Hm i can Test DSS.

i have a spare FIFO somewhere so maybe i can send you one for testing.

Will it explode when we plug it ?
It is your covox with variable frequency speed ?
I am much more motivated by Tandy DAC DMA implementation and 16 bit mixing, after I stabilize the pattern compression so there is time before working on it.
I have no Tandy 1000 with DAC, but it seems that it is implemented in DOSBOX
I need to change a lot of code in the Buffers management, as the Tandy DAC send an IRQ after a fixed number of samples: 2Kb

I Saw this video from DaveJustDave some days ago:
https://www.youtube.com/watch?v=0oBoZTO0GB0
He demonstrate TanTrack playing Space Debris at 22KHz, but there are horrible replay bugs…
Dave explained that is is a an impressive performance for the hardware...

I downloaded the source code and saw it contains the direct access code to the DAC and the mixing code is Slow (Probable 3x slower than mine)
At 22KHz, the Tandy RL can surely play 10 or 12 channels, it is a 286.

This is shockwav3s dss clone, without the mod, somewhere i also have one where i replaced the NE555 with a tiny13 allowing for variable output rates.
No it won`t burn or explode 😀
The Part afterwards is simply a AD7524 with Preamp and Amplifier.

In the meantime I tried it on yet another 286, Excel Tiger-12, for the fun of it. Despite having the fastest ISA implementation I've ever seen (VGA speed is on par with an overclocked OPTi 895) it worked with GUS classic 3.74 as flawlessly as... well, as it would work on any other properly working hardware. On the other hand, GUS Max 1.8 simply couldn't cope with the system. So don't worry about it.

I've tried many XMs but I won't report all the instability and flawed playback, I see that MODM v11 has a rather rudimentary support. Which is still way better than what GLX offers which I could never get to play any XM at all.

matze79 wrote:

This is shockwav3s dss clone

Does the clone implement the 16-byte FIFO @ 7khz of the real DSS? (if not, then it is a covox clone, not DSS clone)

Its complete FIFO with attached Covox. On The Picture are 2 Devices, the FIFO PCB and my Covox PCB 😉

Hi,

Thanks Alvaro for the test 😀 I believe I am still the only one who tested a GUs on a 8088.
I already corrected many problems on XM files Loading: 16 bit samples, loading bug if there are instruments with multiple sample, Add for some more effects and so on.

.S3M File compression is completed (Note Cut, with Note=255 was causing the last bug)
I am working on the .XM bugs because it is also frustrating for me to have so many .XM not playing at all.

I also progressed on the tests with real hardware, I was able to play modules on a Tandy 1000 EX using the Cvx4: 7.5KHz for 4 Channels and 5KHz for 8 channels. Thanks to the pattern compression, Lots of .MOD files can be loaded completely.
Even at 7MHz, the Tandy 1000 EX is slow, slower than the Atari PC2 or Amstrad 1640.

The Tandy 1000 EX is a 7.16 MHz 8088 whereas the 1640 is an 8MHz 8086. That doesn't sound like much of a difference, but the extra 1 MHz, 2 bytes of prefetch queue, and faster word-aligned accesses add up.

Does the modm code attempt to play .XMs without a GUS? I would think the >64k samples, 16-bit samples, etc. would cause issues for code that was optimized for 8-bit mixing.

Hi,

Tandy 1000 EX also has a slow memory. It is supposed to be faster with the Memory extention card I added. I haver not done full bechmark. (This is weird to do benchmark on 35 YO Machines)
Derek did it at 7KHz on this video with GLX: https://www.youtube.com/watch?v=jVgY7fRG0Vs
BTW, I use his Parallel port connector.

Mod master convert the 16Bit samples to 8 Bit. It also cut the >64Kb Samples
There are not that many .XM with >64Kb samples anyway and 16Bit is useless on 8Bit DAC.

Mod Master XT will never mix 16Bit samples, as it requires to do a MUL to apply the volume.
I also did not support 8Bit samples mixing in USMPlay I guild for 486 computers and nobody complained 😀

Edit: Some more things improved in .XM : Extended effect were not implemented (Bug) Improved period precision (for XM & S3M) Extra fine portamento in progress...

FreddyV wrote:

I believe I am still the only one who tested a GUs on a 8088.

It's still along the road - I'll try it but I only have one non-permanent old PC on test bench at a time and I haven't tried any XT class boards these months. It's also more complicated to get progs/data on them as a shorted Audigy killed the only PCI slot in my main PC and it stopped me from directly copying anything on its SCSI boot HDD. I still can use floppies and a bridge machine or I'll have to flash an XT-IDE ROM and use it in an 8-bit NIC or test a PCI SCSI in my P3 build and copy via that one.

Hi,

Your way to copy files is complex.
I use XT IDE on the XT Machines and simply copy the files using an USB Card reader.

I hope that some more youtubers will start to discover My SW and use it as a test in their videos 😀
Inspire other coders to start wortking on XT / 286 can be nice as well. It has never been as simple to code on these machines with DOSBOX and XT IDE.
I also plan to go to some demo party in france next year and show what XT can do.
The Old scool Demo party are almost only on Amstrad CPC and Amiga party here.

If someone gives me a GUS I can test, too. 😉

FreddyV wrote:

Inspire other coders to start wortking on XT / 286 can be nice as well. It has never been as simple to code on these machines with DOSBOX and XT IDE.

Once you're finished and release the final code, being able to play digital sound effects and/or mods will definitely inspire some people, most likely making new DOS games.

alvaro84 wrote:

FreddyV wrote:

I believe I am still the only one who tested a GUs on a 8088.

It's still along the road - I'll try it but I only have one non-permanent old PC on test bench at a time and I haven't tried any XT class boards these months. It's also more complicated to get progs/data on them as a shorted Audigy killed the only PCI slot in my main PC and it stopped me from directly copying anything on its SCSI boot HDD. I still can use floppies and a bridge machine or I'll have to flash an XT-IDE ROM and use it in an 8-bit NIC or test a PCI SCSI in my P3 build and copy via that one.

Get a nullmodem cable, if needed a USB Serial Converter, Dosbox and Fastlynx / Laplink and here you go.
Dos Machine only needs mode.com on it, so you can bootstrap fastlynx.

Hi,

Here is the release 12, with far better .XM loader and mainly internal changes, like partition compression.

It took a big amount of time to code and debug the compression, but this is really usefull.

22/12/19 Beta 12
- Corrected a .MOD File loading problem (Sequence With patterns=255)
- Increased the UMB Block numbers from 16 to 64. (Memory)
- 16Bit samples loading corrected.
- .MOD, S3M and XM Files Pattern compressed: Loading DOPE.MOD Patterns is no more a problem
-> Partition decoding is much faster.
- Lot of .XM Files Loading bug corrected and Extra Fine portamento Up/Down Added
- Added the Free DOS Memory size on the Option Menu.

FreddyV wrote:

Hi,

Here is the release 12, with far better .XM loader and mainly internal changes, like partition compression.

It took a big amount of time to code and debug the compression, but this is really usefull.

Oh nice. An early Christmas present. I will try to test it later tonight on the BlasterBoard and SnarkBarker.

Oh nice,

i will try it out 😀

hey you implented EMS, why not implent Disk Pagging too ? its not very different =)

As i develop the version 2.2 for 386/486 EMS and UMB is supported.
I have an EMS card in my PC1640.
I tested more the rev 12 yesterday. Partition readind is slower with stardust.mod this music bas à complex partition.

The Mono Display Support reallly needs some polish 😀

Yes.
The display is different between DosBox and a read pc.
But at least it is here. Better than nothing
It was a fast patch made in 1 hour