Giana Sisters STE - New Game!

[AtariZoll]

By my experiences need to correct Microwire mixer settings - default one is with silencer DMA . And correcting volume settings too. May need some experimenting to get good balance.

[Cyprian]

I think all samples are already at max volume and also the microwire. The YM and DMA sound volume don't match up so good.

Did you try changing the mixing method from 1 to 0 ? It should lower the YM volume

Code: Select all

   move.w   #%10000000001,d0   ;mix DMA+YM equally
becomes
   move.w   #%10000000000,d0   ;mix DMA + YM-12dB

Is that really works? I did some tests few years ago with no success - In case of "YM-12dB" I heard only DMA sound

[Zamuel_a]

I tried it and it made no difference. I'm not sure if STEEM is emulating the microwire correctly? I changed the master volume and it affected the DMA samples, but not the YM chip. Changing the bass didn't do anything what I could hear.

[npomarede]

I tried it and it made no difference. I'm not sure if STEEM is emulating the microwire correctly? I changed the master volume and it affected the DMA samples, but not the YM chip. Changing the bass didn't do anything what I could hear.

I'm not sure about Steem, but I know Hatari is emulating it and will lower the YM output in that case;
But it should be tested on a real STE to check if the difference can be heard (there could be an error in the volume level in Hatari)

[Zamuel_a]

I can't assemble it in Hatari. I get program buffer full in devpac and I have exactly the same settings and everything. Strange.

[AtariZoll]

In Steem (v 3.2) DMA is much silenter than normal. In Hatari it is better emulated.
Can't assemble ? Assemble in Steem, and run in Hatari

[Zamuel_a]

Can't assemble ? Assemble in Steem, and run in Hatari

Yes, it's just that it's so much work just for a test. To make a runnable program I first must assemble the main file. When with a hex editor, cut away the TOS header. Count the amount of bytes the file takes and put into the loader program and assemble that, so it's not so fun, hehe.

But if the sound will be different depending on emulator or real machine, it's no use anyway.

[junosix]

vasm http://sun.hasenbraten.de/vasm/ is a good cross-assembler that supports DevPac standards.

Hatari definitely emulates bass/treble so should probably be better at YM/DMA mix levels too.

[dma]

I can confirm that Hatari sound is very close to real STe audio ouput for YM/DMA/Microwire.

[npomarede]

But if the sound will be different depending on emulator or real machine, it's no use anyway.

Not necessarily : if sound is good on real machine, then it will be a real use case of the mixing and it should be easy to fix in Steem (or in Hatari in case it's wrong too, but I think mixing is OK).
Lowering YM when mixing in the emulator is just a matter of applying a coeff < 1 to the YM sample data and a coeff=1 to DMA data, I guess it should be easy to add such a case in Steem if needed.

Targeting the real machine will help improving the emulators when possible.

[Eero Tamminen]

I can't assemble it in Hatari. I get program buffer full in devpac and I have exactly the same settings and everything. Strange.

Do you have same memory amount and TOS version?

What means "buffer full"? It sounds more like Devpac configuration issue (e.g. not reserving enough memory for assembling) than Hatari one...

What Hatari version you're using? Latest Hatari version (v1.7.0) doesn't have any known issues with Devpac, it fixed one GEMDOS HD emulation issue which concerned Devpac (flushing of Fwrite()s so that reads from the same file through another file handle get right contents also before the file gets closed).

Getting OS call traces ("trace gemdos" or "trace os_all" in Hatari debugger) may tell something about the problem.

[punkrulesok]

i've noticed that the STE sound in Steam SSE is better .. not quite as good as SainT though ...

[Zamuel_a]

I can't assemble it in Hatari. I get program buffer full in devpac and I have exactly the same settings and everything. Strange.

Do you have same memory amount and TOS version?

What means "buffer full"? It sounds more like Devpac configuration issue (e.g. not reserving enough memory for assembling) than Hatari one...

What Hatari version you're using? Latest Hatari version (v1.7.0) doesn't have any known issues with Devpac, it fixed one GEMDOS HD emulation issue which concerned Devpac (flushing of Fwrite()s so that reads from the same file through another file handle get right contents also before the file gets closed).

Getting OS call traces ("trace gemdos" or "trace os_all" in Hatari debugger) may tell something about the problem.

Yes it's a Devpac issue, but the files and configuration is the same. I use the same "harddrive" for both emulators. I'm using 1.7.0 and I'm close to filling up the entire memory, so there must be something that takes more RAM in Hatari than in Steem. I'm using the same TOS file to. 1.62

[Zamuel_a]

Ok, here comes two versions. One with the DMA and YM sound equal in volume and one there YM is lower (-12dB). It worked in Hatari, but not in Steem (not tried the SSE version). I haven't tested it on a real machine so maybe it doesn't work at all...

I don't know which version I liked best. So you can decide I think I like it best with the YM volume reduced, but the Giana jump sound is to high. Is it the original Amiga volume or are all samples resampled to max volume?

GianaSTE_DMA.zip

[junosix]

Sounds ace! I haven't adjusted the volume of the samples, the only change I made was to try to match the frequency then resample each to 12.5kHz. There's no headroom to make them louder, so the best option would be to use the lower YM volume version and I can reduce the volume of the samples. I'll have a play with the DMA version over the next day or so until I'm happy about the frequencies I'd chosen and make edits for that if necessary, then reduce the volume a little then upload here.

Good work!

[Cyprian]

Ok, here comes two versions. One with the DMA and YM sound equal in volume and one there YM is lower (-12dB). It worked in Hatari, but not in Steem (not tried the SSE version). I haven't tested it on a real machine so maybe it doesn't work at all...

I don't know which version I liked best. So you can decide I think I like it best with the YM volume reduced, but the Giana jump sound is to high. Is it the original Amiga volume or are all samples resampled to max volume?

GianaSTE_DMA.zip

checked on Mega STE and in case of YM_Lower there is no YM music during horizontal scroll
tomorow I can check on TT and STE

[MrMaddog]

Ok, here comes two versions. One with the DMA and YM sound equal in volume and one there YM is lower (-12dB). It worked in Hatari, but not in Steem (not tried the SSE version). I haven't tested it on a real machine so maybe it doesn't work at all...

Yep, works fine on Steem SSE as well as Hatari, in fact the game feels much smoother and faster on SSE at 60Hz refresh rate. Plus the YM reduced version sounds better because I can make out the jump sounds better...

Great work, this is exactly how I imagine a STe port of Giana Sisters would be!

[Zamuel_a]

checked on Mega STE and in case of YM_Lower there is no YM music during horizontal scroll
tomorow I can check on TT and STE

Hmm strange? What could cause that.
TT will not work since it has no blitter.

[Cyprian]

checked on Mega STE and in case of YM_Lower there is no YM music during horizontal scroll
tomorow I can check on TT and STE

Hmm strange? What could cause that.

did you use that code in "YM_Lower"?

Code: Select all

   move.w   #%10000000000,d0   ;mix DMA + YM-12dB

"mix DMA + YM-12dB" mode didn't work for me on real hardware. I'll try to prepare a small test APP to re-check that.

TT will not work since it has no blitter.

ok, I forgot that game needs blitter.

[npomarede]

did you use that code in "YM_Lower"?

Code: Select all

   move.w   #%10000000000,d0   ;mix DMA + YM-12dB

"mix DMA + YM-12dB" mode didn't work for me on real hardware. I'll try to prepare a small test APP to re-check that.

As said above, it's possible mixing mode 0 doesn't produce any YM sound on Atari STE and Hatari could be wrong ; the STE doc is not very precise and it seems people had different results when using it (perhaps YM sound is lowered too much on a real STE, and you can only hear it when volume reg is between 10 and 15).

Things to test on an STE : use mixing mode 0 and play an empty sample (bytes=0) + YM sound on voice A. Now change the YM volume (eg reg 8 for voice A) between 0 and 15 and check if any YM sound can be heard

[Cyprian]

Things to test on an STE : use mixing mode 0 and play an empty sample (bytes=0) + YM sound on voice A. Now change the YM volume (eg reg 8 for voice A) between 0 and 15 and check if any YM sound can be heard

thanks for the hint, I'll check that.

[junosix]

Had a decent go on the STE version this morning, works really well with the samples!

There are a couple of samples I need to redo (the warp sound is slightly lower pitch when compared to the Amiga version, I just need to redo the sample rate on that one, and I'm not quite so sure about the enemy death sound. The Amiga versions of the lowest sample-rate sounds are quite muffled compared to when they're upsampled and played on the STE though so it might just be that making it sound a little different).

To get the samples to settle with the audio on the YM_LOWER version I'd need to reduce them by around -7 to -8dB, which would have quite an effect on the dynamic range of the sample (sounds ridiculous to worry about dynamic range on a low sample rate 8-bit sound effect!), but if the EQUAL version was modified to adjust the volume data just before it gets sent to the YM chip (either by subtracting a little or shifting the data to the right) then could the music just be lowered slightly to match the samples at their unmodified volume? Keeps it as pure as possible without both -12dB'ing the YM and -8dB'ing the samples. Are there enough cycles available in the VBL period to do that rather than having to hunt through the music data and change all the volume levels by hand?

Let me know if you want me to reduce the volume or whether you think the other method is better and I'll make the changes and upload them this evening.

[Zamuel_a]

If I can find were in the code to reduce the volume of the three YM channels, when it would be possible. It's only a simple sub, so it won't take much cpu time.

[npomarede]

If I can find were in the code to reduce the volume of the three YM channels, when it would be possible. It's only a simple sub, so it won't take much cpu time.

You need to intercept the part that writes to reg 8,9,10 in $ff8802.
Doing a sub is not very good, because it can give negative volume values that you must correct and also because the volume is not linear in the YM2149 (subtracting 3 to vol=15 will lower the sound much more than subtracting 3 to vol=8).
The better approach is to have a conversion table with 16 entries for each volume ; to lower volume, you will for example do "-1" to volume 0-7, "-2" to volume 8-12 and "-3" to volume 12-15 (those values are not correct, that's just an example because the volume variation is logarithmic in the YM, not linear). The conversion table will depend on how much percentage you want to reduce the global volume.

[AtariZoll]

What about doing sound in play part complete with DMA ? Then would be no problems with sound levels, only 'small' problem of mixing diverse samples if they play at once. Or just using L-R channels ?