STE DAC circuit

[exxos]

dac.JPG

I was just having a quick look at the STE DAC circuit..

Is the LMC chip really used for anything ? I was thinking the DAC has a lot of stuff already, just have those amps as the main audio output line, and mix the YM audio into it as well. We don't really need the LMC to mix or switch audio I think... ?

I know the LMC can be used for audio selection, but don't really see why this would be needed anyway. It does have bass and treble but I think it is probably never been used. I know there was some talk while ago about the chip being a total overkill for what it does.

I'm not really sure what programs even use the DAC circuit ?. The only one I know of is where Petari patched Xenon2 play audio and YM at the same time.

My thoughts are that the Suska MMU is actually a STE MMU.. Ultimately when this core is used, we should be easily able to add the DAC playback circuit.. Though I don't really want to use the LMC chip (hard to get for one thing) as I think it just over complicates things. The DAC circuit itself can be updated with more modern parts, and likely with fewer parts making the circuit a lot simpler and easier, aka, lower PCB footprint and costs.

[exxos]

Actually... While half thinking about the DSP with external RAM....

There probably isn't any reason not to add a super dac circuit with its own ram. slowdown playing music from hard drive is very low anyway.. so not much CPU overhead using DAC...but of course needs fast hard drive.. But we could add its own 16MB RAM and have a switch to play from "DAC RAM" instead of ST RAM. We could also set whatever sample speeds we wanted..

It would take more RAM, but even 16bit audio could be done (currently 16bit is into 2x8bit dacs). Of course data would need to be loaded from hard drive into dac ram...

Also possible we could use 4x8bit dacs, and configure them however we want.. maybe 4x8bit channels, 2x16bit channels.. or both configs set in a config register...

Or do both STE DAC playing from ST RAM, and have 4 more 8bit channels on external "DAC RAM" making 6x8bit channels of audio. Then at least we retain some STE backwards compatibility and bring in new channels in addition.

... just thinking out loud :)

[chlu600]

I think that the most interesting sound addition is to "simply" add a SID-chip.
That will for sure create a big interest, even outside the Atari community. :yay:

https://en.m.wikipedia.org/wiki/MOS_Technology_SID

[exxos]

I think that the most interesting sound addition is to "simply" add a SID-chip.

I did look at SID about a week ago actually, but the best demo music with it, didn't sound anything past YM stuff with digidrums to me :shrug:

[troed]

The STE does not have an MMU, or a GLUE. It has both combined into the GSTMCU. The STE DMA sound is done by the much expanded Shifter (GST Shifter).

https://info-coach.fr/atari/hardware/STE-HW.php

I'm assuming that's what you meant, just pointing it out for clarity :)

For our video mode expansion thoughts we need to stay with the separate MMU and GLUE. I don't see any harm in using the GST Shifter anyway though - expanded palette is a nice bonus in any case.

(Unless Smonson's FPGA Shifter, or a future variant of it, is simply the better solution as default)

[exxos]

Yeah, I know what you mean. I'm just saying the Suska cores are actually STE cores, so DAC audio can be added. The suska MMU core has the "STE" stuff built into it. It just needs the actual DAC hardware adding.

MMU & GLUE will need to be separate for a while, while the cores are being tested out anyway. Long term, we need a better shifter with proper colour & resolution increases. We will have several times the bandwidth with a 50MHz system.. we just need a shifter to make use of it..