Two Issues about the AWE32

[mace]

Here's a picture:

I got an AWE64 Value card with something really neat called the "SimmConverter" which was a homemade board that allows the use of any old 72-pin SIMM you want!  No expensive specialty RAM modules!

So my AWE64Value has a 32MB SIMM, allowing 28MB of soundfont RAM.  Great for tracker music! 
[/quote]

Yes I've seen that.
It eats up alot of space because the simm sticks out at a 90 degree angle from the board.
So if you have say, a LAPC-I under it, it wont fit.

[Juho Sippola]

I've never found much use for the AWE RAM. When I first got the AWE32 around 1998 I loaded it with 8 megabytes of RAM and used it with various more or less mediocre selfmade soundfont banks. The bigger Creative banks from SB Live! weren't much of a delight either, especially since soon afterwards I got the SCC-1 which blew them all out of the water.

Later in 2001 I was able to obtain AWE64 Gold which I still have. I wanted it for its better output quality and it indeed didn't hiss. The board came with 8 megabyte expansion (total 12 megabytes) and to date, I haven't found ANY viable use for that memory capacity. The MIDI wasn't any better than before.

Can you guys list me some viable uses for extended AWE memory?

[mace]

I've never found much use for the AWE RAM. When I first got the AWE32 around 1998 I loaded it with 8 megabytes of RAM and used it with various more or less mediocre selfmade soundfont banks. The bigger Creative banks from SB Live! weren't much of a delight either, especially since soon afterwards I got the SCC-1 which blew them all out of the water.

Later in 2001 I was able to obtain AWE64 Gold which I still have. I wanted it for its better output quality and it indeed didn't hiss. The board came with 8 megabyte expansion (total 12 megabytes) and to date, I haven't found ANY viable use for that memory capacity. The MIDI wasn't any better than before.

Can you guys list me some viable uses for extended AWE memory?

Well, the chaos banks are supposedly good.
It's 8 MB and you can get it  here

Haven't tried it myself tho.

[BlueMax]

The AWE32/64 memory was great for those of us who wrote tracker music - instead of mixing the channels with the CPU, the samples were loaded into soundcard RAM and mixed on the CARD instead!  Zero CPU usage!  (A Big deal in the 386/486 era!)

They also let you use the hardware chorus/reverb of the card, too!

[Cloudschatze]

Sheesh I wish people would quit beating around the bush and just say what card they recommend and why, rather than saying which cards they don't.  I don't know where to begin to respond to these tomes of information that conclude nothing.

Have an opinion guys!

As far as the CQM goes, if you say it is different, then it is different.  Perhaps the time would be better spent comparing the output of the CT-1747 with the YMF-262. 

Here, perhaps I can address both of these.

I installed a CT2760 AWE32, which has the CT1747 IC, and... it sounds like a Yamaha OPL3. Though I stopped short of making any comparison recordings, I didn't hear any differences that would lead me to believe that it is anything but OPL3 output.

As a pleasant surprise, it also did not exhibit the nasty, high-pitched, metallic "aliasing" that the CQM-based AWE64 cards produce during FM playback (that I've tested) .

Here is an example of what I'm referring to, recorded from an AWE64:
http://www.reama.com/elw/pied3.wav

Don't feel bad if you can't hear it - I can only hear it with one ear.


Now then, about that opinion... 

I have a rather low regard for most of Creative's cards, especially the SB16 line. Lately though, certain things are changing my opinion (like the CT2502).

Today, during my testing, I think I found the one Creative card that, in my opinion, deserves its name:

The CT2760 - AWE32.

It might just be the spiritual successor to the original, "killer card."

Why?

• An onboard OPL3, to which reverb and chorus can be applied.
• An onboard EMU8000 synthesizer, expandable with RAM.
• A WaveBlaster header.*
• SB16-compatible digital audio.
• Surface-mounted ASP/CSP (not that it matters...)
• Non plug-and-play.
• Doesn't smell like pancakes.

* Since I know it is going to be asked, I mounted an SCD-15, and tested with both Quest for Glory III and Tyrian. I didn't encounter any problems, hanging notes or otherwise.  I can't attribute this to just the AWE32, at this point, and need to do some additional testing, inclusive of other daughterboards and software. Switching Tyrian's output, on-the-fly, between FM/EMU8000/SCD-15 is rather neat.

And the downsides?

• Output is not as quiet as that produced by the CT2502-based cards.
• Several daughterboards cannot be properly mounted, including the RIO and HOMAC.
• Becomes quite warm (and it appears that the onboard amplifier cannot be disabled).
• Doesn't smell like pancakes.

Though far from perfect, the CT2760 does live up to some degree of "AWE-someness," and I think I may just leave it in the "synthe" PC..

For a little while, at least.

[mace]

-SNIP-

• Doesn't smell like pancakes.

No pancakes? 
Does the AWE64 gold smell like pancakes? Because I like pancakes. 

Anyway.
I think I know the aliasing effect you describe. It's a very high pitched tone, so my young ears can hear it,
but I can imagine the "older" folks won't be able to hear it.

Maybe adding a simple lowpass filter can solve it, who needs anything over 15 Khz for FM music anyway?

About the hanging notes bug: Hexen seems to be most affected by this.
Use the "-warp 02" command switch, you can also use the demo for this purpose.
The hanging notes should be immediately apparent.

[NewRisingSUn]

As a pleasant surprise, it also did not exhibit the nasty, high-pitched, metallic "aliasing" that the CQM-based AWE64 cards produce during FM playback (that I've tested).

You've said that you couldn't hear it in the AWE64 Gold recordings I sent you.

Don't feel bad if you can't hear it - I can only hear it with one ear

This is a bad example. The high-pitched sound here is the instrument itself, even on a real OPL3 chip. CQM just sounds brighter in general, which is why it's more apparent, but it's there when playing this song on the original chip as well.

And don't blame my ears --- I can hear all those stupid "teenage repellant" things as well, so my ears are not the problem.

But while we're at it --- is it possible to replace the CQM-generating chip on the AWE64 Gold with either a real OPL3 or that exactly-the-same-sounding CT chip?

About the hanging notes bug: Hexen seems to be most affected by this.
Use the "-warp 02" command switch, you can also use the demo for this purpose.
The hanging notes should be immediately apparent.

Is that the one with the cheesy "Greetings, mortal" speech? Just to make sure I've got the right level.

I first listened to the playback from a Roland SCC-1 at i/o address 0x330. No hanging notes there, obviously.
Then I listened to the playback from a Roland CM-500 connected to the AWE64 Gold's external MIDI out, accessed through the MPU-401 UART at i/o address 0x300. Exact same output, no hanging notes there.

So the AWE64 Gold does not seem have the problem. Since it has no Wave Blaster connector, I can't try that.

Any other games to check?

[Cloudschatze]

As a pleasant surprise, it also did not exhibit the nasty, high-pitched, metallic "aliasing" that the CQM-based AWE64 cards produce during FM playback (that I've tested).

You've said that you couldn't hear it in the AWE64 Gold recordings I sent you.

You're right. I had hoped to exclude your AWE64 Gold recordings with the statement, "that I've tested," but perhaps I should have just singled out the AWE64D instead.

Don't feel bad if you can't hear it - I can only hear it with one ear

This is a bad example. The high-pitched sound here is the instrument itself, even on a real OPL3 chip. CQM just sounds brighter in general, which is why it's more apparent, but it's there when playing this song on the original chip as well.

I realize that CQM is brighter, but I've never heard the same noise from anything OPL-based. I've wondered if, perhaps, the OPL3 incarnations I've heard are just more-heavily filtered. It does seem to be the instrument itself, as you've said, but again, as an overtone that shouldn't be there. I'd write it off as something possibly wrong with my hearing, but for the fact that, when I had originally made that WAV recording, my sister-in-law complained of the high-pitched sound from the opposite end of the house.

I'm just glad to not hear it with every card, else I'd probably look for a new hobby. 

But while we're at it --- is it possible to replace the CQM-generating chip on the AWE64 Gold with either a real OPL3 or that exactly-the-same-sounding CT chip?

Sounds messy. I wonder if it would be easier to somehow disable the CQM on the Gold, and use another card for FM playback. Sure would be nice to see Creative's datasheets.

[mace]

About the hanging notes bug: Hexen seems to be most affected by this.
Use the "-warp 02" command switch, you can also use the demo for this purpose.
The hanging notes should be immediately apparent.

Is that the one with the cheesy "Greetings, mortal" speech? Just to make sure I've got the right level.

I first listened to the playback from a Roland SCC-1 at i/o address 0x330. No hanging notes there, obviously.
Then I listened to the playback from a Roland CM-500 connected to the AWE64 Gold's external MIDI out, accessed through the MPU-401 UART at i/o address 0x300. Exact same output, no hanging notes there.

So the AWE64 Gold does not seem have the problem. Since it has no Wave Blaster connector, I can't try that.

Any other games to check?

I think so, Great H. has more info.
Please see this vogons thread for more info.

Also: the waveblaster conn. is basically hardwired to the gameport, so it should not make any difference.

But while we're at it --- is it possible to replace the CQM-generating chip on the AWE64 Gold with either a real OPL3 or that exactly-the-same-sounding CT chip?

Sounds messy. I wonder if it would be easier to somehow disable the CQM on the Gold, and use another card for FM playback. Sure would be nice to see Creative's datasheets.

I've been thinking, and I want to track down a (cheap, because they aren't very great) FB-01, and somehow get adlib to work on that, maybe cutting traces and rerouting them to a MIDI interface somehow. But it's probably too difficult.

[NewRisingSUn]

Please see this vogons thread for more info.

I have. No hanging notes with Hexen -warp 02 or DOOM E1M1. AWE64 Gold: CT4540, DSP 4.16 with both MIDI music and PCM sound.

I wonder if it would be easier to somehow disable the CQM on the Gold, and use another card for FM playback.

I only have two ISA slots, one for the AWE64 Gold, one for the SCC-1. I could replace the SCC-1's sound source with the CM-500's excellent GS part, but I'm not satisfied with the AWE64 Gold's UART-only MPU-401 interface.
I would need a card that has both a real OPL3 and an intelligent-mode-compatible MPU-401, or failing that, one that at least returns ACK on all MPU-401 commands.

[5u3]

As a pleasant surprise, it also did not exhibit the nasty, high-pitched, metallic "aliasing" that the CQM-based AWE64 cards produce during FM playback (that I've tested) .

Here is an example of what I'm referring to, recorded from an AWE64:
http://www.reama.com/elw/pied3.wav

Don't feel bad if you can't hear it - I can only hear it with one ear.

This is a bad example. The high-pitched sound here is the instrument itself, even on a real OPL3 chip. CQM just sounds brighter in general, which is why it's more apparent, but it's there when playing this song on the original chip as well.

I realize that CQM is brighter, but I've never heard the same noise from anything OPL-based. I've wondered if, perhaps, the OPL3 incarnations I've heard are just more-heavily filtered. It does seem to be the instrument itself, as you've said, but again, as an overtone that shouldn't be there. I'd write it off as something possibly wrong with my hearing, but for the fact that, when I had originally made that WAV recording, my sister-in-law complained of the high-pitched sound from the opposite end of the house.

I agree with NewRisingSUn on this, the "high-pitched noise" in the pied3.wav example sounds normal for OPL3. But it surely is more present on AWE64 cards than on old SB16s, probably because in contrast to many SB16s, the AWE64's mixer and output amps are actually capable of reproducing sound above a few kHz.

Maybe we're all talking of different things when we mention the "high-pitched aliasing noise" on CQM-based cards. Here are my example files for comparison:

DOOM E1M1 on AWE64 with high-pitched aliasing noise
DOOM E1M1 on AWE64 without high-pitched aliasing noise

The CT2760 - AWE32.
[...]
* Since I know it is going to be asked, I mounted an SCD-15, and tested with both Quest for Glory III and Tyrian. I didn't encounter any problems, hanging notes or otherwise.

Board member Locutus recently tested a CT2760 and reported hanging notes:

CT3990, CT3980 and CT2760 all have the hanging notes bug and SBMPU401.EXE or MPU-ON.COM won't fix it. 

[NewRisingSUn]

Maybe we're all talking of different things when we mention the "high-pitched aliasing noise" on CQM-based cards. Here are my example files for comparison:

Indeed, mine always sounds like the example without aliasing. So do these two examples come from different cards, or what did you do to turn on/off the aliasing? Considering that you can add reverb/chorus to the FM output, maybe under some circumstances the FM reverb/chorus registers contain crap values, causing these artifacts. Just my speculation, I haven't looked deeply into this.

[Cloudschatze]

I agree with NewRisingSUn on this, the "high-pitched noise" in the pied3.wav example sounds normal for OPL3. But it surely is more present on AWE64 cards than on old SB16s, probably because in contrast to many SB16s, the AWE64's mixer and output amps are actually capable of reproducing sound above a few kHz.

But it isn't from an OPL3. The example was recorded from an AWE64D, with CQM. I've never heard anything like that from an OPL.

I'm convinced that you guys are not hearing it. I work in a white-noise saturated environment, and at a low volume, from half a room away, and when all of the other frequencies in the recording have been drowned-out, the tones are still piercingly audible to my right ear, clear as day, non-musical as ever. My left ear, however, I can place right up to the speaker and still not hear them.

I'm not sure if it's possible, but I'll try and filter everything else out of the recording this evening.

Maybe we're all talking of different things when we mention the "high-pitched aliasing noise" on CQM-based cards. Here are my example files for comparison:

DOOM E1M1 on AWE64 with high-pitched aliasing noise
DOOM E1M1 on AWE64 without high-pitched aliasing noise

I'll have to listen to these later...

Board member Locutus recently tested a CT2760 and reported hanging notes:

CT3990, CT3980 and CT2760 all have the hanging notes bug and SBMPU401.EXE or MPU-ON.COM won't fix it. 

I know. I am going to test with Hexen later as well...

Busy evening ahead.

[NewRisingSUn]

I'm convinced that you guys are not hearing it.

Nonono. We are hearing it; we just don't think that it's something out of the ordinary. See it this way: PIEP3.BMF sounds exactly with my AWE 64 Gold as in your .wav file. You acknowledged that my AWE 64 Gold does not have the aliasing problem. Therefore, it follows that the obnoxious frequencies in your .wav file are not aliasing but just what the song sounds like with CQM.
(DosBox sounds basically the same, but of course that's not saying much, since its FM core is crappy as hell. Not that they give a damn or anything.)

[5u3]

Indeed, mine always sounds like the example without aliasing. So do these two examples come from different cards, or what did you do to turn on/off the aliasing? Considering that you can add reverb/chorus to the FM output, maybe under some circumstances the FM reverb/chorus registers contain crap values, causing these artifacts. Just my speculation, I haven't looked deeply into this.

Both examples were recorded on the same card, a CT4390 AWE64 Gold. The aliasing noise occurs when I initialise or reset the EMU8K with the Master Mixer 5.11 utility I mentioned earlier in this thread. This also happened with old versions of Creative's AWEUtil program. Unfortunately I can't remember the version numbers any more, and I have not been able to reproduce the issue with newer AWEUtil 1.3x versions.
Affected cards are AWE32/64 and SB32 with CQM synth, so the obvious explanation is that it really has something to do with the EMU8K reverb/chorus effects applied to the FM output.

But it isn't from an OPL3. The example was recorded from an AWE64D, with CQM. I've never heard anything like that from an OPL.

Yes, I'm aware of that, but this kind of noise can also be heard on real OPL3 chips. It's just more apparent with the CQM synth.

Do you mean this sound?

[Cloudschatze]

Do you mean this sound?

Yes, exactly. What is that?

If an OPL3 will produce the same, albeit less noticeably, I'm glad to have never noticed.


By the way, I tested the CT2760 (DSP 4.12) with "hexen -warp 02," and didn't experience any hanging notes with either the SCD-15 or Soundscape DB that I tested with. Anything else I should try?

Is the MIDI issue processor-speed related, perchance?

[mace]

I've recieved my AWE64 Gold today, and I'll do some testing as well, but first I'm going on vacation for a week.
So in a week or two I'll try to get to the bottom of the problem.

I'll use praat which is a scripting language designed for audio processing, and maybe I can make a filter for that noise.

PS: Too bad the seller of the AWE64 was a smoker, no pancakes for me. 

[Locutus]

By the way, I tested the CT2760 (DSP 4.12) with "hexen -warp 02," and didn't experience any hanging notes with either the SCD-15 or Soundscape DB that I tested with. Anything else I should try?

Is the MIDI issue processor-speed related, perchance?

To your first question: You could try X-Wing Collector's Edition. If you don't encounter hanging notes during the intro, you will most likely hear some during missions. On my Pentium 200 I have tested several Sound Blaster revisions and they produced quite different results. CT3990, CT2940 very quickly came up with some hanging notes. The supposedly bugfree CT3980 (according to tests on Vogons) also gave me stuck notes with X-Wing CD, just took some minutes of gameplay.

CT2760 also has the bug. I don't know what's the case with the AWE64s, since I haven't tested them.

The only Sound Blaster with WB connector I have tested that didn't produce stuck notes, was the Non-PNP CT1740 SB16. Unfortunately it is very noisy and I would never put a DB50XG on it... maybe a Waveblaster II, but not a Yamaha board.

I suppose, this issue could be CPU-speed-related. Maybe I'll do some testing with an older 486 machine soon, but my SW60XG and SCC-1 both work fine with X-Wing CD.

Regards,
Christoph

[Cloudschatze]

I'd love to hear a definition as to what the "bug" really is.

Roland's MPU-401-compatible devices have a 1700 byte UART MIDI IN buffer. My understanding is that the Creative cards either do not implement the same size buffer, or do not implement it "correctly." I have not found any definitive information to support either statement.

I tested the CT2760 in a 486-based PC. Perhaps it's reasonable to assume that games running on my system are simply not sending MIDI data to the UART interface as quickly as yours, and so, do not overrun whatever buffer is actually present on the Sound Blaster. Perhaps not.

So... who has an X-Wing ISO for me?

[5u3]

Is the MIDI issue processor-speed related, perchance?

I tested various cards at system speeds between 25 and 500 MHz and ISA bus speeds between ~8 and 12 MHz. The bug was always the same.