file is just the right one. So to anyone reading this thread. Use that anorm file. It worked even if I scaled the model. Sorry.
Everything is peachy now. Have a wonderful day. Any chance I could pass you lunch money over PayPal?  ;D

https://i.imgur.com/QalXkJm.png?1
17
Share your code / Re: Model converter (.ZTP) -0.1 - WIP
« Last post by XL2 on June 21, 2018, 07:25:42 pm »
Did you keep the anorm.h file?
Try to comment out the part of the code that updates the normals to see what's wrong.
If it still does it, then maybe the issue is the scale.
18
Share your code / Re: Model converter (.ZTP) -0.1 - WIP
« Last post by ponut64 on June 21, 2018, 07:18:18 pm »
I am assuming to fix the patchy-ness like this I'm going to have to increase the precision?
Or perhaps DUAL. Hmm.

https://i.imgur.com/4NGPF34.png?1

Really great so far though.

/e: DUAL attribute on the textures sort of gets around it
19
Share your code / Re: Model converter (.ZTP) -0.1 - WIP
« Last post by ponut64 on June 21, 2018, 06:41:49 pm »
Reducing the scale (incl. scale done in blender and in converter, but not scale in-engine) did put it back into shape. Thank you.
I guess that highlights the fact that I am running my crap on a pretty big scale, and it is OK to scale up the stuff in-engine for me.

It's not moving yet because I am a fool and somehow my OBJs are exporting as identical to one another. Note to self: Tick "Apply Modifiers", save headache.

Given how patchy the model is, I think I need to take its scale down even more...
20
Share your code / Re: Model converter (.ZTP) -0.1 - WIP
« Last post by XL2 on June 21, 2018, 05:42:01 pm »
It compresses animations (interpolation ) and vertices (12 to 6 bytes).
You should try to scale down your model to see if it works (I'm sure it will).
That's one issue since I only keep a 7.9 fixed point instead of 16.16.
In other words, if you want a different ratio, you will have to play with the source code.
You can't really avoid loss of precision, so it's all about choosing what you need.
Pages: 1 [2] 3 4 ... 10
SMF spam blocked by CleanTalk