GUS Daily Digest            Tue, 24 May 94  9:37 PST     Volume 13: Issue  21 

Today's Topics:
						32-bit Sierra drivers
						  Dune II revisited
					   GUS Daily Digest V13 #20
				   If Raptor+GUS hangs, read this!
					  MegaRace & Pacific Strike
						  PAT to raw samples
							Pirates! Gold
					   Privateer music is yuck
						 Protected mode games
						  Q4G-IV and X-Wing
				  Re Cheap memory chips for upgrade
		  Sam 'n Max CDROM - anyone succeeded with the GUS?

Standard Info:
	- Meta-info about the GUS can be found at the end of the Digest.
	- Before you ask a question, please READ THE FAQ.

----------------------------------------------------------------------

Date: Mon, 23 May 1994 14:22:05 -0700
From: Scott.Jordahl@Eng.Sun.COM (Scott A. Jordahl)
Subject: Re: 32-bit Sierra drivers

There is a LARGE patch available from the Sierra BBS for GK, both a CD
and floppy version (Sorry, can't remember the file names). Have you
installed it?  I too installed the 32bit driver with my CD version GK.
The MIDI sounds GREAT, but the problem I've been having is that the DAC
voices (characters talking) cut in-and-out. Does anyone have a
quick-fix solution for this problem?

-- Scott

> 
> Date: Sun, 22 May 1994 17:01:10 EDT
> From: lpkruger@phoenix.Princeton.EDU (Louis P. Kruger)
> Subject: 32-bit Sierra drivers
> 
> First of all, for those who haven't heard yet, they're OUT!!!
> Look on epas for the file gus32.exe
> 
> Second, I'm having a bit of trouble with the DAC driver in Gabriel
> Knight. (the midi driver works fine)  When I try to leave the map scene,
> I get a "Internal Game Error #135, except.cpp:36:64989:ed"
> Not fun.  I tried moving my IRQ's around, but it didn't seem to help.
> Any suggestions, anyone?  It works fine if I just use the midi driver,
> however.
> 
>   - Louis
> 


|=|=|=|=|=|    Scott A. Jordahl - SunSoft      |=|=|=|=|=|
|==|=|=|==|    Sun Microsystems - Mt. View, CA |==|=|=|==|
|=|=|=|=|=|    VOICE: [415] 366-5659           |=|=|=|=|=|
|==|=|=|==|    INTERNET: scottj@Eng.Sun.COM    |==|=|=|==|

If you give a man a fish, he'll fish for a day.
If you teach a man to fish, he'll fish for a life time....
 .. And he'll live for a life time.            -- Dan Quayle 10/13/92

------------------------------

Date: Tue, 24 May 1994 14:55:20 +0200
From: Afzal Ballim <afzal@divsun.unige.ch>
Subject: Dune II revisited
Message-ID: <3052*/S=afzal/OU=divsun/O=unige/PRMD=switch/ADMD=arcom/C=ch/@MHS>

Over the weekend I got Dune II and copied the patches,
etc to get it to work with the GUS. Well, I instantly
got great effects and digitised voices, but no music.
Great, I thought, not enough memory. So I switched off
smartdrive and tried again.

Nope, still no music. Well I played around with all
sorts of things, and eventually out of desperation
I did:

	ultramid -f
	ultramid -16 -m100

and tried again.

Yes! Effects, speech, AND music. So, the question is
why does "ultramid -c -m100" give me no music, while
"ultramid -16 -m100" does?

BTW, I have a PC with 4M of memory, and can have smartdive
running with a 256k cache without loosing music, although
I can "pauses" in the game.

-Afzal

------------------------------

Date: Tue, 24 May 94 09:16:51 +0300
From: Yossi Oren <LIOREN1@WEIZMANN.WEIZMANN.AC.IL>
Subject: Re: GUS Daily Digest V13 #20

>>The AWE has 11MB of patches compressed using some lossy format into 1MB.
>
>I think this may just be an urban legend.  Some have posted that the
>Waveblaster has 11MB of samples compressed to *4MB*.  Others who have it
>say this is nonsense.  The AWE32 uses 1MB, maybe the same 4MB of the WB
>compressed.  I'm not sure of that.  But 11MB compressed to 1MB??  Sounds
>improbable, doesn't it?
You can get up to 36:1 lossy compression with existing standards.  11:1 is
breakfast for them.  Only by this time, you SEE it's lossy.

>Again, I question this.  I haven't seen anyone else post this.  Are you
>getting it confused with the Turtle Beach Rio and Monterey?
Sorry, seen it on csips - somebody bought an AWE32 real cheap and is now
ranting about it.  Anyway, there SHOULD be an API for it somewhere - but CL
will only SELL it to you.  MIDI dumps are what you're SUPPOSED to do, it's
just that they're DAMN SLOW, so forch-nut-lee we have an alternative.
Yossi.

+---+-----------------------------------------------------------+---+---+
| = | Yossi Oren, Al-Daf Technological Mercenaries. [diet .sig] | v | ^ |
+---+-----------------------------------------------++----------+---+--+|
|LIOREN1@weizmann.weizmann.ac.il (or just @weizmann)||GUS-sound that's ||
|The people are with the Golan Heights!             ||worth hearing!   ||
+===================================================++=================++

------------------------------

Date: Mon, 23 May 1994 18:27:27 -0500
From: Terry Lee <terry@uivlsi.csl.uiuc.edu>
Subject: If Raptor+GUS hangs, read this!

	There is a problem of Raptor+GUS hanging on some systems when the
number of voices is greater than 4.  My system used to be one of those, and I
did resent Apogee for not getting the GUS driver "right".  But now I think my
resentment was misplaced, and that Apogee actually did me a favor by writing
their suboptimal GUS driver.  In my case, the error was actually caused by
very subtle chipset configuration errors.  It is possible that this is true for
all systems where Raptor+GUS hangs.

	I run a lot of DOS games with my GUS, and it seemed that only Raptor
was giving me problems.  I can also run QAplus all day without getting errors.
So I thought Raptor was to be blamed.  It wasn't until I started running
FreeBSD (a free 32-bit BSD OS) heavily that I noticed very occasional parity
errors.  I set out to discover the cause of the parity errors, and when I
finally fixed it, my Raptor problem was gone, too.

	The reason I described the Raptor GUS driver as suboptimal is because
every sound effect is DMAed to the card each time the sound effect is played.
With 8 voices, the system is subjected to constant heavy DMA traffic.  As it
turns out, this effect does a good job of exposing chipset configuration
errors.

	WARNING:  Try the following procedures at your own risk.  Variations
among motherboards also exist.  Standard disclaimers apply.

	To try the steps I will describe next, you will need a copy of AMIsetup
(amise260.zip from many ftp sites).  It should work on most systems.  Even if
your BIOS setup appears to be pretty comprehensive, I would still try AMIsetup
to see if any option is hidden.  Check to see if there is an
"Auto configuration" option; you must disable it for the wait state and bus
clock settings to take effect.  One motherboard I've owned actually had the
"Auto configuration" option hidden (and enabled), so that no matter how I used
the BIOS setup to change memory wait states and bus speed, nothing happened.

	My current motherboard is the AIR UMC VLB motherboard (highly
recommended!).  It has 15ns SRAM cache with 12ns tag SRAM.  The motherboard
doesn't have the 16-bit DMA bug.  (Even if your motherboard has the 16-bit DMA
bug, you can still get around it by using 8-bit DMA channels 1 or 3.  If your
system still hangs with Raptor, read on.)  My SIMMs have 70ns access time.
The CPU is an AMD 486DX/40 overclocked to 50MHz.

	On most motherboards (including mine), the default setting enables
"Auto configuration".  One would think that the default setting would take
the CPU clock speed into account, but this is not necessarily true.  The
default setting is often too conservative, and sometimes wrong (especially at
50MHz).  On my system, the default setting has 3-2-2-2 cache read burst, 2ws
cache write, 2ws DRAM read, 2ws DRAM write, and DMA CAS delay enabled.

50MHz operation (486DX50 or DX4 50/100)
 --------------------------------------

	The default setting is not correct at 50MHz unless the SIMMs have 60ns
access time.  With 2ws DRAM read, each read access take 3 cycles.  At 50MHz,
each cycle is 20ns.  Thus each read access takes 60ns, which is too short for
70ns SIMMs.  The correct setting should have been 3ws.  The 2ws DRAM write, on
the other, is probably too conservative.  Each DRAM write access would be
4 cycles (minimum possible is 2 cycles), or 80ns.  However, DRAMs can be
written faster than read, so 1ws probably would have been OK.  If you don't
have VLB boards (the delay going through the VL bus can be significant), you
might even get away with 0ws (40ns) DRAM writes.

	With 15ns SRAM and 12ns tag SRAM, 3-1-1-1 cache read burst and 0ws
cache write is possible.  I think 10ns tag SRAM is required to reliably achieve
2-1-1-1 cache read burst.  Most motherboards, however, have 20ns SRAM cache and
15ns tag SRAM.  With 20ns SRAM cache, probably only 3-2-2-2 cache read and
1ws cache write is safe.  Some motherboards don't even have tag SRAM (slower
cache performance), but they should still work.

	On many motherboards, the DMA clock is always half the ISA bus clock.
Therefore, don't overclock the bus too much.  The DMA controller is probably OK
up to 5MHz (10MHz ISA bus), but not too much beyond.

	Even after all these changes, my system is still not 100% reliable
under extremely extensive FreeBSD testing.  However, it probably can run Raptor
reliably now.  My conclusion is that this AMD 486DX/40 cpu cannot be 100%
reliably overclocked to 50MHz even with a Peltier junction cooler.

40MHz operation (486DX/40 or 486DX2/80)
 --------------------------------------

	The default setting does work, but it is too conservative.  The
settings I am using now is 2-1-1-1 cache read burst, 0ws cache write, 2ws DRAM
read, 0ws DRAM write, and DMA CAS delay enabled.  One danger is that 1ws (50ns)
DRAM read will appear to work, but it is not reliable.  If you have any VLB
cards, you may have to use 1ws DRAM write.  All my problems are gone now.

33MHz operation (486DX/33, 486DX2/66, or DX4 33/100)
 ---------------------------------------------------

	I have not tested this extensively, but the same settings as 40MHz
operation should be used.  If you have 60ns DRAM, you might be able to use
1ws DRAM read reliably.

25MHz operation (486DX/25, 486DX2/50, or DX4 25/75)
 --------------------------------------------------

	I have not tested this, but 2-1-1-1 cache read burst, 0ws cache write,
1ws DRAM read, and 0ws DRAM write should work.

Conclusion
----------

	Even after you get tired of Raptor, it (with a GUS) still serves as
an extremely good system stress test.  If I were to buy a new computer or
motherboard, I definitely would pound it with Raptor+GUS at 8 voices as a test.
Also, be careful if you overclock your CPU.  From a performance perspective,
I believe 40MHz operation best matches standard hardware.  It's too bad that
the DX4 40/100 isn't working.  To really take advantage of 50MHz operation,
you need pretty fast parts (15ns SRAM, 10ns tag SRAM, and 60ns SIMMs).  I hope
the above information is of some help.

					Terry Lee
					terry@uivlsi.csl.uiuc.edu

------------------------------

Date: Mon, 23 May 94 16:36:02 EET DST
From: s106275@ee.tut.fi (Saari Anssi)
Subject: Re: MegaRace & Pacific Strike

>Regarding Pacific Strike, its surprising to see that Origin is using =
>EMS again, which means the game works with MEGA-EM...the music's real=
>ly good with GM emulation but the digital sounds and Speech crashes q=
>uite often or just disappears during the game. Anyone else has this p=
>roblem? If someone manages to make it work properly..please post how.=
> I hava a really old Sound Blaster (v1.5) in the same computer (could=
> be the problem?)=20

Mine worked fine with the SB doing the digital fx and GUS doing the music.

Anssi

------------------------------

Date: Tue, 24 May 94 11:04:30 EDT
From: Akintunde Omitowoju <ZAO1@ETSU.EAST-TENN-ST.EDU>
Subject: PAT to raw samples

Hello GUSsers,
I was wonder whether there are any programs that convert the .PAT sample
files that come with the GUS to raw sample files.  Any info would greatly
be appreciated.  Thanks!!! =)

Akintunde Omitowoju
zao1@etsu.bitnet           (BitNet)
zao1@etsu.east-tenn-st.edu (InterNet)
omit@wiw.org               (InterNet)

------------------------------

Date: Tue, 24 May 94 15:47:59 METDST
From: <martinus@morra.et.tudelft.nl> (Martijn de Jong)
Subject: Pirates! Gold

Hi Gussers,

I have a problem with the music from Pirates! Gold.
When I start the game with megaem and emuset emulating the
Roland Sound Canvas There are some patches who can't seem to
kep the right tone. They start allright but then the tone
gets higher making the music really terrible. Also when you
find a big treasure (song B in Pgjuke) the strings need a
lot of time to give some sound. More than they get cause
they stop before they're on the right volume. 
I believe the first time I played the game everything sounded
allright. After then I've cleaned my HD and put the game on
again and I've upgraded to 1024K gus memory.
Anyone else had this problem? Comments on this? I really
would like the music to be well again, cause I do like this
game.

Martinus.
--
************************************************************
*                   Hi, this was me!!!                     *
*                   Martinus himself!!                     *
*         e-mail: martinus@morra.et.tudelft.nl             *
************************************************************

------------------------------

Date: Mon, 23 May 94 16:38:07 EET DST
From: s106275@ee.tut.fi (Saari Anssi)
Subject: Re: Privateer music is yuck

>
>Hi,
>
>When I play Privateer with the GUS, the music is really REALLY
>bad! Speech and effects work fine. The GUS has 256k mem.
>SBOS v3.7. Would more ram fix the problem?

SBOS only uses 256K. For better music, you need a real SB or
GM device.

Anssi

------------------------------

Date: Mon, 23 May 1994 16:44:47 -0700 (PDT)
From: David.Roberts@amd.com (Dave Roberts)
Subject: Protected mode games

> Date: Sun, 22 May 1994 10:07:53 -0700 (PDT)
> From: Peter C. Chien Jr. <pchienjr@earthquake.berkeley.edu>
> To: rock@POOL.Informatik.RWTH-Aachen.DE
> Subject: Protected mode games
> 
> I assume 486-40 is the Cyrix DLC chip.  In general, I have arrived at the 
> conclusion that non-Intel CPUs may have problems running protected mode 
> games, since I have problems with my IBM 486-SLC2/50.  My own experience 
> is that I couldn't get the CD game Microcosm to work properly.  It would 
> crash, and it uses protected mode.  Then Microcosm came with a patch for 
> that, it runs longer, but still crashes.  I know it's not the UltraSound, 
> because I would play the game without any sound, and it still crashes.  
> Rebel Assault is another problematic game.  In general, if a game uses a 
> DOS4GW.exe extender, use v1.8 of it and copy over the problematic 
> version.  Game may then work.  (Older DOS extenders are found with the 
> Rebel Assault patch, or older games like Syndicate.)  Also, don't bother 
> loading HIMEM.SYS and EMM386, because these may conflict, plus DOS4gW 
> doesn't need them.

Ummm... as an employee of a company that makes a great number of
non-Intel CPUs (;-), I'd like to ask that people not generalize the
way you do in your second statement.  I've run protected mode games on
AMD CPUs (both 386 and 486) and have never had any problems with the
CPU and any protected mode games.  I *have* had problems with
motherboards, however.  I can't say whether the Cyrix or IBM designs
have any problems, but I wouldn't be on it.  If they had, you'd have
heard a lot more screaming when those designs first shipped.  Frankly,
games are often one of the best test vehicles out there.  They beat on
the hardware in very interesting ways.

I'd bet that your problems may be related to some motherboard quirks
that the particular games bring out.

Anyway, in the future, please cite specific examples and don't
generalize beyond your experience.  I can't speak for Cyrix or IBM but
AMD CPUs have always served me well.

Dave Roberts
Advanced Micro Devices, Inc.
I/O and Network Products Division
david.roberts@amd.com

------------------------------

Date: Mon, 23 May 1994 20:42:35 -0400 (EDT)
From: gt5543b@prism.gatech.edu (Antonio C. Rodriguez)
Subject: Q4G-IV and X-Wing

Anyone gotten the 32 bit drivers to work for Quest for GLory 4? Oh and I
was also wondering if the Import Patch is available anywhere on the
Internet. I know it's on CompuServe but I don't have access to that.
Also, I haven't been reading the mail for a while so if anyone can
remember how to fix the problem about the digital sound in x-wing
(targeting computer makes screen freeze for .5 second misses a few frames
of action) I'd like to know. Thanks.

Tony Rodriguez

------------------------------

Date: Tue, 24 May 94 10:03:40 CDT
From: Daniel A Nielsen <dnielsen@ria-emh2.army.mil>
Subject: Re Cheap memory chips for upgrade
Message-ID: <9405241456.aa00989@VGR.ARL.MIL>

I got mine from a neighbor.  There are alot of 1Meg 286 systems out there
using 256K chips for their first meg of memory that has to be replaced
with SIMMs to go to 2M or more of RAM.  You just have to ask around.

regards
dan

------------------------------

Date: Tue, 24 May 94 8:11:08 BST
From: cgd@ecmwf.co.uk (Dick Dixon)
Subject: Sam 'n Max CDROM - anyone succeeded with the GUS?

Although the G-list suggests that Megaem should be fine with Sam'n Max,
the CDROM version is something else.  Nothing so far seems to work - 
anyone with experience to report, please do so.

The game uses the Rational Systems DOS extender DOS4GW, which seems to
have a terminal dislike of Megaem.  Even the sound card setup routine
won't run if Megaem is loaded - just hangs up solid. GDT options on 
megaem seem to make no difference.

There are things that can be done with DOS4GW via the environment string
'DOS16M=`; e.g. DOS16M=@2M forces DOS4GW to load above 2 meg.  However,
still nothing works (ideas, anyone?).

SBOS (both 1.2 and 3.8) give ADLIB music OK but digital sound doesn't work 
properly - broken up and highly intermittent.  All single options and all
combinations of pairs fail to fix this - o4 comes closest, making the sound
on the initial logo sequence almost intelligible; but then the game fails
to start.

The sound drivers look somewhat like AIL drivers, but the DOS extender
falls over if I try to load a miles driver.  (Not obvious how to make the
game run either, then, since this is a CDROM). 

All this is particularly annoying, since the sound is perfectly intelligible
to PLAYFILE, being a collection of normal WAVs (180 Megs of it).  It
can't be far from working if one could patch their bloody driver to
handle SB clone output in a somewhat standarc way.

Has anyone succeeded with this annoying game, or do I have to buy a real SB?
(The Rebel Assault demo on the same CDROM has a really neat setup that
recognises and sets up the GUS immediately.  Oh, and Lucas Arts BBS seems to
have no files which might be relevant.)

All suggestions gratefully received.

Cheers,

====
Dick Dixon

------------------------------

End of GUS Daily Digest V13 #21
*******************************
