2005-12-18  Russell Marks  <russell.marks@ntlworld.com>

	* Version 0.5.

	* sound.c (sound_start_fade): remembered why fadetime was declared
	as static - it was a different variable. As such, version 0.4 got
	released with a zero fade-out time. :-( Fixed again by renaming
	sound.c's fadetime to sfadetime, which should hopefully avoid
	future confusion...

2005-07-14  Russell Marks  <russell.marks@ntlworld.com>

	* Version 0.4.

	* sound.c: `fadetime' was declared as static when it shouldn't
	have been. Thanks to Daniel Baumann for pointing this out.

2004-05-15  Russell Marks  <russell.marks@ntlworld.com>

	* main.c (action_callback): fixed unhelpful interactive
	stop-after-setting behaviour when started with stop-after set to
	something not a multiple of 30 seconds; now the first interactive
	change will set it to the nearest multiple in the specified
	direction. Thanks to Bartlomiej Ochman for this.

2004-05-01  Russell Marks  <russell.marks@ntlworld.com>

	* main.c: added `-t' option, to play only a given track (actually
	slightly different, see the man page). Thanks to Bartlomiej Ochman
	for this.
	
2004-03-26  Russell Marks  <russell.marks@ntlworld.com>

	* Added 16-bit support. Now defaults to this when possible. New
	option `-e' forces 8-bit playback (even that is improved, due to
	the 16-bit mixing now done). Thanks to Stuart Brady for inspiring
	this change.

2003-11-30  Russell Marks  <russell.marks@ntlworld.com>

	* aylet.1: fixed typo.

2003-11-27  Russell Marks  <russell.marks@ntlworld.com>

	* ui.c (ui_frame): fixed minor bug when checking value of KEY_DC
	at compile time. Unlikely to affect anything in practice.

2002-09-14  Russell Marks  <russell.marks@ntlworld.com>

	* Version 0.3.

	* main.c (out): added partial port-bitmask to allow for certain
	less-than-ideal .ay conversions. (This is only done when the low
	byte is FDh, and even then doesn't disable CPC autodetect, so it
	shouldn't break CPC tunes.) Thanks to Vt Hotrek for reporting
	the files that (previously) didn't work.

2002-06-24  Russell Marks  <russell.marks@ntlworld.com>

	* z80.c (z80loop): fixed silly typo which meant that L and L'
	weren't set correctly when starting up the Z80. (Though curiously,
	this bug didn't seem to break any .ay files.) Thanks to Patrik Rak
	for spotting this.

2002-05-25  Russell Marks  <russell.marks@ntlworld.com>

	* sound.c (sound_frame): previously, when a track stopped and
	happened to do so leaving high/low level `silence' (e.g. a few AY
	tracks, and all beeper tracks after the recent beeper-fade
	removal), if this change happened during a fade the fading level
	would screw up the silence detection and give (with default
	settings) up to ten seconds of extra `silence'. Fixed that by
	detecting silence before fade filter is applied, and overriding
	the not-silent result if the track has been completely faded out.

2002-05-05  Russell Marks  <russell.marks@ntlworld.com>

	* sound.c (sound_frame): removed beeper fading, which wasn't
	actually necessary and was causing problems with some tracks, most
	noticeably in Trantor. The rest position is still central for
	AY-only and CPC tracks, though, so the change shouldn't affect
	those.

	* uigtk.c (DO_TBL_RIGHT): long file details (e.g. track name) no
	longer expand the window to fit, but are clipped. You can still
	manually resize the window to see the rest of the text, if you
	like.

2002-03-16  Russell Marks  <russell.marks@ntlworld.com>

	* sound.c (sound_ay_overlay): now emulates tone counters at all
	times, rather than just doing it when producing a tone and
	attempting to compensate later. I'm not sure if this makes any
	practical difference, but it's a more accurate emulation.

2002-02-26  Russell Marks  <russell.marks@ntlworld.com>

	* aylet.1: added explanation of what ACB/ABC actually mean, rather
	than just pulling the terminology out of thin air. :-)

2001-12-21  Russell Marks  <russell.marks@ntlworld.com>

	* sound.c (sound_ay_init): finally added decent AY levels. Thanks
	to Matthew Westcott for the measurements these were based on. The
	change from the previous values isn't all that big to be honest,
	but we should at least get a faithful `blip' now. :-)

2001-12-11  Russell Marks  <russell.marks@ntlworld.com>

	* Version 0.2.

	* sound.c (sound_ay_init): the calculated AY volume levels used
	previously were wrong, even though they were based on the levels
	claimed by official documentation. (!) The new precalculated
	levels should be a lot more accurate, but don't yet reproduce a
	curious `blip' in the AY DACs. (I may use values as measured from
	a speccy in future, in order to do this.)

2001-11-13  Russell Marks  <russell.marks@ntlworld.com>

	* drv-oss.c: the OSS driver has now moved here.

	* drv-obsd.c: OpenBSD sound driver. Thanks to Chris Cox for this,
	and separating out the OSS code.

2001-11-11  Russell Marks  <russell.marks@ntlworld.com>

	* main.c (out): whoops, no *additional* t-states for OUT... this
	explains why some beeper tunes with very tight loops over OUT
	instructions (e.g. Heavy on the Magick) were slow before.
	(out): ok, we need to disable CPC detection when *any* speccy port
	is written to.

	* sound.c (sound_ay_overlay): fixed a strange clicking problem;
	this was probably most noticeable in Defenders of the Earth,
	especially the drum solo bit, but there were several other cases.
	I've been struggling with this one for maybe a week now (bleah),
	but I think I've finally got it. :-)
	(sound_ay_overlay): ok, that fix wasn't quite right; changed
	tone_tick workings so it should work properly now.

2001-11-09  Russell Marks  <russell.marks@ntlworld.com>

	* sound.c (sound_ay_overlay): fix for high-frequency noise.

2001-11-04  Russell Marks  <russell.marks@ntlworld.com>

	* sound.c (sound_ay_overlay): rewrote envelope emulation, the
	old one couldn't be made to cope with high-speed envelopes (as
	used in some demos, e.g. Binary Love track 2). Also fixes
	accidental (?) zero-period envelope use with FFh volume (e.g.
	Afterburner).

2001-11-03  Russell Marks  <russell.marks@ntlworld.com>

	* Changed field label from "Title" to "Misc" throughout. Some
	files use it for Title, some Copyright, some both. So "Misc" is
	about the only reasonable label.

	* main.c (out): added support for CPC tunes. Unfortunately these
	have to be auto-detected (!) so it starts off emulating the usual
	cut-down speccy, then switches to a cut-down CPC if those ports
	are used, slowing down the AY and speeding up the Z80. Also, if
	the (speccy) beeper is used (which could theoretically clash with
	the CPC stuff) and it's not in the CPC mode, the CPC ports are
	disabled for the current track.

2001-10-25  Russell Marks  <russell.marks@ntlworld.com>

	* sound.c: beeper tones were *also* upside-down, not just AY -
	fixed.

	* aylet.1: again, it's 44.1kHz, not 44 (see below).

	* main.c (usage_help): oops... `-s' outputs a stereo sample by
	default, not mono. Also, it's 44.1kHz not 44.

2001-10-21  Russell Marks  <russell.marks@ntlworld.com>

	* Version 0.1.

	* sound.c (sound_frame): correctly detect silences at a non-zero
	level - some .ay files have tracks which end like this, most
	likely due to the original author's misunderstanding/ignoring an
	AY chip oddity (the whole `turning the channel tone and noise off
	in the mixer doesn't disable it, you have to zero the volume'
	thing). Understandable, really, it's not an *audible* problem...

2001-10-17  Russell Marks  <russell.marks@ntlworld.com>

	* sound.c: tweaked beeper/AY volume balance. I decided not to make
	the beeper the same volume as all 3 AY channels together, it was
	just annoyingly loud like that. But I did try to give beeper and
	AY tunes more comparable levels, so the beeper is a bit louder
	now. Agent X 2 still sounds odd, but I think I may just have to
	add an option for the proper balance at some point, because it'd
	be a crazy default for a player given how loud the beeper really
	is.
	(sound_ay_overlay): fixed behaviour when neither tone nor noise is
	selected on a channel, meaning that e.g. Chase HQ's samples work.
	(sound_ay_overlay): AY tones were upside-down :-), fixed.

2001-10-16  Russell Marks  <russell.marks@ntlworld.com>

	* sound.c: added `-N' option for narrower AY channel stereo
	positioning.
	(sound_beeper): fix for slight spikes on ordinary beeper tones.

2001-10-15  Russell Marks  <russell.marks@ntlworld.com>

	* sound.c: improved ACB stereo, and made it the default. (Use `-m'
	to force mono playback.) Also added a pseudo-stereo option for the
	beeper sound; not enabled by default (use `-S') as it can make
	things sound a bit shaky and rough in some cases.

	* main.c: default stopafter reduced from 5 minutes to 3. I think
	this is probably a decent default, even if it is too short for
	some things (e.g. Agent X 2).

	* sound.c (sound_ay_overlay): removed tone/noise counter
	wraparound (inherited from z81), which wasn't necessary and
	screwed up (e.g.) Marauder and Cybernoid 2.

2001-10-14  Russell Marks  <russell.marks@ntlworld.com>

	* main.c: added `-s', a play-to-stdout option.
	(do_interrupt): fixed bug in non-UI mode, where it could have been
	left on a non-existent file.
	(do_list): added `-l', to list the contents of files rather than
	playing them.

	* ui.c (ui_frame): fixed backspace/del, which I'd managed to
	break.

2001-10-13  Russell Marks  <russell.marks@ntlworld.com>

	* main.c (do_interrupt): now returns to first file/track (rather
	than sticking on the last one) as well as stopping when it's got
	no more to play.

	* ui.c: added non-UI `UI'. :-) It's a bit like using mpg123. If
	you press ^C (well, send SIGINT) in the first half-second of a
	track the program exits, otherwise it starts the next track.

	* main.c (action_callback): prev/next track now cross files.

	* uigtk.c: a GTK+ UI. Probably matches the curses one a bit too
	closely, making things like tweaking stopafter/fadetime rather
	clunky, but I think it'll do for now.

2001-10-10  Russell Marks  <russell.marks@ntlworld.com>

	* sound.c (sound_ay_overlay): whoops, get that frequency right -
	not on a ZX81 any more... :-)

	* main.c (mainloop): now supports prev/next file. Pretty hairy,
	but it's working. The playlist comes from the command-line at the
	moment; that behaviour's probably going to stick, but I might add
	an alternative way of specifying a playlist eventually.

2001-10-09  Russell Marks  <russell.marks@ntlworld.com>

	* sound.c (sound_ay_reset): fixed, previously didn't always reset
	due to possible non-zero tstates when calling.

	* First essentially working version, based on nc100em and z81.
	(z81's sound emulation being based on the one I wrote for Fuse.)