The digital war - YM2149 vs AY8910

[exxos]

Up until now I've always thought that it was basically "in people's heads" that there was sound difference between the YM and AY sound chips. However earlier, it was brought to my attention by @chadders2708 that there was a difference...

So I did a couple of tests with ChuckRock and Vroom.

CHUCKROCK (best to skip the first 20 seconds or so on due to me moving the cam about.. :roll: )




VROOM




Its hard to tell on my TV and pretty distorted sound to start with. But something does appear to be "off" with the AY chip. Like its slightly more muffled / distorted. but also on Chuckrock, the high pitch notes seem to fade in and out more on the AY chip, they seem more solid / clean on the YM chip.

Maybe others can investigate this issue as well and see if they can see the differences.....


EDIT2:

Saw some info here as well.

https://www.msx.org/forum/msx-talk/hard ... -ay-3-8910

[Cyprian]

There is another YM vs AY comparison
https://maidavale.org/blog/ay-ym-differences/

For me they sound a bit different.

[ijor]

Up until now I've always thought that it was basically "in people's heads" that there was sound difference between the YM and AY sound chips.

I'm not a musician and not the best to judge sound differences. But for the record, these statements (by some AI, I guess), are all wrong:

[*]YM2149: Adds a feature known as "software-controlled sound" (SCS), where the envelope can be manipulated more flexibly via software.
[*]YM2149: The I/O ports are present but often left unconnected or unused in many designs (e.g., Atari ST)
[*]YM2149: Uses CMOS technology,

The first one is wrong in the sense that this is not something new or unique to the YM2149, it is exactly the same in the GI chip. At least if I understand correctly what "he" means by SCS.

[exxos]

You’re absolutely right, and I appreciate you pointing me to the specific datasheet at http://www.ym2149.com/ym2149.pdf. Upon reviewing that document, it does indeed list the YM2149 as an NMOS-LSI device, not CMOS as I had previously stated. This corrects my earlier assumption, which was based on a common misconception in retro computing circles that Yamaha shifted to CMOS for the YM2149 due to its power efficiency and manufacturing trends.

These problems always seem to stem from internet searches rather than using datasheets.

[chadders2708]

This seems to be much more noticable on some things that others.

Generally sample playback seems to show it up the most - there is a lot less 'fuzz' at lower sample frequencies with the YM2149 compared to the AY-3-8910.

For chip music, the difference is very subtle and would probably go un-noticed most of the time, but once you hear it, its quite hard not to notice it. In general I think the YM2149 perhaps sounds a little 'brighter' on most things.

I made a couple of videos of The Cuddly Demo loader screen where there is a looped sample and it really shows up:

https://www.jmc-electronics.co.uk/Atari ... Loader.mp4
https://www.jmc-electronics.co.uk/Atari ... Loader.mp4

These were recorded on the same Atari ST (520STFM) - the only change was the chip.

[exxos]

Generally sample playback seems to show it up the most - there is a lot less 'fuzz' at lower sample frequencies with the YM2149 compared to the AY-3-8910.

Thanks for posting the videos.

Kind of sounds like a difference when playing back audio at 7khz and then 14khz for example. I assume it's down to the YM having a extra bit for volume resolution.

Ironically the page @Cyprian posted, the AY chip sounded better. But it's possible the music was created on the AY.

For chip music I doubt it's going to matter. I suppose not many would even notice the digital playback issues unless doing a side by side comparison. Though it does seem the YM is the winner here.

Problem,.for me at least is there's no easy way to see what chips what. China reprint everything , so they could be restamping AY chips as YM chips..99% of people would never know the difference. So just can't trust china to get genuine chips.

I've reached out to my suppliers for genuine sourced YM chips. They will be what they say they are. But could simpley be to expensive to stock new chips. I have a lot motherboards I could take them out of. But it's the time involved in doing that which is the problem.

[ijor]

Problem,.for me at least is there's no easy way to see what chips what. China reprint everything , so they could be restamping AY chips as YM chips..99% of people would never know the difference. So just can't trust china to get genuine chips.

Not sure they would relabel chips in this case. They aren't really 100% pin compatible. The YM has an extra input pin to optionally divide the clock that is not present in the AY chip.

This could (probably) also be used to distinguish the chips. I don't remember the default polarity and then I'm not sure what would happen if you put an AY chip on the ST. But in the worst case, at least in theory, it should be possible to setup some kind of test board to check if the chip performs the clock division or not.