Inertia's picture

Eax, documentation and more

Since you don't join IRC anymore, I'll just ask here:

-did you spend any time on the Eax->Efx reverb conversion?
-the AudioContext class should probably expose the Alut.CreateBuffer* functions, else you will break all OpenAL samples that require loading a .wav from disk.
-i really think the naming should be consistent, either Audio (latin: "i hear") and Video (latin: "i see"), or use the english terms. not a mix inbetween. (the latin is from wikipedia :p )
-i've edited the geometry pages in the book to use 0.9.0 enums, should the part about GL.VertexPointer() get it's own page? (it's a bit hidden behind all the GL.MapBuffer() explanation)
-isn't it too early to show the russian translation in the menu already? there's very little translated so far, and noone has proofread it.
-attached is the SM 3.0+ julia set fragment shader, untested ofc. might crash&burn. It *might* work with higher MAXIteration values, if you set the 1D texture filter to linear in the app.

JuliaSet_SM3_FS.txt1.02 KB


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
the Fiddler's picture

I'd love to join IRC, but my internet access is down and my wireless proxy cuts the necesssary port.

You can always contact me on my email s_t_ap|ostol_@|g_mail.|com (remove the _ and | characters).

On the rest:
[Eax->Efx] and [.wav loaders]
Will do these during the weekend.

[Graphics vs Video]
Personally, video brings to mind "motion picture" while graphics brings "rendered picture". A namespace like OpenTK.Video.OpenGL (== Open Graphics Library) would look a little funny. Another solution would be to change Audio to Sound, but OpenTK.Sound.OpenAL sounds a little strange too (although I'd prefer that over OpenTK.Video).

Or we could just follow the conventions laid down by OpenGL and OpenAL and have Graphics and Audio.

The VBO page is growing a little long, it might be nice to split it up.


[Julia S&M, err SM3]
Is it just me, but doesn't SM look a little *too* close to S&M? Will test & add, thanks.

Edit: Tested, and it only displays a black screen. It compiles fine though, so I must be missing something obvious, I'll try to debug it. Btw, Ati's GPU ShaderAnalyzer is a great tool!

Inertia's picture

I know the address, but most of these topics might have gotten a 3rd opinion on IRC, or maybe here. If someone from russia proofreads the translation, my objection becomes obsolete.

If you could link the .lib into a .dll i'll do the conversion bit (found some EAX lowpass filter presets too i'd like to test out and include). I've tried doing this with Mingw and tcc, but with no success. It might be related to the .lib being compiled in VS6.0?

It would also be good if the Extension encapsulation is looked at from a different perspective too, with the AudioContext class it might be reasonable to instantiate them by default. The way they currently work was solely determined by easy development.

[Graphics vs Video]
I don't really want to argue about this, just wanted to point out that Graphics&Sound or Video&Audio would be more consistent. You're right, the latin is used in a more abstract sense. Maybe do it like the Pros and flip a coin? :p

Indeed, but I lack a better word to describe this hardware detail without citing GL_TOKENS

Inertia's picture

It may not work inside a Shader IDE due to the projection matrix setup, or maybe there's a link error? I've read that ATi hardware isn't very forgiving if you declare uniforms only in the FS, but not the VS - that could be an issue too. My remote debug skills aren't that far developed to figure this out, but functionality of the SM2 and SM3 fragment shaders should be identical from the GLSL logic.

Edit: Most likely it's the missing " #version 120 " at the top of the shader. The uniforms are probably not the problem, as they work in the SM2 shader. Don't spend too much time on this unless you do enjoy debugging shaders, I can still add the 2nd shader when my new rig arrives and the SM2 one does the job just fine for now, but I admit: the SM3 shader is alot less confusing because the loop isn't unrolled.

The shaderanalyzer looks very cool on the screenshot (especially the asm output), but crashes when I try to load the shader.

the Fiddler's picture

It probably needs an Ati card (although I just updated to the latest version, and it crashes here too).