shapefusion-devel — Development-related list

Re: [Shapefusion-devel] sounds

[tiennou <tie...@gm...>]

Le 12 avr. 07 =C3=A0 00:11, Tito Dal Canton a =C3=A9crit :

> I have two more sound files that ShapeFusion fails to load: Evil and
> RED. Evil says "incorrect version/tag", RED says "incorrect group
> offset / total length". Can you look up these too?

I fixed RED file, caused by a negative groupOffset in sound 6. Wail =20
had the test I added, and strangely, Anvil don't complain when this =20
sound is opened, but it won't play it...
Anyway class 6 is 'Unused', so it won't break anything.
My EVIL sound file doesn't show this error... I just tried with the =20
Unimap version, but same... I don't get any error.

Your fix to the PlaySound() method causes an error here, but I don't =20
know what triggers it (I get a File not found, but the file is =20
clearly in /private/var/tmp/folders.501/TemporaryItems). I tracked =20
the error to the QuickTime call but now I'm wondering why it fails...

I've started working on a WadReader class, to provide abstract access =20=

to Wad data (use for Physics & Map). I didn't understood when you =20
told me that Physics are actually stored in Wad format, opposed to =20
Sound/Shapes.
I find this so strange that Bungie used Wad for some files, but not =20
for others...

tiennou

Re: [Shapefusion-devel] sounds

[Tito D. C. <ti...@da...>]

I have two more sound files that ShapeFusion fails to load: Evil and
RED. Evil says "incorrect version/tag", RED says "incorrect group
offset / total length". Can you look up these too?

Bye
T
-- 
Physics is reverse engineering

Re: [Shapefusion-devel] sounds

[<ti...@da...>]

> I have found this
> (http://www.borg.com/~jglatt/tech/aiff.htm). There   is
> some code at the end, but I wasn't able to make it work. I
> haven't   looked carefully though...
I've found everything within FFMPEG sources. It's all GPL so
we can freely use it. I'll do it, probably by introducing
another source file (MathUtils or something like that,
including code for doing IO to fixed fields).

T

Re: [Shapefusion-devel] sounds

[tiennou <tie...@gm...>]

Le 11 avr. 07 =C3=A0 16:32, Tito Dal Canton a =C3=A9crit :

> On Wed, 2007-04-11 at 13:08 +0200, tiennou wrote:
>
>
>> If you could have a look at my AIFF code, I don't know what I do
>> wrong, but Quicktime won't open it, and it crashes VLC...
> The sampleRate field is a 10-byte SANE extended field. I'd say INSANE
> instead. I don't know this format nor how to convert it to/from =20
> standard
> float, but I'm investigating.

I have found this (http://www.borg.com/~jglatt/tech/aiff.htm). There =20
is some code at the end, but I wasn't able to make it work. I haven't =20=

looked carefully though...

tiennou

Re: [Shapefusion-devel] sounds

[Tito D. C. <ti...@da...>]

On Wed, 2007-04-11 at 13:08 +0200, tiennou wrote:


> If you could have a look at my AIFF code, I don't know what I do
> wrong, but Quicktime won't open it, and it crashes VLC...
The sampleRate field is a 10-byte SANE extended field. I'd say INSANE
instead. I don't know this format nor how to convert it to/from standard
float, but I'm investigating.

T
-- 
Physics is reverse engineering

Re: [Shapefusion-devel] sounds

[tiennou <tie...@gm...>]

Le 11 avr. 07 =C3=A0 12:34, Tito Dal Canton a =C3=A9crit :

> On Tue, 2007-04-10 at 21:41 +0200, tiennou wrote:
>> I've implemented the SoundsHeader class, to handle converting to/from
>> System 7 Sounds format. For now I'm just accessing a BigEndianBuffer
>> as you did, because I don't want to mess with loading/saving such
>> things. Oh well, I guess I'll have to do it for importing, so...
> I'll commit some necessary fixes ASAP. Also, I think SoundsHeader =20
> should
> better be called System7SoundBlock, MacSound or something like that,
> since it contains everything (header+waveform). I can do this if you
> agree.

Arg, I just committed my stuff, but please do ! I've changed some =20
stuff to wave export (mainly putting ASCII values instead of =20
hexadecimal).
If you could have a look at my AIFF code, I don't know what I do =20
wrong, but Quicktime won't open it, and it crashes VLC...


tiennou

Re: [Shapefusion-devel] sounds

[Tito D. C. <ti...@da...>]

On Tue, 2007-04-10 at 21:41 +0200, tiennou wrote:
> I've implemented the SoundsHeader class, to handle converting to/from  
> System 7 Sounds format. For now I'm just accessing a BigEndianBuffer  
> as you did, because I don't want to mess with loading/saving such  
> things. Oh well, I guess I'll have to do it for importing, so...
I'll commit some necessary fixes ASAP. Also, I think SoundsHeader should
better be called System7SoundBlock, MacSound or something like that,
since it contains everything (header+waveform). I can do this if you
agree.

T
-- 
Physics is reverse engineering

Re: [Shapefusion-devel] sounds

[tiennou <tie...@gm...>]

Hi !

I fixed the problem with Rubicon sounds. It was caused by my loading  
code trying to load 16-bit sounds that didn't have permutations, but  
whose offsets where still present.
This one triggers the "8-bit/16-bit sounds differents" warning, and  
it works well (doesn't do anything yet).

I've implemented the SoundsHeader class, to handle converting to/from  
System 7 Sounds format. For now I'm just accessing a BigEndianBuffer  
as you did, because I don't want to mess with loading/saving such  
things. Oh well, I guess I'll have to do it for importing, so...

tiennou

Re: [Shapefusion-devel] sounds [was 0.3 release]

[Tito D. C. <ti...@da...>]

Here's the diff.

T
-- 
Physics is reverse engineering

Re: [Shapefusion-devel] sounds [was 0.3 release]

[Tito D. C. <ti...@da...>]

On Sun, 2007-04-08 at 19:51 +0200, tiennou wrote:


> Ok, I'll work on coding this. I'll be glad if you sent me your code,
> so I'll have a correct starting point.
Come on #alephone, I'm there. BTW I'm cleaning the code I just wrote,
then I'll send.

T
-- 
Physics is reverse engineering

Re: [Shapefusion-devel] sounds [was 0.3 release]

[tiennou <tie...@gm...>]

Le 8 avr. 07 =C3=A0 19:44, Tito Dal Canton a =C3=A9crit :

> On Sun, 2007-04-08 at 19:07 +0200, tiennou wrote:
>> I was thinking that I could write a class for SoundHeaders =20
>> (instead of
>> using BigEndianBuffer, that was to haste sound support), so we could
>> write input/output code from/to System 7 sounds... I'll write one
>> based on Aleph code...
> I'm experimenting right now. I wrote
> SoundsDefinition::SavePermutationToWAVE today, it parses the sound
> header in the BigEndianBuffer and writes a WAVE file. It works. =20
> However
> maybe your idea is better, since I'd like to provide the GUI with some
> feedback about what's actually inside each sound. You know, selecting
> the sound makes some text field pop up saying sample rate, etc. It =20
> would
> be nice. For detailed info about the sound header, download Inside
> Macintosh: Sound and see chapter 2. Let me know, I won't commit in the
> mean time.

Ok, I'll work on coding this. I'll be glad if you sent me your code, =20
so I'll have a correct starting point.

> Also, I'm realizing that broken/damaged sounds are not so rare between
> different scenarios. Rubicon sounds don't load at all. Can you take =20=

> care
> of this?

I had noticed this, but hadn't got time to investigate. I'll take =20
care of this too...

>
>> I'm not sure about embedded... IIRC, Physics files can be standalone
>> too ? I think we shouldn't bother about embedded Physics files yet,
>> they should be taken care of when we get to editing map files. I'll
>> look at the file loading classes you're talking about...
>>
>>
>> BTW, since I'm talking about map editing (and this is my ultimate
>> goal) :
>> I'm thinking that we should focus on editing raw
>> (non-precalculated/merged) map files before introducing full map file
>> support (Opening a map file and providing a 'Level' menu). I don't
>> want to go the Pfhorge way (redoing each map structs in Obj-C Data
>> classes), and I think "merging" was a good thing.
>> I also want to take a look at wx 3D capabilities for a visual mode,
>> but that's only for testing... I think the doc/view framework permits
>> having two 'views' : a visual/3d editor, and  a Grid view. What is
>> cool is that we should be able to move map vertexes in 3d view, and
>> provide real-time update on the Grid view, and the other way =20
>> around...
>> What to you think about this ?
> Well, honestly I think you are running too fast. My initial plan =20
> was to
> make a good Shapes editor (hence the name ;-) and I'm not really =20
> sure I
> want to include a map editor. A map editor is *way more complex* than
> our Shapes/Sounds class bunch, I won't have time for such a big thing.
> Think about that: terminals, geometry, platforms, light definitions,
> visual mode, merging, handling both AppleSingle / AppleDouble /
> MacBinary / separate forks... Of course, it would be *very* =20
> interesting,
> with 3D math and all that, but I can't maintain such a big =20
> creature. If
> ShapeFusion should become a map editor, we should better rename it to
> AlephOne Editor or something like that and we should bring other =20
> people
> to work with us. I prefer if we keep on creating an ideal Shapes/=20
> Sounds
> editor... For now.

Yes, you're right, I'm running too fast... I'll work on this on spare =20=

time, so I'll see what can be done (and how, because this is a really =20=

big task to handle)...

tiennou

Re: [Shapefusion-devel] sounds [was 0.3 release]

[Tito D. C. <ti...@da...>]

On Sun, 2007-04-08 at 19:07 +0200, tiennou wrote:
> I was thinking that I could write a class for SoundHeaders (instead of
> using BigEndianBuffer, that was to haste sound support), so we could
> write input/output code from/to System 7 sounds... I'll write one
> based on Aleph code...
I'm experimenting right now. I wrote
SoundsDefinition::SavePermutationToWAVE today, it parses the sound
header in the BigEndianBuffer and writes a WAVE file. It works. However
maybe your idea is better, since I'd like to provide the GUI with some
feedback about what's actually inside each sound. You know, selecting
the sound makes some text field pop up saying sample rate, etc. It would
be nice. For detailed info about the sound header, download Inside
Macintosh: Sound and see chapter 2. Let me know, I won't commit in the
mean time.

Also, I'm realizing that broken/damaged sounds are not so rare between
different scenarios. Rubicon sounds don't load at all. Can you take care
of this?



> I'm not sure about embedded... IIRC, Physics files can be standalone
> too ? I think we shouldn't bother about embedded Physics files yet,
> they should be taken care of when we get to editing map files. I'll
> look at the file loading classes you're talking about...
> 
> 
> BTW, since I'm talking about map editing (and this is my ultimate
> goal) :
> I'm thinking that we should focus on editing raw
> (non-precalculated/merged) map files before introducing full map file
> support (Opening a map file and providing a 'Level' menu). I don't
> want to go the Pfhorge way (redoing each map structs in Obj-C Data
> classes), and I think "merging" was a good thing.
> I also want to take a look at wx 3D capabilities for a visual mode,
> but that's only for testing... I think the doc/view framework permits
> having two 'views' : a visual/3d editor, and  a Grid view. What is
> cool is that we should be able to move map vertexes in 3d view, and
> provide real-time update on the Grid view, and the other way around...
> What to you think about this ?
Well, honestly I think you are running too fast. My initial plan was to
make a good Shapes editor (hence the name ;-) and I'm not really sure I
want to include a map editor. A map editor is *way more complex* than
our Shapes/Sounds class bunch, I won't have time for such a big thing.
Think about that: terminals, geometry, platforms, light definitions,
visual mode, merging, handling both AppleSingle / AppleDouble /
MacBinary / separate forks... Of course, it would be *very* interesting,
with 3D math and all that, but I can't maintain such a big creature. If
ShapeFusion should become a map editor, we should better rename it to
AlephOne Editor or something like that and we should bring other people
to work with us. I prefer if we keep on creating an ideal Shapes/Sounds
editor... For now.

T
-- 
Physics is reverse engineering

Re: [Shapefusion-devel] sounds [was 0.3 release]

[tiennou <tie...@gm...>]

Le 8 avr. 07 =C3=A0 13:19, Tito Dal Canton a =C3=A9crit :

> On Sun, 2007-04-08 at 11:26 +0200, tiennou wrote:
>
>> I'll be glad if you can handle this... I'm at a loss concerning file
>> formats for export. I know sounds are stored as System 7 sound file
>> (from Aleph sources), but I haven't found anything to read them after
>> export. Maybe export to AIFF too ?
> I already tried to feed a sound exported by ShapeFusion into baudline
> (my favourite signal analysis tool) and everything plays fine with the
> right parameters. I just have to extract the relevant info from the
> System 7 sound header and use it to write the WAVE headers.

I was thinking that I could write a class for SoundHeaders (instead =20
of using BigEndianBuffer, that was to haste sound support), so we =20
could write input/output code from/to System 7 sounds... I'll write =20
one based on Aleph code...

>> I'll look at Physics file handling too... Net players will
>> appreciate ;-).
> I think you better ask on the A1 list for opinions before... Let's =20
> know
> what people wants, then we may code it regardless ;-) Also keep in =20
> mind
> that physics data is not in a simple, separate file like sounds and
> shapes: it's embedded in the map, which is also a forked file. We'd =20=

> have
> to engineer a good framework for handling that. I've had a chat with
> treellama and he says we could benefit from the file loading classes
> he's introducing in A1. We should think about that expecially for =20
> things
> like physics (our shapes code rocks already ;-)

I'm not sure about embedded... IIRC, Physics files can be standalone =20
too ? I think we shouldn't bother about embedded Physics files yet, =20
they should be taken care of when we get to editing map files. I'll =20
look at the file loading classes you're talking about...

BTW, since I'm talking about map editing (and this is my ultimate =20
goal) :
I'm thinking that we should focus on editing raw (non-precalculated/=20
merged) map files before introducing full map file support (Opening a =20=

map file and providing a 'Level' menu). I don't want to go the =20
Pfhorge way (redoing each map structs in Obj-C Data classes), and I =20
think "merging" was a good thing.
I also want to take a look at wx 3D capabilities for a visual mode, =20
but that's only for testing... I think the doc/view framework permits =20=

having two 'views' : a visual/3d editor, and  a Grid view. What is =20
cool is that we should be able to move map vertexes in 3d view, and =20
provide real-time update on the Grid view, and the other way around...
What to you think about this ?

tiennou

[Shapefusion-devel] sounds [was 0.3 release]

[Tito D. C. <ti...@da...>]

On Sun, 2007-04-08 at 11:26 +0200, tiennou wrote:

> I'll be glad if you can handle this... I'm at a loss concerning file
> formats for export. I know sounds are stored as System 7 sound file
> (from Aleph sources), but I haven't found anything to read them after
> export. Maybe export to AIFF too ?
I already tried to feed a sound exported by ShapeFusion into baudline
(my favourite signal analysis tool) and everything plays fine with the
right parameters. I just have to extract the relevant info from the
System 7 sound header and use it to write the WAVE headers.

> I'll look at Physics file handling too... Net players will
> appreciate ;-).
I think you better ask on the A1 list for opinions before... Let's know
what people wants, then we may code it regardless ;-) Also keep in mind
that physics data is not in a simple, separate file like sounds and
shapes: it's embedded in the map, which is also a forked file. We'd have
to engineer a good framework for handling that. I've had a chat with
treellama and he says we could benefit from the file loading classes
he's introducing in A1. We should think about that expecially for things
like physics (our shapes code rocks already ;-)

Bye
T
-- 
Physics is reverse engineering

Re: [Shapefusion-devel] 0.3 release

[tiennou <tie...@gm...>]

Le 7 avr. 07 =C3=A0 23:12, Tito Dal Canton a =C3=A9crit :

> On Sat, 2007-04-07 at 20:39 +0200, tiennou wrote:
>
>> Done. The file ShapeFusion_0.3_OSX.zip is on Sourceforge. I'm trying
>> to download it, but I think mirror won't update until 1 hour. If you
>> can try on your side, I'll be glad to hear how it's going on your OSX
>> side (There shouldn't be any problem, but we never know ;-)).
> It works like a charm. I can't wait trying trunk and all its new
> features ;-)
>
>> I've tagged it in svn (read: moved it to the tags folder for future
>> reference), so we can keep an eye on it...
> I like this.
>
> Wee, we can now concentrate on 0.4. I'll keep on cleaning shapes code
> and implementing a preliminary color table editor. If you want I can
> also code a routine for saving sounds to WAVE format.

I'll be glad if you can handle this... I'm at a loss concerning file =20
formats for export. I know sounds are stored as System 7 sound file =20
(from Aleph sources), but I haven't found anything to read them after =20=

export. Maybe export to AIFF too ?

I'll look at Physics file handling too... Net players will =20
appreciate ;-).

tiennou

Re: [Shapefusion-devel] 0.3 release

[Tito D. C. <ti...@da...>]

On Sat, 2007-04-07 at 20:39 +0200, tiennou wrote:

> Done. The file ShapeFusion_0.3_OSX.zip is on Sourceforge. I'm trying
> to download it, but I think mirror won't update until 1 hour. If you
> can try on your side, I'll be glad to hear how it's going on your OSX
> side (There shouldn't be any problem, but we never know ;-)).
It works like a charm. I can't wait trying trunk and all its new
features ;-)

> I've tagged it in svn (read: moved it to the tags folder for future
> reference), so we can keep an eye on it...
I like this.

Wee, we can now concentrate on 0.4. I'll keep on cleaning shapes code
and implementing a preliminary color table editor. If you want I can
also code a routine for saving sounds to WAVE format.

T
-- 
Physics is reverse engineering

Re: [Shapefusion-devel] 0.3 release

[tiennou <tie...@gm...>]

Le 6 avr. 07 =C3=A0 18:52, Tito Dal Canton a =C3=A9crit :

> I've released source for 0.3. You can edit the release and add the =20
> MacOS
> X binary. Please name it ShapeFusion_0.3_OSX.zip to keep consistency.
> I'll post to the A1 list when you're done.

Done. The file ShapeFusion_0.3_OSX.zip is on Sourceforge. I'm trying =20
to download it, but I think mirror won't update until 1 hour. If you =20
can try on your side, I'll be glad to hear how it's going on your OSX =20=

side (There shouldn't be any problem, but we never know ;-)).

> What about the 0.3 svn branch now? Should it be deleted?

I've tagged it in svn (read: moved it to the tags folder for future =20
reference), so we can keep an eye on it...

tiennou

Re: [Shapefusion-devel] 0.3 release

[Tito D. C. <ti...@da...>]

I've released source for 0.3. You can edit the release and add the MacOS
X binary. Please name it ShapeFusion_0.3_OSX.zip to keep consistency.
I'll post to the A1 list when you're done.

What about the 0.3 svn branch now? Should it be deleted?

Bye
Tito
-- 
Physics is reverse engineering

[Shapefusion-devel] indexed bitmaps and LittleEndianBuffer

[<ti...@da...>]

I've implemented exporting bitmaps to indexed BMP files.
Since I couldn't find how to do this using wx, I've created
a new ShapesBitmap method which directly writes in BMP
format to a file. Everything seems to work fine and
exporting bitmap masks is now straightforward to code. I'll
commit ASAP.

To aid in direct BMP access I've also created a new class,
LittleEndianBuffer, with obvious purpose. You can use it to
write sounds to WAVE files.

Bye
Tito

[Shapefusion-devel] exporting bitmap masks [Fwd]

[Tito D. C. <ti...@da...>]

I'll take care of this...

-------- Forwarded Message --------
From: Blayne Scott <bla...@gm...>
To: ti...@da...
Subject: ShapeFusion
Date: Sat, 31 Mar 2007 18:28:42 -0400

Hi Tito,

I'm the lead developer in a Marathon scenario titled "Where Monsters
are in Dreams". Treellama suggested I ask you this question about your
OSX /  Windows app "ShapeFusion":

While Shapefusion can do a full export of the shapes file as BMP-24,
are there any plans to modifying it to add the mask feature as well?

I'm in the process of converting every frame of animation into
MML-anchored .PNG's.

Although, the work of generating a Black /  white Alpha map for each
frame by hand is enormous. The Marathon engine automatically selected
the first color table slot (Blue) as the mask. Any hope of having an
automated separation of that color into a separate mask file?

Thanks in advance,
-Blayne Scott
http://bighouse.bungie.org/wmaid/

-- 
Physics is reverse engineering

Re: [Shapefusion-devel] 0.3 release

[tiennou <tie...@gm...>]

Le 31 mars 07 =C3=A0 12:17, Tito Dal Canton a =C3=A9crit :

> On Fri, 2007-03-30 at 20:39 +0200, tiennou wrote:
>> I'm testing a full build (Development & Deployment) before I commit.
>> No problem, I just got prebinding disabled because wx overlaps with
>> Shapefusion...
>> tiennou
> Fine, I'll release the sources ASAP when you succeed in building the
> binary. Let me know. You'll be able to add the binary without problems
> then.

Then you can go, I have a working Shapefusion UB (Works on my MBP, =20
and on my iBook).
I'm not sure about 10.3 though, because it's there that prebinding is =20=

used, but I don't have a test system with this... Anyway I think OS X =20=

can manage prebinding issues by itself.

tiennou

Re: [Shapefusion-devel] Sound File support

[tiennou <tie...@gm...>]

Le 31 mars 07 =C3=A0 12:31, Tito Dal Canton a =C3=A9crit :

> What about adding a frame around the Flags checkboxes, like the Volume
> radio buttons? It should look better. Also I'd add a bit more space
> around controls, because under Linux they look all packed together. I
> can do this.

I'll put a Static Box around the Flags, that's right it should look =20
better... As for space between controls, things look good on OS X, =20
but some more spacing can't be bad ;-)

tiennou

Re: [Shapefusion-devel] Sound File support

[Tito D. C. <ti...@da...>]

What about adding a frame around the Flags checkboxes, like the Volume
radio buttons? It should look better. Also I'd add a bit more space
around controls, because under Linux they look all packed together. I
can do this.

Bye
T
-- 
Physics is reverse engineering

Re: [Shapefusion-devel] new wx features

[Tito D. C. <ti...@da...>]

On Fri, 2007-03-30 at 20:48 +0200, tiennou wrote:
> Strange, I got it there : http://wxwidgets.org/manuals/stable/ 
> wx_wxtextctrl.html#wxtextctrlchangevalue
> I used this because http://wxwidgets.org/manuals/stable/ 
> wx_wxtextctrl.html#wxtextctrlsetvalue is marked as deprecated...
Oh, very interesting! This exactly fixes the problem with
wxEVT_COMMAND_TEXT_UPDATED in my shapes code! Too bad it's "new since
wxWidgets version 2.7.1". I really need to upgrade, I checked around and
wx has improved a lot since my current version. But even the latest
Ubuntu has quite an old version (don't think it's newer than 2.7.1). I
have just installed a desktop with the latest Ubuntu, I'll try. I'll
consider compiling the latest wx on it if that's needed.

> Feel free to change my code if you have a better way (or if you can't  
> use ChangeValue at all...)
I've included an #ifdef for now, so I can at least build.

> Hmm I think this is wx deprecating identifiers... I think you can  
> replace those with wxSAVE...
#ifdef'ed this too.

> Sorry, I forget every time... What a pain to add those macros  
> everywhere. I'm wondering why I don't get any errors about them when  
> I build, whereas they drive you crazy...
I'm not surprised, it depends on a lot of factors. Moreover, I'm
paranoid and I build with -Wall which triggers every sort of warnings.
That's why I complain so much about new code ;-)

Bye
T
-- 
Physics is reverse engineering

Re: [Shapefusion-devel] 0.3 release

[Tito D. C. <ti...@da...>]

On Fri, 2007-03-30 at 20:39 +0200, tiennou wrote:
> I'm testing a full build (Development & Deployment) before I commit.
> No problem, I just got prebinding disabled because wx overlaps with
> Shapefusion...
> tiennou
Fine, I'll release the sources ASAP when you succeed in building the
binary. Let me know. You'll be able to add the binary without problems
then.

T
-- 
Physics is reverse engineering