Release Name: 6.0
Notes:
Short: Retargetable audio v6.0, Developer's Archive
Type: dev/misc
Author: Martin Blom
Uploader: Martin Blom
This archive contains the developer documentation and
example source code for AHI, both for programmer who wish
to use AHI in their own applications and those who want
to write hardware drivers.
Changes since last release
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
* Anti-click mixing routines, which makes AHI search for a
zero-crossing before playing a new note
* Playback of 32 bit samples.
* Multi-channel audio. Eight channels are supported.
* Support for four more architectures: MorphOS/PPC, AmigaOS4/PPC,
Amithlon/ix86, AROS/ix86.
* Support for PowerPC accelerator cards (MorphOS and AmigaOS4
only).
* A new driver, device.audio, can be used to make all low-level AHI
applications, which normally open AHI in exclusive mode, behave
as if they used the high-level API which allows several AHI
applications to work simultaneously. (Note that this driver
currently does not work with applications that require less than
20 ms latency.)
* The 68000 processor is no longer supported; a 68020 or better is
required.
For a full log of everything that has been updated, see the
ChangeLog file.
--------------------------------------------------------------------------------
TITLE
AHI Audio system
VERSION
Release 6.0
AUTHOR
Martin Blom
DESCRIPTION
Retargetable Audio for AmigaOS
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
(When refering to this software, the correct term is
'AHI audio system' or just 'AHI', never 'Audio Hardware
Interface'! 'AHI' is pronounced 'atchii', as in 'God bless!'.)
AHI now supports the PowerPC processor. Thanks to the users who
donated money and Syscom & Mematex AB who sponsored the hardware,
I was finally able you buy a Cyberstorm PPC card. The current
version is for MorphOS only, but a WarpUp version will hopefully
follow later.
Version 6.0 supports 68020, 68030, 68040 and 68060 from
the m68k family and 603 and 604 from the PowerPC family.
Staring with version 6, the 68000 processor is no longer supported.
68000 owners should use version 4 instead.
Quick overview
¯¯¯¯¯¯¯¯¯¯¯¯¯¯
* Driver based
Each supported sound card is controlled by a library-based audio
driver. For a 'dumb' sound card, a new driver should be written in
a few hours. For a 'smart' sound card, it is possible to utilize an
on-board DSP, for example, to maximize performance and sound quality.
For sound cards with own DSP but little or no memory, it is possible
to use the main CPU to mix channels and do the post-processing
with the DSP. Available today are drivers for
· Aura (sampler only)
· Clarity (sampler only)
· Concierto (Picasso IV sound module)
· Delfina DSP and Delfina Lite
· DraCo Motion
· Maestro Pro
· Melody Z2 and 1200
· Paula (8/14/14c bit)
· Prelude and Prelude 1200
· Toccata
· Wavetools
· 8SVX (mono) and AIFF/AIFC (mono & stereo) sample render
· ...and more.
* Fast, powerful mixing routines (yeah, right... haha)
The device's mixing routines mix 8- or 16-bit signed samples, both
mono and stereo, located in Fast-RAM and outputs 32-bit mono or stereo
(with stereo panning if desired) data, using any number of channels
(as long as 'any' means less than 128...). Tables are no longer
used, which might affect performace slightly. Instead, version 6
forcuses slightly more on the quality of the sound.
* Support for non-realtime mixing
By providing a timing feature, it is possible to create high-
quality output even if the processing power is lacking, by saving
the output to disk, for example as an IFF AIFF or 8SXV file.
There are so-called HiFi mixing routines that can be used, which
use linear interpolation and gives 32 bit output.
* Audio database
Uses ID codes, much like Screenmode IDs, to select the many
parameters that can be set. The functions to access the audio
database are not too different from those in 'graphics.library'.
The device also features a requester to get an ID code from the
user.
* Both high- and low-level protocol
By acting both like a device and a library, AHI gives the programmer
a choice between full control and simplicity. The device API allows
several programs to use the audio hardware at the same time, and
the AUDIO: dos-device driver makes playing and recording sound very
simple for both the programmer and user.
* Future Compatible
When AmigaOS gets device-independent audio worth it's name, it should
not be too difficult to write a driver for AHI, allowing applications
using 'ahi.device' to automatically use the new OS interface. At
least I hope it wont.
Funny how things turned out, isn't it? Sigh.
SPECIAL REQUIREMENTS
AmigaOS release 2.
MC68020 processor or better.
AVAILABILITY
Aminet, for example:
ftp://ftp.germany.aminet.org/pub/aminet/dev/misc/ahi6dev.lha
ftp://ftp.germany.aminet.org/pub/aminet/dev/misc/ahi6src.lha
ftp://ftp.germany.aminet.org/pub/aminet/mus/misc/ahi6usr.lha
WWW:
http://www.lysator.liu.se/~lcs/ahi.html
PRICE
Free. As in speech. And beer too! However, should you want to
make a donation, my PayPal account is martin@blom.org.
DISTRIBUTABILITY
Copyright ©1994-2005 Martin Blom.
AHI version 6 is (L)GPL:ed. See COPYING and COPYING.LIB inside the
archive for more information.
Changes:
2005-11-20 Martin Blom
* Released AHI 6.0. w00t!
2005-11-06 Martin Blom
* Docs/readme.changes: Added note about device.audio's 20 ms
limit. [Ian Greenway]
2005-10-09 Martin Blom
* Drivers/Common/library.h: Avoid using assembly symbol
redirection in AROS.
2005-10-05 Martin Blom
* Docs/readme.changes, Docs/readme.usr: Updated.
2005-10-01 Martin Blom
* COPYING.DRIVERS: Fixed a typo and added a small plea for open
source drivers.
2005-09-28 Martin Blom
* Bumped just about every single revision counter ...
* Drivers/Common/Makefile.common.in: Arranged to that the
"library.o" object comes first when linking, so that the _start()
function is at the very beginning of the driver. [Gunther Nikl]
* Drivers/Common/library.c: Made rawputchar_m68k const. [Gunther
Nikl]
* AHI-Handler: Don't compile debug support code if not debug
build. Made rawputchar_m68k const. [Gunther Nikl]
* AHI Preferences, AddAudioModes, AHI-Handler: Define __main() for
libnix to make the binaries smaller. [Gunther Nikl]
2005-09-2 Martin Blom
* configure.in: Use AC_SEARCH_LIBS so that -lauto, -lraauto and
-lmui only gets added if required. That fixes a potential
ixemul/libnix conflict for m68k-amigaos. [Gunther Nikl]
* configure.in: Don't add include path /gg/os-include per
default. [Gunther Nikl]
* devcommands.h: Removed duped prototype for FindNode(). [Gunther
Nikl]
2005-09-27 Martin Blom
* Released AHI 6.0rc4.
* Released ahi.device 6.6.
* ahi.device: Fixed a long-standing bug, first reported on
2002-08-17, about an enforcer hit when AbortIO()'ing an I/O
request. Thanks for the excellent test program, Harry! [Marcin
Kurek, Sensei, Harry Sintonen]
2005-09-22 Martin Blom
* Released AddAudioModes 6.2.
* Released ahi.device 6.5.
* Released ac97.audio 6.2.
* Released AROS.audio 6.2.
* Released device.audio 6.1.
* Released emu10kx.audio 6.4.
* Released filesave.audio 6.2.
* Released void.audio 6.2.
* Docs/ahiusr.texinfo: Added Gunther Nikl and Davy Wentzler to
"Contributors". Removed all contributor e-mail addresses.
* ahi.device, Drivers/*: More OS4 porting.
2005-09-21 Martin Blom
* Imported AROS updates to aros.audio.
2005-09-20 Martin Blom
* Added build environment support for OS4.
* AHI Preferences (ClassAct): Ported to Reaction. At least it
compiles without errors. If I had libreaction.a and libraauto.a
for clib2/OS4 it would even link.
* ahi.device: Now builds for OS4 using the OS4 cross compiler,
the OS4 SDK from GoldED and sfdc 1.4.
* AddAudioModes: Ported to OS4.
2005-09-11 Martin Blom
* Bumped revision of ahi.device to 6.4.
* Bumped revision of AHI Preferences to 6.1.
* Bumped revision of AHI-Handler to 6.1.
* Bumped revision of AddAudioModes to 6.1.
* Bumped revision of ac97.audio to 6.1.
* Bumped revision of aros.audio to 6.1.
* Bumped revision of emu10kx.audio to 6.3.
* Bumped revision of filesave.audio to 6.1.
* Bumped revision of void.audio to 6.1.
* Rebuilt using a gcc that does not have emulated instructions in
libgcc.a.
2005-09-11 Martin Blom
* Released emu10kx.audio 6.2.
* emu10kx.audio: OS4 version use 80 Hz interrupt instead of 1
kHz. [Davy Wentzler]
2005-08-11 Martin Blom
* Documentation: Added PDF versions of the User's And Developer's
guides. [Thor]
2005-06-19 Martin Blom
* Released toccata.audio 4.7.
* ahiusr.texinfo: Added notes about AHItoccataPlayBufferSize and
AHItoccataRecordBufferSize.
* toccata.audio: Added the variables AHItoccataPlayBufferSize and
AHItoccataRecordBufferSize.
2005-04-13 Martin Blom
* AHI preferences: Added updated german translation by Stefan
Bojartschuk.
2005-04-06 Martin Blom
* ahi.device: Harry's ixemul/SIGBREAKF_CTRL_F fix seem to have
fixed the infamous DisplayBeep()/IPrefs/sound.datatype/whatever
bug that several people have reported. This must be my lucky week.
[Luca "Hexaae" Longone]
2005-04-05 Martin Blom
* Device/header.c: The __attribute__ is back again. [Davy
Wentzler]
* ahi.device/concierto.audio: Oh, I forgot to mention: That
PreTimer/PostTimer fix Gunther helped with in January with made
the Concierto driver work again! Fett med credz till Gunther!!
2005-04-03 Martin Blom
* ahi.device: The copyright year accidentally went back to 2004
when merging Gunther's patch. Fixed.
* Released AHI 6.0rc3.
* Released ahi.device 6.2.
* ahi.device: Removed the beta requester from the Amithlon build.
* ahi.device: Applied an old clean-up patch from Gunther Nikl to
header.c.
2005-03-29 Martin Blom
* Released ahi.device 6.1.
* Device/header.c: Instead removed the 'static' keyword on RomTag
and undid the __attribute__ fix to avoid a warning when using gcc
2.x.
* ahi.device: Applied a patch from Piru that removed the usage of
SIGBREAKF_CTRL_F in ahi.device, working around a problem in the
latest MorphOS ixemul.library. Thanks! [Harry Sintonen]
2005-03-26 Martin Blom
* Device/header.c: Added __attribute__((used)) to RomTag. [Davy
Wentzler].
* Released emu10kx.audio 6.1.
* emu10kx.audio: Added write-only (for now) access to internal
DSP volume registers via the AC97 mixer interface.
* configure.in: Temporarily added -msoft-float to OS4 compiler
flags.
* ac97.audio: Bumped version number to 6.0.
* ac97.audio: Added this AROS driver to the distribution, with
permission from its author Michal Schulz. Thanks Michal!
* emu10kx.audio: Added small patch from Davy.
2005-03-24 Martin Blom
* emu10kx.audio: Removed unnecessary OS4-only interrupt disable
code.
2005-03-21 Martin Blom
* emu10kx.audio: Applied a fix from Davy for the recording problem
on Audigy cards. Thanks!
2005-02-02 Martin Blom
* Updated all copyright years to 2005.
* Bumped all version numbers to 6.0.
* Docs/ahidev.texinfo (The Author): Updated the authors address (5
years later).
* Added updated/new italian translations by Paolo Besser. Thanks!
* AHI preferences: Update the strings in the unit cycle gadget
when loading new prefs from disk or restoring to the
defaults. [Gunther Nikl]
2005-01-28 Martin Blom
* Updated the docs and .readme files a bit.
2005-01-23 Martin Blom
* ahi.device: That CPU overload protection (PreTimer/PostTimer)
thing didn't work ... at all. Duh. [Gunther Nikl]
2005-01-14 Martin Blom
* Updated the README files. The language source files are now
included in the ahidev archive.
2005-01-13 Martin Blom
* AHI preferences: Localized the anti-click time and CPU load
values as well. [Michael Malyshev]
* Added updated Russion translations by Michael Malyshev. Thanks!
2005-01-12 Martin Blom
* The binary user archive now includes the .ct files for easy
translation updates.
* Updated to FlexCat 2.6.6-cvs-2005-01-12 and regenerated language
files.
* Released AHI preferences 5.15.
* Released ahi.device 5.56.
* Added 'FILL' switch to FlexCat when generating the catalogs in
order to make missing strings use the built-in English defaults
instead of nothing. [Pierre Giroux]
* Localized Hz, dB and the number of channels. [Michael Malyshev]
* Only try to build emu10kx.audio for AROS if is found.
2005-01-11 Martin Blom
* filesave.audio: Fixed a compiler warning. (No bumprev.) [Gunther
Nikl]
* AHI preferences: Used floating point instruction introduced by
the -m68020-60 switch by accident. Added -msoft-float for m68k to
prevent this. [Gunther Nikl]
* ahi.device: Tweaked the m68k gateway stubs so no nops are used
anymore. [Gunther Nikl]
* Applied more patches (mainly gcc 3.x-related but also
C_c_noauto.sd and gas stab) from Gunther Nikl. Thanks!
2005-01-09 Martin Blom
* Released AHI 5.15b (ix86be only).
* Released ahi.device 5.55.
* ahi.device: Minor modifications to make the ix86be version build
again, this time using a non-broken version of the compiler. :-)
2005-01-08 Martin Blom
* Released AHI 5.15.
* Released ahi.device 5.54.
* Released AHI preferences 5.14.
* AHI preferences: Moved the selection of scale modes to the
global, advanced prerference page. Changed the default to "Full".
* ahi.device: Now used global scale mode settings. Changed the
default to AHI_SCALE_FIXED_0_DB.
* ahi.device: Tweaked the default "dizzy" tags for BestAudioID()
to award the user's preferred audio mode. [Fab]
2004-11-08 Martin Blom
* Released AHI-Handler 5.5.
* AHI-Handler: Fixed bad IFF chunk lengths for AIFF/AIFC
files. [Patch from Stefan Haubenthal]
2004-11-03 Martin Blom
* Released AHI-Handler 5.4.
* AHI-Handler: Fixed bad struct alignment for non-m68k
builds. [Stefan Haubenthal]
* Released ahi.device 5.53.
* ahi.device: 16 bit samples were broken in 7.1 mode. [Marcin
'Morgoth' Kurek]
2004-11-02 Martin Blom
* Released ahi.device 5.52.
* ahi.device: DSP echo and master volume did not work in 7.1
mode. Fixed now.
* Updated copyright year and removed almost all CVS/RCS tags from
the source code.
2004-11-01 Martin Blom
* Released AddAudioModes 5.6.
* Applied a couple of patches from Gunther Nikl. Thanks Gunther!
2004-10-31 Martin Blom
* Released AHI 5.14.
* Imported the rest of the AROS modifications from AROS
CVS. Mostly Makefile mods.
* Released emu10kx.audio 5.22.
* emu10kx.audio: Reduced the volume boost by 6 dB. According to my
voltmeter, the driver now outputs around 2 Vpp when the master
volume is set to 0 dB and around 8 Vpp when the master volume is
+12 dB. If you have a real oscilloscope, I'd be happy to hear your
verification.
* emu10kx.audio: Merged the AROS port by Georg Steger. Thanks
Georg!
2004-10-29 Martin Blom
* Released emu10kx.audio 5.21.
* emu10kx.audio: Made the +12 dB boost work on Audigy too. [Marcin
'Morgoth' Kurek]
2004-10-28 Martin Blom
* Released emu10kx.audio 5.20.
* emu10kx.audio: Added front channel downmix to center and LFE
when playing mono or stereo sounds (-3 dB), just like the driver
used to do for Audigy cards before. This is disabled in
multichannel mode. [Marcin 'Morgoth' Kurek]
* emu10kx.audio: Boosted AHI output by 12 dB (!).
2004-10-27 Martin Blom
* Released ahi.device 5.51.
* Device/devsupp.c: I obviously have a nasty habit of not testing
the code I write. "for (i = 0; i < size; ++size)"?! Yeah,
right. And you know what's REALLY scary? It's been like that since
late 2001! Anyway, recording in device mode or via AUDIO: should
now work again. [Vladimir Javorski, Denis Markovic, Alexey Ivanov,
Stefan Haubenthal]
2004-10-23 Martin Blom
* Docs/ahiusr.texinfo: Fixed some typos. [Stefan Haubenthal]
2004-09-20 Martin Blom
* ahi.device: Minor bugfixes from Davy.
* emu10kx.audio: Added an OS4 reset handler patch from Davy.
2004-09-08 Martin Blom
* Docs/ahiusr.texinfo: Added description of the default anti-click
time. [Luca 'Hexaae' Long]
2004-08-13 Martin Blom
* Added an updated german ahiprefs translation. [Stefan
Bojartschuk]
2004-08-07 Martin Blom
* emu10kx.audio: Applied even more OS4 patches from Davy
Wentzler. Thanks!
2004-08-06 Martin Blom
* ahi.device: Applied even more OS4 patches from Davy
Wentzler. Thanks!
2004-08-05 Martin Blom
* Released emu10kx.audio 5.18.
* emu10kx.audio: Remade the routing once again. Tried to make the
routing code work on Audigy. Testing required.
2004-08-04 Martin Blom
* Added the string "\nAnnotation" to all prefs catalogs'
msgProperties string. Real translations are more than
welcome. [Michael Malyshev]
* Pass CPU option to linker when building. [Gunther Nikl]
2004-08-02 Martin Blom
* Added updated Russion translations by Michael Malyshev. Thanks!
* emu10kx.audio: Don't print debug info if no cards found. [Harry
Sintonen]
2004-07-27 Martin Blom
* Released AHI 5.13.
* Released emu10kx.audio 5.17.
* Released ahi.device 5.49.
* ahi.device: Added backward mixing in 7.1 modes.
* ahi.device: Fixed a couple of bugs in the 7.1 code.
* emu10kx.audio: Replaced the EMU10k1 routing completely. Disabled
most of the patch manager. Somebody needs to fix it for Audigy
cards.
* emu10kx.audio: Added 5.1 multichannel support for Live!
cards. And there was much rejoicing.
2004-07-26 Martin Blom
* Examples/Low-level/Test-7.1: Added a 7.1 test example.
* AROS, DEVICE, EMU10KX, FILESAVE, PAULA, TOCCATA, VOID,
WAVETOOLS (mode files): Enabled panning for all mono modes.
* Released void.audio 5.8.
* void.audio: Added a comment about locking the actual hardware in
AHIsub_AllocAudio().
* void.audio: Added 7.1 "support".
2004-07-25 Martin Blom
* Released AHI preferences 5.13.
* Released ahi.device 5.48.
* AHI preferences: Added selection of volume scaling mode. And
there was much rejoicing.
* ahi.device: Implemented a couple of alternatives to that
horrible "safe scaling" thing. [Daniel Westerberg, Hubert Maier
and a thousand others]
2004-07-07 Martin Blom
* emu10kx.audio: AHIDB_MaxRecordSamples returned twice the number
of samples actually used. [Davy Wentzler]
2004-06-26 Martin Blom
* Released AHI 5.12.
2004-06-12 Martin Blom
* ahi.device: Avoid memory trashing when opening AHI_NO_UNIT using
a V2-sized AHIRequest. [Harry Sintonen]
2004-06-09 Martin Blom
* Released device.audio 5.1.
* device.audio: Clear audio buffers before playing the first time
(removes the noise in the beginning).
2004-06-08 Martin Blom
* Added 7.1 multichannel support. And there was much rejoicing.
* Added missing TAG_DONEs to ScanAudioModes.
2004-06-04 Martin Blom
* Forked off BRANCH_6.
2004-06-02 Martin Blom
* Released AHI preferences 5.12.
* AHI: Added a not-so-beautiful hack to make sure modes from
device.audio modes does not show up in the Unit mode lists.
* Released device.audio 5.0.
* Added a new driver, device.audio, that plays sound using one of
AHI's units. And there was much rejoicing.
2004-06-01 Martin Blom
* Added the good old ScanAudioModes to the example collection.
2004-02-19 Martin Blom
* Docs: Added note about AddAudioModes and sound card/AC97
initializing.
2004-02-14 Martin Blom
* ahi.device: I forgot to update dspechofuncs_68k.s when removing
the need for gasp. Fixed.
* ahi.device: Applied a new batch of AmigaOS 4.x pathches from
Davy. Thanks!
2004-02-09 Martin Blom
* Released filesave.audio 5.5.
* filesave.audio: Applied AOS4 patches from Davy Wentzler.
2004-02-04 Martin Blom
* Released aros.audio 5.1.
* aros.audio: Added "100 Hz" patches from Georg Steger.
2004-01-31 Martin Blom
* aros.audio: Minor OSS tweaks.
2004-01-25 Martin Blom
* Released emu10kx.audio 5.16.
* emu10kx.audio: Applied AmigaOS 4 patches (with minor
modifications) from Davy Wentzler.
* ahi.device: Removed gasp dependency. gasp is no longer required
for building AHI.
* Drivers: Applied patches to make mode file generation work
without using (the now obsolete tool) gasp. [Fabio Alemagna]
2004-01-18 Martin Blom
* Drivers: Removed faulty alignment at the end of the mode files.
* Device/audioctrl.c (_AHI_AllocAudioA): Changed 'case NULL' to
'case 0'.
2004-01-17 Martin Blom
* The top Makefile called 'make' directly and didn't use
$(MAKE). Fixed.
2003-12-22 Martin Blom
* Released aros.audio 5.0.
* Added an AROS driver based on oss.library.
2003-11-28 Martin Blom
* Released emu10kx.audio 5.15.
* Clear DMA buffers when they are allocated, in case they
are accidentally played before filled (GREX problem??)
* emu10kx.audio: Detect GREX at loading time, not on every
interrupt.
2003-11-24 Martin Blom
* Released emu10kx.audio 5.14.
* emu10kx.audio: Added GREX support (apparently, it has no bus
snooping)
2003-11-23 Martin Blom
* Released emu10kx.audio 5.11, 5.12 and 5.13.
* emu10kx.audio: Resynced with the latest Linux drivers in order
to make it work on Audigy cards.
2003-10-31 Martin Blom
* Released emu10kx.audio 5.10.
* emu10kx.audio: Don't show a requester if no sound card can be
found. [Nicolas Sallin]
2003-10-16 Martin Blom
* Released paula.audio 4.31.
* paula.audio: In MorphOS, audio.device is no longer used as a
lock for the audio hardware. Instead, an internal flag is used
2003-09-30 Martin Blom
* Released ahi.device 5.46.
* Obviously, I left a KPrintF in the code ...
* Released ahi.device 5.45.
* ahi.device: Test: Don't increase volume of playing requests
untill all outstanding requests are finished. Should fix that
annoying volume fluctuations?
2003-09-29 Martin Blom
* Released ahi.device 5.44.
* ahi.device: Fixed a memory trashing bug introduced with the
dynamic volume update (ahi.device tried to play a sound on channel
65535.) [Kev Harrison and Michael Malyshev]
* Released void.audio 5.7.
* Released filesave.audio 5.4.
* Released emu10kx.audio 5.9.
* The introduction of --addvectors made a bug in the driver's init
code visible. Fixed. [Marcin Kurek]
* The MorphOS versions of the different modules now use
RTF_EXTENDED and rt_Revision.
2003-09-28 Martin Blom
* Released emu10kx.audio 5.8.
* emu10kx.audio: Added a public semaphore with hooks to control
the AC97 mixer.
2003-09-22 Martin Blom
* Applied initial AmigaOS4 patches from Davy Wentzler.
* Does not use CompilerSpecific.h anymore.
* Now uses the sfdc switch --addvectors instead of ugly
internal .sfd files.
2003-07-30 Martin Blom
* Bumped revision of emu10kx.audio to 5.7.
* Bumped revision of filesave.audio to 5.3.
* Bumped revision of void.audio to 5.6
* Bumped revision of AHI-Handler to 5.3.
* Bumped revision of AddAudioModes to 5.5.
* Released AHI preferences 5.11.
2003-07-28 Martin Blom
* Released ahi.device 5.42.
* ahi.device: Re-disabled the powerpc assembler optimized mixing
routines, since they're totally fubar. I'll look into it until
next release, I promise ...
* Rebuilt everything with an updated GG-cross build enviroment and
using the new SDK for the MorphOS version.
2003-07-24 Martin Blom
* Bugfixed the ix86be-amithlon version and removed that beta
requester at startup.
* Removed void.audio from default build, since it's just an
example driver.
* Changed TARGET in Makefiles to EXECUTABLE, since TARGET
interferes with the AROS build environment. [Adam Chodorowski]
* ahi.device: Most of the library function gates are now
auto-generated.
* Nuked files that are now auto-generated (FD and clib includes).
* No separate SAS/C and Storm C files anymore. The proto and
pragma files work for all compilers now.
* The build process no longer uses fd2inline and fd2pragma. I
switched to SFD files and sfdc instead.
* emu10kx.audio: Don't build unless OpenPCI is found.
* ahi.device: More embarrassing bugs fixed: Due to a type in the
Makefile, the powerpc assembler optimized mixing routines were
never included!!
2003-07-23 Martin Blom
* ahi.device: The linear interpolation code was like totally
fucked up when sounds were looping or joined. The test sine in the
preferences program finally sounds ok at other frequencies than 48
kHz.
* AHI: "Play a test sound" now plays a 16 bit sample instead of 8
bit, and it's always played at full volume.
2003-07-14 Martin Blom
* Minor updates to the SFD files.
2003-07-13 Martin Blom
* Added SFD files contributed by Steffen Haeuser to the developer
archives.
* Include/C/clib/ahi_(|sub_)protos.h: Added extern "C" if C++
detected. [Steffen Haeuser]
2003-07-07 Martin Blom
* Released emu10kx.audio 5.6.
* The EMU10kx CAMD driver is now part of the m68k build.
2003-07-06 Martin Blom
* Released ahi.device 5.41.
* ahi.device: Reworked the dynamic volume so that the master
volume is scaled based on the number of device openers. Then, for
each opener, each request is individually scaled based on the
number of active requests. This way, different programs will not
affect each other.
2003-07-05 Martin Blom
* emu10kx.audio: The newly added CAMD init code would crash the
driver would not load.
* Released ahi.device 5.40.
* ahi.device: Made a subtile change to the linear interpolation
behaviour when a looped sound is played. Let's hope I made it
better and not worse ...
* Released AHI preferences 5.10.
* AHI preferences: Added annotation to the properties box.
* Released emu10kx.audio 5.5.
* Released void.audio 5.5.
* Released filesave.audio 5.2.
* Drivers/Common/library.c: Fixed init code for MorphOS and
Amithlon.
* Drivers/Common/library.h: Added PROCGW macro to handle
CreateNewProc() entry for all targets.
2003-07-04 Martin Blom
* emu10kx.audio: Added CAMD support.
* emu10kx: A CAMD driver that works with both V37.14 and V40.4 of
camd.library.
2003-06-30 Martin Blom
* Released ahi.device 5.39.
* Resurrected the old m68k assembler version of the echo routines.
2003-06-23 Martin Blom
* emu10kx.audio: Added protection against multiple calls to
AHIsub_AllocAudio().
2003-06-22 Martin Blom
* Released emu10kx.audio 5.3.
* emu10kx.audio: Ported to openpci.library.
2003-06-20 Martin Blom
* Released emu10kx.audio 5.2.
* emu10kx.audio: Ported to use the common driver architecture
shared with void, filesave and now emu10kx.
2003-05-03 Martin Blom
* Minor Makefile fixes for building in the the source tree.
2003-05-01 Martin Blom
* Released ahi.device 5.38.
* ahi.device: Fixed an interpolation bug in the mixing routines
for 32 bit samples that appeared when 32 bit samples were used
with other sample types.
* The PlaySineEverywhere example now plays the sine using all
available sample types (mono/stereo, 8/16/32 bits). It also slides
the pan from the extreme left to the extreme right.
2003-04-30 Martin Blom
* Added catalonian translations by Josep Subirats i Ferré. Thanks!
2003-04-29 Martin Blom
* Released ahi.device 5.37.
* The non-hifi audio modes are now assembler optimized in the m68k
version.
* Updated autodocs to say that V6 drivers do not need to preserve
all registers in AHIsub_Disable()/AHIsub_Enable().
* Added ahiac_PreTimerFunc and ahiac_PostTimerFunc to
AHIAudioCtrlDrv structure (available to V6 drivers only, of
course).
* When building the developer's archives, only relevant header
files are included now. For example, only the m68k archive
contains header files for SAS/C, Storm and assemblers.
2003-04-25 Martin Blom
* Released ahi.device 5.36.
* ahi.device: The dynamic volume code now assumes that each
application will play at least one sound. (It's just a test.)
* The m68k version is now compiled for 060 per default, but will
still run on 020+ of course.
2003-04-24 Martin Blom
* Released filesave.audio version 5.1.
* Released ahi.device 5.35.
* Released void.audio 5.4.
* Added a new example: PlaySineEverywhere.
* CloneAudioModes: Removed warnings when BPTR is a pointer.
* ahi.device: The workaround for the m68k compiler bug introduced
in ahi.device 5.25 assumed a big endian CPU. Now, the workaround
is only applied when compiling for the m68k.
* filesave.audio: Now endian aware. Avoided varargs in order to
make it compile for AROS.
* filesave.audio: Can now be compiled using gcc.
2003-04-23 Martin Blom
* Released void.audio 5.3.
* void.audio: Moved generic code parts to Drivers/Common, which
can now be used by other drivers too.
2003-04-22 Martin Blom
* Released void.audio 5.2.
* Released AHI-Handler 5.2.
* Released ahi.device 5.34.
* Released AddAudioModes 5.4.
* Released AHI preferences 5.9.
* The endian update from 2003-04-17 broke all big endian
targets. Fixed.
* void.audio: This example driver can now be compiled using gcc
and builds for all targets (m68k-amigaos, ppc-morphos,
i?86be-amithlon and i386-aros).
2003-04-18 Martin Blom
* AHI: If no mode is available, the "Play a test sound" button is
disabled.
2003-04-17 Martin Blom
* Ported AHI to AROS.
* The device and preference program are now endian aware when
reading IFF files.
2003-04-13 Martin Blom
* ahi.device: For some weired reason I apparently disabled
PreTimer()/PostTimer() support (aka the CPU overload protection)
in version 5.29. Now it's enabled again.
2003-04-12 Martin Blom
* Released ahi.device 5.33.
* ahi.device: The audio mode requester now resizes to its minimum
width if asked to open a smaller window. Before, it just failed
to open.
* ahi.device: The audio mode requester now uses the ASL
preferences if available (requires asl.library V45 or
later). [Todd A. Oberly]
* ahi.device (AudioRequestA): AHIR_PrivateIDCMP is now honored
even if AHIR_Screen or AHIR_PubScreenName is provided.
* PlayTest: I finally fixed the bug that could make the example
call DeleteIORequest() on a pointer obtained from AllocMem().
* Released ahi.device 5.32.
* ahi.deivce: io_Actual is now valid even if a CMD_READ or
CMD_WRITE request failed or was aborted. [Grzegorz Kraszewski]
* ahi.doc: Don't mention io_Actual for CMD_STOP/CMD_START.
* ahi.device: gw_IndexToFrequency in the MorphOS version now uses
d0/a1 for arguments instead of the stack. [Patch from Harry
Sintonen]
* Released paula.audio 4.29.
* paula.audio: Small cleanup here and there. [Patch from Harry
Sintonen]
* paula.audio: Library cleanup didn't free resources at
failure. Fixed. [Patch from Harry Sintonen]
* paula.audio: DMA playback did CloseDevice with a0 =
p_TimerReq. Fixed. [Patch from Harry Sintonen]
* ahi.device: Added new MOSSYS patches (MorphOS only) from Harry
Sintonen.
2003-01-24 Martin Blom
* ahiusr.texinfo: Updated the Acknowledgments chapter. Added
Donors and Honourable mention sections. Updated my
address. Updated the list of supported sound cards.
2003-01-23 Martin Blom
* Limit lha compression method to lh6.
* Now builds AHI-Handler for ppc-morphos too.
2003-01-22 Martin Blom
* Docs: Now builds AutoDocs for ahi.device and drivers.
* Released AHI-Handler 5.1.
* Drivers: source code for a couple of drivers now included in
bindist (ahidev.lha).
* Examples: source code now included in bindist (ahidev.lha).
2003-01-20 Martin Blom
* AHI-Handler: Now part of build tree, ported to gcc (m68k only!),
GPL:ed and will no longer accept 24 bit AIFF/AIFC files (which AHI
cannot play -- but 32 bit files are still accepted).
2003-01-19 Martin Blom
* Released ahi.device 5.31. This is actually all I had in mind for
V6 (except for a proper distribution with installation scripts
etc, which I don't feel like doing anymore), so unless somebody
complains, I'll bump the version to 6.0 and branch the source
tree. Speak now or forever hold ...
* Updated the User's and Developer's guides slightly (dynamic
volume, AntiClickSamples, AHISF_NODELAY, 32 bit samples).
* ahi.device: Device mode now automatically scales the volume to
the number of playing requests. This means that you can now set
the number of channels in the preferences program as high as you
wish without getting annoyingly low volumes when playing just one
audio stream.
* ahi.device: Added support for playing 32 bit samples in both
low-level and device mode. Note that 32 bit samples are truncated
to 16 bit in non-HiFi modes and there are no assembly mixing
routines for this sample format (yet?). [Aaf]
* Updated copyright year in all files.
2002-12-14 Martin Blom
* Released ahi.device 5.30.
* ahi.device: Now looks into MOSSYS:DEVS/AudioModes and
MOSSYS:DEVS/AHI in MorphOS build. [Harry Sintonen]
* Released AddAudioModes 5.3.
* AddAudioModes: Now looks into MOSSYS:DEVS/AudioModes in
MorphOS build. [Harry Sintonen]
2002-11-20 Martin Blom
* Released toccata.audio 4.6.
* Drivers/Toccata/toccata.c: intAHIsub_GetAttr() could generate
an enforcer hit if no card is present. [Sensei]
2002-10-16 Martin Blom
* ahi.device/paula.audio: The distorsion bug reported by Sensei
(see 2002-07-06) turned out not to be the 50 Hz bug. In fact, it
turned out not to be a bug at all. It was just paula.audio that
couldn't play at 44100 Hz and instead using 44336 Hz, causing AHI
to resample all sounds that should play at 44100 Hz. Haha, I knew
it. Not my fault, not my fault, not my fault!! (See the
documentation for the AHIpaulaFakeMixFreq environmant variable for
more information.)
2002-09-30 Martin Blom
* Released ahi.device 5.29.
2002-07-08 Martin Blom
* ahi.device: Did you know, by the way, that AttemptSemaphore() is
not safe to call from interrupts? Oh, you did? Proves what a sad,
miserable excuse for a developer I am. This bug has been there for
about five years now ... [Nicolas Sallin]
2002-07-07 Martin Blom
* Released AHI preferences 5.7.
* AHI: No enforcer hits if no modes are available. [Don Cox]
2002-07-06 Martin Blom
* Released AHI preferences 5.6.
* AHI: Don't crash if no modes are available. [Don Cox]
* [ahi.device: 5.25 IR]
* ahi.device: The sample pointer would drift 1/(2^32) samples for
each interrupt. No big deal, since it would take about 3 years
until it actually caused a sample frame to be skipped, but a bug
is a bug and deserves to die.
* ahi.device: In addition to this, due to a (m68k) compiler bug,
the sound would be *1.0* sample off for each interrupt. Ouch! This
should fix that "50 Hz click bug" (50 Hz is the default buffer
"size"). [Przemyslaw 'SENSEI' Gruchala, Bernd Roesch]
* ahi.device: Removed checks for unused (since v5) speed-up flags
(surround, echo and clipping).
2002-07-01 Martin Blom
* Released emu10kx.audio 4.11.
* emu10kx.audio: Set all active AC97 inputs to 0 dB at startup.
2002-05-28 Martin Blom
* Released emu10kx.audio 4.10.
2002-05-25 Martin Blom
* emu10kx.audio: Fixed the low PCM volume by using the full
digital dynamic range. Set the default analog PCM mixer to 0 dB
(was +6 dB).
2002-05-13 Martin Blom
* Released emu10kx.audio 4.9.
* emu10kx.audio: Reordered the recording sources so the "Mixer"
comes first (ProStation Audio workaround).
* emu10kx.audio: The microphone was unmuted by accident when the
playback was stopped.
2002-05-12 Martin Blom
* Released emu10kx.audio 4.8.
* emu10kx.audio: Added code to handle recording and the selection
of input sources.
2002-05-11 Martin Blom
* AHI-Handler: Imported into CVS.
* Examples: Added all developer examples to the CVS repository.
2002-03-17 Martin Blom
* Released AHI preferences 5.5.
* AHI: Disabled unused advanced options (anti-click time when
using AHI v4, echo, surround and master volume when using AHI v5).
2002-02-27 Martin Blom
* Released emu10kx.audio 4.6.
* emu10kx.audio: Playback works. AC97 analog mixer initialized to
reasonable defaults. I made it in time for AmiGBG 2002! :-)
2002-02-24 Martin Blom
* Released AHI preferences 5.4.
* AHI: Increased driver info buffers from 32 to 64 characters.
2002-02-15 Martin Blom
* emu10kx.audio: Initial check-in. A header file is a good start,
no?
2002-02-09 Martin Blom
* toccata.audio: The CPU load check now also includes the
PlayerFunc.
2002-01-20 Martin Blom
* The '-s' flag is not used when stripping the binaries anymore.
This way, the Amithlon binaries can be stripped too.
* ahi.device: Say hello to the "Fast" modes again. This time,
instead of being removed, they are added to the database but never
returned by AHI_NextAudioID(), meaning that they will never show
up in mode requesters etc. They are still available in
AHI_AllocAudio(), though, and behave just like the "non-HiFi"
modes: one multiplication/sample and no assembler optimization.
2002-01-06 Martin Blom
* Released ahi.device 5.21.
* ahi.device: Re-enabled optimization and removed all 64 bit
kludges for the Amithlon version. It's still just using the C
version of the mixing routines, but is already about 2.5 times
faster native (compared to the JIT-compiled 4.180 version). I like
it.
* paula.audio: When opening timer.device (used in the DMA mode),
I didn't clear the flags parameter and opened UNIT_MICROHZ instead
of UNIT_ECLOCK. [Sigbjørn Skjæret]
2002-01-03 Martin Blom
* toccata.library: Added LGPL license.
* toccata.library: Applied the latest fixes I got from Teemu.
* toccata.library: Imported the old emulation library to CVS.
* toccata.audio: The PreTimer() and PostTimer() calls did not pass
the required argument.
* ahi.device: "elfloader.h" was included even when it's not used.
2002-01-02 Martin Blom
* Updated the Hungarian catalogs. [Márton Dósa]
2001-12-29 Martin Blom
* When cross-building, the wrong 'strip' was used on the binaries.
2001-12-22 Martin Blom
* Added a note about AHIA_PlayerFunc in the autodocs.
2001-12-21 Martin Blom
* Released AHI preferences 5.3.
* Released ahi.device 5.20.
* Added alpha support for Amithlon/native x86. Go ahead, try it,
but do save your work first ... Then enjoy the fireworks!
* Fixed some minor stuff in the AHI preferences so I can build
native MorphOS/Amithlon versions of it. Why? Because I can.
* The change in the mode file handling on 2001-07-29 seems to have
made AHI_LoadModeFile() return failure if the argument was a
directory. Now it always returns success in that case.
2001-12-19 Martin Blom
* If the device's init routine failed, the device memory was not
freed and the libraries that had been opened were not closed.
* Tweaked some things so I can cross-compile better.
* Rewrote the last m68k assembler-only file in C. Whoo... The
MorphOS version is finally feature-complete.
2001-12-14 Martin Blom
* Released AHI preferences 5.2.
* AHI: It's now possible to test the selected audio mode by
pressing the "Play a test sound" button.
2001-12-14 Martin Blom
* Released ahi.device 5.19.
* When adding an audio mode file, ahi.device now verifies that the
driver can be opened before adding the modes to the database. If a
driver refuses to load if its hardware is not present, the useless
modes will thus not be added. Driver authors, feel free to make sure
your drivers behave like this.
2001-12-13 Martin Blom
* Released filesave.audio version 2.11.
* filesave.audio: Added RIFF WAVE output.
2001-11-26 Martin Blom
* ahi.device: Added a couple of validation checks for mode files.
* Released AddAudioModes 5.2.
* AddAudioModes: The REFRESH flag now removes the audio mode
database before (re-) scanning DEVS:AudioModes for mode files.
2001-09-29 Martin Blom
* Released ahi.device 5.17.
* I tweaked the build environment somewhat: makedepend is out; gcc
-MM is in, even in AmigaOS. "make all" now aborts on errors, as it
should. All automatically generated files (version.h, localize.*)
are now build before the dependencies are created.
* ahi.device: Haha! I found the race condition in question! It
was in AddWriter(), which assumed that no free channel
automatically implied that PlayingList was non-empty. Not so!
[Stephan Rupprecht and Harald Frank]
* ahi.device: Fixed yet another potential race condition (in
SoundFunc()). Sooner or later I *will* find the real one, won't I?
* ahi.device: Fixed a bug in TermIO (ioreq->ahir_Extras is now set
to NULL before its FreeVec'ed, not after).
2001-09-28 Martin Blom
* Released paula.audio version 4.27.
* paula.audio: I broke the DMA playback mode in the last update.
Fixed now. [Sigbjørn Skjæret]
2001-09-24 Martin Blom
* ahi.device: A missing 'break' statement made AbortIO() return
IOERR_NOCMD on CMD_WRITE requests.
* ahi.device: Fixed a non-fatal race condition in DevAbortIO().
It just made the source code look prettier ... :) Now if I could
only find the one that makes AHI crash in AMIthlon ...
2001-09-16 Martin Blom
* Released paula.audio version 4.26.
* Released ahi.device 5.8.
2001-09-09 Martin Blom
* paula.audio: The fixed CMD_STOP/CMD_START in ahi.device
triggered a bug that could make the driver lose track of the
AHIsub_Disable and AHIsub_Enable nesting. Also, these functions
was not "thread safe" (they are not actually called from a thread
but rather an interrupt).
2001-09-04 Martin Blom
* ahi.device: Hopefully, CMD_STOP/CMD_START actually work now.
* ahi.device: Fixed a dead-lock problem that could happen when
more than one double-buffered sound was played using the device
API.
2001-07-29 Martin Blom
* ahi.device: If a file in Devs:AudioModes is not a legal mode
file, the scanning code would just stop. Now, such errors are
silently ignored.
2001-04-22 Martin Blom
* ahi.device: When using a filter hook with the audio mode
requester, and AHIR_DoDefaultMode was TRUE, the hook would receive
AHI_INVALID_ID instead of the user-selected default mode ID.
2001-04-21 Martin Blom
* Released ahi.device 5.7.
* ahi.device The default PlayerFunc (which is used unless the
application provides one) did not have an m68k wrapper when the
device was compiled for ppc, making ReqAttack (and many other
programs, I guess) crash. [Jaca/D-CAPS]
2001-04-12 Martin Blom
* PAULA: Always use calibration tables for 14-bit output.
2001-03-29 Martin Blom
* Released ahi.device 5.6.
* ahi.device: Fixed a bug in the mixing routine that, when using
"anti-click", caused notes to be lost in rare cases. Credits to
Jaca/D-CAPS for spotting it and sending me an example that
reproduced the bug!
2001-03-11 Martin Blom
* Released ahi.device 5.5.
* ahi.device: Nuked one more m68k assembler source file and moved
the code to C.
* ahi.device: Stopped using dprintf in MorphOS.
2001-03-04 Martin Blom
* Released ahi.device 5.4.
* ahi.device Fixed a bug that caused clicks on every new sound
start, if an end-of-sound had been reached and the last sample in
that sound was not 0.
2000-11-28 Martin Blom
* Released AHI 5.1.
* Last-minute bugfixing for AmigsOS 3.9 CD.
Download Latest Version
ahisrc-6.0.tar.gz (685.4 kB)
