Atari ST music vs Amiga music.

[ATARI7800fan]

What is your opinion of the musical capabilities of the two systems. I noticed while listening to the intro's to Turrican one and two that a lot of people. I am guessing amiga fans were saying those videos were fake just because it said original amiga graphics and then listed the designers name. They could not belive the music was coming from a actual Atari ST. Is it a large group of people that underestimate the Atari ST when it comes to sound or just a small group. Love how some programmers could really push the Atari ST. Seems these days developers do not try as hard. Plus it seems with technology, it is always advancing so they do not have to work hard. What happened to having a computer part last 5 or so years instead of becoming obsolete within a few months.

[+wood_jl]

I'm a fan of the ST computers. But I'd be a moron if I didn't say the sound was disappointing. However, sometimes it isn't, and I agree that it's kind of neat when someone can wring out cool sound from the old dog - makes it worthy of serious appreciation. Overall, I'd say that at least the ST sound was "adequate" for gaming and little plinky game music, so it's not that it totally sucked, just wasn't the best.

 

I think it must be difficult to get it to sound quite good, which is why it didn't a lot of the time.

 

I think the Amiga sound was seriously good for the era, though.

[svenski]

I think in the days when the Amiga and ST were selling side by side it was obvious which one had the better sound and graphics capabilities (apart from the early titles that were just quickly ported onto the Amiga from the ST version) and the Amiga , in my opinion, would win pretty much hands down every time.

 

Having said that, there were and still are good and bad examples of music/sound on both both platforms and not having had any Amiga for years I find I now appreciate the ST a lot more than I did back in the day.

[+wood_jl]

I think in the days when the Amiga and ST were selling side by side it was obvious which one had the better sound and graphics capabilities (apart from the early titles that were just quickly ported onto the Amiga from the ST version) and the Amiga , in my opinion, would win pretty much hands down every time.

 

Having said that, there were and still are good and bad examples of music/sound on both both platforms and not having had any Amiga for years I find I now appreciate the ST a lot more than I did back in the day.

 

 

Indeed. Through the prism of history, the gulf between them seems relatively small now that modern hardware is millions of light-years beyond them. Some fanbois will argue to their deaths, and somewhere grumpy old men are still arguing whether or not the 1928 Ford is better than the 1928 Chevrolet, while most drivers on the road would regard them as strikingly similar relics.

[oky2000]

Hmmm well depends how you mean.

 

1. ST samples will never sound as good as Amiga samples end of story, regardless of how many channels you want.

2. Amiga has no actual soundchip just 4 digital to analogue converters...ST chiptunes can only sound like the YM chip

 

But this doesn't mean anything artistically.

 

But the point is the ST will sound like an Amstrad/Sinclair 8bit micro and the Amiga can sound like anything you want (ie haunting pan pipes of Shadow of the Beast of nightclub quality trance tunes of Super Stardust)

 

If you are asking which machine had the best tunes, and therefore the greater number of talented composers, this is a whole new kettle of fish to what each has technically too

[+DarkLord]

I'd have to agree with the consensus...and having owned both machines,

I'd say that the Amiga's sound was generally better.

 

Having said that though, I still like a lot of the ST's sounds. The

pipes from SOTB were mentioned. I've played the game on both, and

there's no doubt the Amiga version is better, but...I *like* the ST

versions music in the intro. Also, I love the pounding soundtrack to

StarDust on my Mega STe (of course that's an STe, but SD was brought

up). Several arcade games with the "chiptune" sound were fine by me,

I like the music to "Venus, the Flytrap", for example. Ultima 3's

dungeon music is unforgettable. I love the bard's music from the

Bards Tale, etc, etc,...

 

Keep in mind - a trained musician, I'm not, just "IMHO" and "YMMV".

 

[kool kitty89]

Hmmm well depends how you mean.

 

1. ST samples will never sound as good as Amiga samples end of story, regardless of how many channels you want.

2. Amiga has no actual soundchip just 4 digital to analogue converters...ST chiptunes can only sound like the YM chip

Yes, and you can software mix Amiga much more easily too, more than the STe for that matter (since you have 2 hardware stereo channels or 4 hardwired semi-stereo rather than 1/2 hardwired stereo channels on the STe -or 2 25 kHz multiplexed/interleaxed stereo/4 semistereo channels)

 

And why downplay the Amiga's sound hardware? It's not just some bare DACs like some arcade games (usually MCU driven), Covox/PC parallel port hacks, or what the Genesis has in the YM2612 (or PC Engine/TG-16 with direct 5-bit DAC mode enabled -also able to pair for 10-bits like the Amiga does for 14), it's full hardware PCM generation/mixing with pretty high resolution/range sample rate (used for pitch control as well for hardware 4 channel MOD -opposed to software scaling or fixed pitch playback needed for mixing more channels): hardware DMA loading 4 streams of signed linear 8-bit PCM samples with 0 CPU overhead (other than setting the samples -or for musical instruments, setting loops and sample rate for notes) and 64 step logarithmic (decibel) volume.

The only thing really lacking is that they hardwired the stereo and didn't allow per hardware channel panning (not even hard panning, let alone proper 4-bit L/R volume like the PCE/TG-16 or SNES -a shame they didn't replace the per-channel volume with per-channel L/R smooth panning).

Per-channel filtering would also have been nice. (especially SID-like filter control)

 

But the point is the ST will sound like an Amstrad/Sinclair 8bit micro and the Amiga can sound like anything you want (ie haunting pan pipes of Shadow of the Beast of nightclub quality trance tunes of Super Stardust)

Amiga can't fully emulate the YM2149 at its highest strengths: the sample rate is too low, and while a human won't be able to tell the difference at the high pitches, they may at low sample rates (you'll either hear the harmonics of sampled square waves -or other chip sounds- or you'll have muffled sound with the LED filter enabled -and the forced filtering on top of that). On top of that, there's more CPU resource to manage very small, very tight samples for such small waveforms (or you could use redundantly long samples).

 

Likewise you won't be able to fully match any high-sample rate sound chip/synthesizer or synths with hardware effects not very practical to do in software or presampled (unless you're talking demos and not in-game or parallel with lots of other processes). It can do a lot of things that plain hardware synth can't (depending on the case of said synth... 6-op FM is pretty damn capable for a huge range of instruments, then there's wavetable -not sample- based additive stuff and various combinations of other forms of synth -like the MT-32, various Ensoniq synth chips, etc but those were all much later and/or more expensive -vs sommon arcade/home computer level synth chips like Yamaha's 4 and 2-op FM offerings -they launched the YM2151 and YM2203 a bit before the Amiga was released -the latter is actually a fully compatible successor to the ST's YM2149 I/O and all )

 

Even the SNES's sound system has considerable disadvantages in replicating hardware synth, but with its own advantages as well, jsut as all PCM based sound has. (like the Amiga, the SNES also has sample rate limitations, but the forced filtering is the bigger issue for ruining any chance of sharp, crisp sounds -the forced interpolation also hurts in some cases- and the forced compression adds to that -no support for uncompressed PCM and only the BRR ADPCM format used, better for some things, worse than linear 4-bit PCM in some others -namely very low rate samples where uncompressed 8-bit would also be inferior to 4-bit at the same bitrate and double the sample rate- )

 

OTOH, it's also a lot more foolproof as much synth hardware takes a lot more skill and experience to use well while PCM/sampel based stuff is much more flexible in terms of the user literally selecting what they want (unless they're stuck with a default sample set or such). But that doesn't mean making the most of the system still doesn't take a lot of skill and experience, but still more foolproof for some things. (for the skilled stuff, it definitely would take experience and hard work to get really good sounding samples working within the RAM and resource limitations of the stock A500 for games -things like picking and choosing what sample sneed to be longer and what can be short/lower quality,

 

 

 

 

And since we're just talking the ST's stock sound hardware specifically, there's only very limited and extremely specific advantages over the Amiga: the Amiga can't fully match the YM2149, but it can come very close while the YM2149 can't come close at all unless hacked with a hefty amount of CPU resource to use it as a DAC. (or 3 poorer sounding hardware channels to avoid scaling/mixing)

Still, the ST's sound was pretty much the best single chip off the shelf sound hardware available in quantity in 1984, even if you didn't restrict cost -Ymaha's FM chips were not available in quantity until '85 -everything better was custom and not offered to 3rd party buyers . . . then again it was the better of 2 mass market options with the only other being the weak SN76489 -weaker noise generation, volume control -crappier PCM hacking, no ADSR, lower frequency resolution, etc -and other than being in-stock already, POKEY probably wouldn't have been a cost-friendly option unless they could have stripped it down to a cheap form factor later on, but that wouldn't have been a design consideration untiil after Atari Corp was formed -ie earlier RBP work might have already gotten pretty far on paper by mid '84; you could argue a very basic custom chip with a small array of direct DAC channels and logic to interface to the CPU could have been a good idea on top of a basic PSG chip -without DMA you'd be limited to CPU heavy PCM options, but at least with a few hardware mixed direct DAC channels it would be better sounding and less intensive than PSG hacking -and even more practical to use in-game to a limited extent while still making full use of the PSG -and could have matched the amiga much better for dedicated sound/music programs due to varying sample rate/loop times on the hardware channels rather than scaling and mixing, and given the ST managed better than 16 kHz 4 channel MOD on the YM there's a real chance that you could maeet of exceed 28 kHz 4 channel mod with 4 hardware DACs; DMA would have been great, but a bare DAC hack could have been a stop-gap solution due to time constraints, and aside from PCM, you could also do simple PWM based sound stuff like the A8 does with POKEY using the ST's interval timers -except that's still a fair bit of CPU overhead and proportionally more so than the A8 due to the slower 68k interrupts, but probably a bit less than PCM at least)

The ST could have used something like the Lynx's sound chip.

 

 

 

 

 

Also, you can indeed have better samples on the ST than the Amiga, but only if all else isn't equal (ie better sample conversion/formatting, larger/onger samples in more RAM, etc).

And there are cases of the Amiga MOD/tunes are actually weaker than the ST AY chiptunes too. (cases where the ST -and sometimes Speccy/MSX/CPC- chiptunes sound better than the C64 for that matter)

 

 

 

 

Did the ST have support for Covox DAC type devices on the parallel port? (not a big jump over using the AY alone, but better nonetheless) Given the sound/music emphasis, popularity in Europe, and extreme low cost/simplicty of the Covox/DIY soundboards (and the fact it appeared on PCs in '86 commercially -and a ton of DIY versions following that), it would be a bit odd that the ST DIDN'T get support for such. (or use the cartridge slot for such, but I think that may be read-only and not really usable for such unless you did a hack to have certain software use one address/select/etc signal on the connector for something else like some VCS games did -or all actually as the 13th address line had to be hacked as chip select; but for that matter, they could have done that for other PCM/synth chip upgrades and such too -the only snag would be mixing that with the ST monitor output, but a simple piggyback plug/camble arrangement might have worked -like Sega did for the FM add-on of the Mk.III)

Edited January 28, 2011 by kool kitty89

[oky2000]

I wasn't downplaying either really. My point was simply that unlike any other micro before it the Amiga had no soundchip as such in the traditional sense that Pokey sounds like Pokey and YM/AY like YM/AY and SID like SID and TI like TI etc etc. Amiga will sound like any sample you feed it but didn't have an 'Amiga sound' unlike other machines, Given the massive frequency range and variable playback speeds it was a cunning use of the DAC technology.

 

My point about 'just 4 DACs' is simply to illustrate it [Amiga] is a blank canvas as far as musicians are concerned limited only by channels (up to 8 in some trackers) and what samples they could feed it. Sadly there is no fancy pants ADSR or sync/ring mod effects so it is quite basic compared to a synth sample keyboard still but as an idea it is genius.

 

Most Amiga tunes sound a bit cheesy now but then that's because the 80s/early 90s WERE cheesy and that's where they got those tacky instruments from....rubbish 80s/90s music

 

Making a DAC sound like a discreet IC that forms a soundchip is very difficult @ 7 or 8mhz. Of note is Per Hakan Sundel's awesome awesome C64 demo, it is the best SID emulation any 7mhz computer in the world can produce.....just happens to be on Amiga. Generally though this will sound terrible (There is no AY/YM emulator I know of for Amiga) On a plain ST samples (ie a sample file on a disk) WILL sound weaker on the ST than Amiga. But then the ST is using the same sort of tricks as 8bit micros used to play samples not a dedicated 8bit DAC (STE has a DAC setup though obviously). But everything in the last two paragraphs is relevant really. If you don't like AY/YM sounds you won't like ST game music ditto if you don't like the Pokey/SID sound you won't like a lot of A8/C64 sfx/music either.

 

Do what I do, buy them all and pick and mix

Edited January 28, 2011 by oky2000

[kool kitty89]

I wasn't downplaying either really. My point was simply that unlike any other micro before it the Amiga had no soundchip as such in the traditional sense that Pokey sounds like Pokey and YM/AY like YM/AY and SID like SID and TI like TI etc etc. Amiga will sound like any sample you feed it but didn't have an 'Amiga sound' unlike other machines, Given the massive frequency range and variable playback speeds it was a cunning use of the DAC technology.{/quote]

DAC technology??? ALL digital audio devices use DACs by definition (at some point it has to be come analog)... Smart use of DMA logic perhaps and forward thinking to design hardware PCM playback logic (and have very flexible DMA sample rates), but not "smart use of DACs." (the Atari 8-bit had smart use of DACs )

 

Hell, they could have removed the DACs from paula and it would still have been quite impressive/innovative as such. (the Mac had done DMA sound first, but not as flexibly, and arcades had done CPU/MCU driven PCM long before that, but that's also different -not sure what the actual first implementation of DMA loaded PCM playback was though -I think arcades were using DMA PCM chips by 1985)

Yeah, take PAULA and remove the internal DACs and have it hardware mix digitally to an external DAC (be it parallel -as AMY required, or serial -like Yamaha usually used) and it still wouldn't have been any less impressive. (other than the weaker integration)

 

 

 

And I'd argue that the Amiga DOES had a fairly distinctive sound to it due to a number of things from the memory limitations forcing samples of a certain quality, the hardwired stereo, the types of filtering used, the pitch control mechanism used by most hardware synth (ie the playback rate changes and thus sounds a bit different than hardware/software PCM mechanism using scaling -let alone interpolation), and then the general style of things done with it.

Style of the music plays a MASSIVE role in it too, more in some respects than the feature set or hardware limitations: platforms with very heavy European support have a very distinctive sound to much of the chiptune stuff done (heavy use -some would say overuse- of fast arpeggio to simulate chords, etc), game consoles dominated by Japanese developers/composers, platforms more tied to the US, etc.

If you start adding in more CPU intensive mixing stuff, yes then it becomes more ambiguous (or emulators with filtering disabled), but that's true for others as well. (like MOD on the ST or other systems hacking it through a PSG or bare DAC)

 

If there had been contemporary machines with similarly specific hardware limits for audio and popularity in a similar region then things might have blurred more, but that didn't happen. (and it's also the reason that several other platforms have such distinctive sound -hell, if US-type DOS PCs had been big in Europe and Japan at the beginning of the 90s, Adlib/Sound Blaster probably would have been pushed a good bit further than it tended to be in the US -then again, there were several US companies that also did much better work on other platforms than SB/Adlib sound -and pretty much none made use of the PCM channel for music along with FM synth -some later games sued MOD players, but that wasted the synth hardware; the OPL3 also got largely unused -Japan's PC-like NEC PC9800 series had a higher end sound system more like the higher end PC8801s -YM2608, and by the time PCs got big in Japan things had moved on to CD audio and general midi -among some other sample synth formats)

 

 

 

You also seem to ignore that while sound chips don't have the PCM/sample replication route, many can be made to sound exactly (or almost exactly) like other sound chips, especially weaker ones: the AY/YM chip can do pretty much everything the SN76489 can do but also much more. (POKEY can also pretty much replicate the SN and do some of the things the AY can but not others -with trade-offs for other things the AY can't do in hardware)

You have cases where POKEY stuff doesn't take advantage of periodic noise and may be approximated very closely by an SN or AY PSG. And the SID which can do most/all of that but with a 3 channel limit. (pulse wave generation can be somewhat used in place of periodic noise as well -especially tweaked a bit by rapidly changing the pulse-width used in a random/pseudo-random fashion -also possible to simulate with a plain 1-bit/square wave channel and CPU resource modulating it -hence Pitfall II on the PCJr and CV both have the periodic noise-like jumping sound of the VCS game -C64 uses pulse too, POKEY oddly opts for Square in the A8/5200 game)

 

The PC Engine's sound chip can replicate almost all of the previous ones, except some of the SID's effects (namely the filter), and with added capabilities of its own, and then you've got various FM synth chips that can do pretty much everything all major PSGs did and a lot more (depending somewhat on the chip in question -the YM2413 was limited to 15 preset instrument options and only 1 user programmable instrument to apply to the 9 hardware channels -any 1 instrument could be used any number of times on those channels; but for other cases, even the simple 2-op FM sine wave OPL, you could match pretty much everythign but filter specific effects on the SID -and even then you could get around some of that).

 

 

 

But going back to the "characteristic sound" thing: NES games are often defined by the distinctive Japanese developer sounds with some US devs that were forgettable, tended to mimic the JP stuff somewhat, or a few with distinctive sound, and then more so for EU developers. (David Crane's A Boy and His Blob doesn't really sound like anything else on the NES style wise -for both sound and music it's actually more like some US 8-bit computer stuff or Atari Console stuff even) Codemasters NES games often have a very C64 like sound to them, not full-on SID sound, but more like SID music that stays away from the filters and focuses mainly on pulse/square/triangle wave stuff. (the DPCM channel can be hacked to do saw to I believe)

 

Seriously, doesn't this have a very C64-like quality to it. (the fact that there's a fair amount of SID stuff that stays away from filtered/ring stuff pushes the similarity all the more -the variable waveforms of the SID and the ADSR tend to make it more distinctive than the filter/ring stuff -the full-on analog synth stuff is distinctive too, but the unfiltered stuff is also a good step of what went before it -even though PWM and software ADSR should have been very possibly on a competitive level with some previous systems -including the A8- but the hardware features make that a bit more foolproof -a shame the AY8930 wasn't out by '84/85 given that should have been a simple alternative -the YM2203 would be much better but should have been more expensive while the 8930 was a low cost enhancement -taking the AY/YM and adding variable pulse width generation and 3 ADSR envelopes rather than 1 -all of which could be hacked into low-res sawtooth/reverse saw or triangle channels, I'm not sure when that became available but it doesn't seem like until the late 80s -way too late and the YM2203 would have been a better option by then anyway to replace the YM2149 with -a shame the STe didn't add that along with the simple DMA PCM hardware)

http://www.youtube.com/watch?v=X8ruaLiGfkA

http://www.youtube.com/watch?v=oiTSG6u0fmE

http://www.youtube.com/watch?v=v6-WeaJNqmY

 

 

 

 

Software ADSR is pretty CPU light and common on a number of platforms, and codemasters seem to push that for their NES music. (on sample based systems like the amiga, it's usually better to use separate samples employed for ADSR or at least an attack and sustain sample -the MT-32 used a very interesting hybrid of samples for attack and simple waveforms added/subtracted/modulated for the sustain as well as other effects, but that's well beyond plain sample synth where you'd use segments of sample played in order for attack and possibly decay -you coudl just use volume control for decay, and also volume for attack to save space for an actual attack sample)

 

 

 

Most Amiga tunes sound a bit cheesy now but then that's because the 80s/early 90s WERE cheesy and that's where they got those tacky instruments from....rubbish 80s/90s music

Then there's all that stuff using samples (or AHX) to try to sound like SID/analog synth stuff. (hardly a bad thing though -Raiden on the Jaguar did that too)

And the inherent memory limitations.

And there's TONs of modern stuff that goes beyond that due to the chiptune/demoscene. (many conforming to the 8/24/48k MOD formats -it's amazing what some manage to cram into 8k too)

http://battleofthebits.org/arena/Entry/Orbital+debris/3625/ http://battleofthebits.org/arena/Entry/greyness/2332/

 

 

There was no other platform pushing MOD-type stuff at the time, so it IS distinctive as such, and the stuff that sounds closest later on is often due to conforming to the Amiga's hardware limitations of 4 channel hard unpanned stereo.

 

Even the examples that went beyond that also have distinctive sounds, often also due to limitations: examples that use high-res long samples (like Shadow of the Beast and Wrath of the Demon -which seems to rip-off SoTB's sample set) sounds extremely limited and repetitive.

Sure the high res pan pipes and a few other instruments at 22 kHz are an interestign change of pace, but the fact that thy overuse the same small set of sounds makes it a major trade-off against some of the better examples that used low quality samples (but a broader set of instruments) and contemporary platforms with hardware synth pushing much more varied sound (though plenty of cases of poor/weaker use -and then cases within the weaker examples were it's up to personal preference; my little brother has an odd preference for even some of the poorer Adlib versions of games with rather well-done amiga soundtracks including wrath of the demon -and many of those are rather poorly optimized for Adlib and not using the hardware effectively, especially sicne they had 9 channels to pare down to 4 -additive synth- to compare with the arrangements or add complexity with more channels).

 

SOTB at least mixes it up near the end of the game with a pretty substantial shift from the pipes, so they DIDN'T stick to one hard set of instruments, but it's a real shame they decided to push ahead with that limited instrument set for most of the game. (the C64 version is a good bit more varied and even the Genesis version with its rather weak use of the EA sound engine was more varied at times -the japanese version seems to have tweaked it slightly but not much -added more PCM and changed the graphics a bit for better color optimization and soem sprites redrawn with the added detail of the PCE and FM Towns ports)

 

Again, plenty of other examples of "samey" sounding stuff (adlib in particular and some 8-bit platforms -MegaDrive to some extent) in other examples too, but that doesn't mean it gets a pass in this example. (it's up to taste too, but it would have made a ton more sense to vary the instruments depending on the environment -plenty of examples of that, but Donkey Kong Country jumps to mind due to being UK developed and with significant pipe usage at times -but none at all where inappropriate)

 

 

Not until streaming recorded audio did you have a truly blank canvas: though in that sense you could argue cassette soundtracks included with some games did that well before CD-ROM. (a damn shame the A8 didn't catch on in Europe, there could have been some really neat use of the computer controlled tape drive with the separate sound channel -the A8 uses 1 stereo channel for data and the other for analog audio; sure, seek/rewind times could be a bit long, but that would have still been a nice potential option for in-game music synced to the game by the computer controlling the tape)

 

 

 

 

Making a DAC sound like a discreet IC that forms a soundchip is very difficult @ 7 or 8mhz.

Not really, just use samples. And technically, you could have the CPU "build" some waveform samples ahead of time and not have to load them from disk/

I'm not sure how AHX works, but some descriptions claim no samples.

Anyway, samples definitely would eb the primary route to emulate such chip sounds with DMA audio. (with a CPU and a parallel interface to a DAC with no added logic -totally different circumstances than the Mac or Amiga- you could do on the fly waveform generation as well for simple square and pulse wave stuff and with a bit more work you could probably manage some sawtooth and triangle stuff depending on the CPU resource thrown out there -pulse and square would be simple though, and software driven pulse and square waves are what the Spectrum and Apple II use as well as some PC speaker stuff -sometimes mixing multiple channels into PWM, and the CoCo can do that or use the full 6-bit DAC with parallel connectivity to the CPU -vs 1-bit toggle/beeper- for PCM playback)

A bare CPU/MCU+DAC (with parallel interface) is all several arcade games used for audio (Donkey Kong among many others), though those usually used low quality PCM rather than modulated waveforms. (and CoCo, Covox DAC, what the Genesis uses via YM2612 -Afterburner plays through the PSG rather than the 8-bit DAC port though)

 

But stop saying DAC if you mean PCM playback system (totally different context than just a DAC -which is a component of any digital sound device with built=-in analog output -again, PAULA could have easily been made all digital with external digital output to tie to a DAC like Amy or most of Yamaha's FM chips ).

 

 

4 hardware PCM channels is significant, 4 DACs means almost nothing. (4 DACs with parallel interfaces to the CPU means a little more, but still not much compared to hardware PCM with or without internal DACs )

Edited January 28, 2011 by kool kitty89

[+rdemming]

Did the ST have support for Covox DAC type devices on the parallel port?

 

Yes, some mod trackers supported a DAC on the printer port. Also some games like Wings of Death and probably Lethal Xcess supported a DAC on the printer port.

I made one myself using a resistor ladder DAC.

 

(or use the cartridge slot for such, but I think that may be read-only and not really usable for such unless you did a hack

 

That was possible too. All digital sample cartridges (to make sound samples) also supported playing digital samples through the cartridge port.

The cart port indeed does not have a write signal. But they got around this by reading specific addresses in the cart. I'm not familiar with the details but it could work by setting aside a block of cartridge memory space. So reading from address $xxxxxx00 will output level 0 and reading from $xxxxxxFF will output level 255.

 

Also those sampler cartridges were supported by many sound trackers and I believe also by some games like Wings of Death.

 

 

My point about 'just 4 DACs' is simply to illustrate it [Amiga] is a blank canvas as far as musicians are concerned limited only by channels (up to 8 in some trackers)

 

8 channel trackers on the Amiga used software mixing, just like the ST(E) does it with playing mod files. So more than 4 channel music on an Amiga eats a lot of CPU time as well. Thats why 8 channel music was only used on title screens etc. and not during game play.

 

 

Making a DAC sound like a discreet IC that forms a soundchip is very difficult @ 7 or 8mhz. Of note is Per Hakan Sundel's awesome awesome C64 demo, it is the best SID emulation any 7mhz computer in the world can produce.....just happens to be on Amiga.

 

The Nostalgic-O-Demo has a demo screen with a SID emulator that works on a regular ST. The sound quality is not so high but it works.

 

 

B.T.W. There were at least two sound expansions for the Atari ST.

First there was the game B.A.T. that included a cartridge for sound output. I don't know if it was just a DAC or contained other hardware as well. Maybe someone can tell more about that.

 

And I remember that in the magazine ST-Format, there was a review of a sound cartridge that had an additional soundchip. Beside the software that came with it, it was afaik never supported.

 

Robert

[oky2000]

No Amiga IS unique in having pure DAC and zero waveform generation as sound hardware @ kitty. Also the filter was aggressive and only had 1bit on/off state unlike SID. The only other control was channel volume and sample playback speed. Bob Yannes is the king of 8bit sample based technology via Ensoniq Mirage synths that sounded 16bit and used by Inner City for chart topping 90s hits

 

Each to their own though.

[atarian63]

I think in the days when the Amiga and ST were selling side by side it was obvious which one had the better sound and graphics capabilities (apart from the early titles that were just quickly ported onto the Amiga from the ST version) and the Amiga , in my opinion, would win pretty much hands down every time.

 

Having said that, there were and still are good and bad examples of music/sound on both both platforms and not having had any Amiga for years I find I now appreciate the ST a lot more than I did back in the day.

When we were retailing them side by side we were selective and used better demos on the ST/STE so the match appeared much closer. In a commercial sense St evened things up a bit with samples in early software. This method seemed to become more common later in the ST's life. StarGlider comes to mind. Sounded better on the ST

[oky2000]

I think in the days when the Amiga and ST were selling side by side it was obvious which one had the better sound and graphics capabilities (apart from the early titles that were just quickly ported onto the Amiga from the ST version) and the Amiga , in my opinion, would win pretty much hands down every time.

 

Having said that, there were and still are good and bad examples of music/sound on both both platforms and not having had any Amiga for years I find I now appreciate the ST a lot more than I did back in the day.

When we were retailing them side by side we were selective and used better demos on the ST/STE so the match appeared much closer. In a commercial sense St evened things up a bit with samples in early software. This method seemed to become more common later in the ST's life. StarGlider comes to mind. Sounded better on the ST

 

Early ST software seemed to really push samples more despite the poor 68k being overworked already, we the consumers expected it I guess. The with quantity increase came quality decrease generally speaking, ditto for Amiga but longer cycle. Look at Gauntlet 1 & then 2 for same apathy towards graphics capabilities.

[oky2000]

One thing that really annoyed me, and nobody has mention which is understandable as this isn't Lemon64, is around the time the ST and Amiga were being dropped to £400-£500 price Hubbard and Galway left the scene!

 

For me this was a disaster as even on Amiga let alone ST all people were doing was loading in a single 30 second sample! Talent was scarce and people left were very blinkered to the true possibilities of new sounds/music types. Even when they tried it they messed it up (took me just 1 day to do a superior and accurate rendition of Megablast as on Xenon II

 

Damn it!

[yerzmyey]

...and we still like ST sound and we don't complain.

http://ym-digital.i-demo.pl/

[kool kitty89]

(or use the cartridge slot for such, but I think that may be read-only and not really usable for such unless you did a hack

 

That was possible too. All digital sample cartridges (to make sound samples) also supported playing digital samples through the cartridge port.

The cart port indeed does not have a write signal. But they got around this by reading specific addresses in the cart. I'm not familiar with the details but it could work by setting aside a block of cartridge memory space. So reading from address $xxxxxx00 will output level 0 and reading from $xxxxxxFF will output level 255.

 

Also those sampler cartridges were supported by many sound trackers and I believe also by some games like Wings of Death.

Yes, hacking it like the VCS, except with fewer problems/headaches in several areas. The R/W line issue is trivial, but for the VCS you also had the lack of Phi-2 that gave the major headache. (something nonexistent with the 68k)

--------------------

64k and larger games with oodles of RAM are not that hard to make inside the cartridge, from a hardware standpoint.

 

If you want to talk historically, it was a big leap at the time. I don't think anyone really knew you could manage to get RAM in a cart without the read/write lines back in '77. And based on what Larry Wagner said, I think the general thinking at the time was that banking hardware was going to be too tricky to implement until he came up with a way to do it with a "couple of ICs" as he said. So what seems obvious to us now required some technical breakthroughs.

Bankswitching hardware is sort of trivial. I wonder if any engineer really had trouble with it.

 

It's not just the R/W line but the lack of phi2 that makes adding RAM to a lot harder than just switching banks. I'm not sure what cwlikson meant by "not that hard" but it is a lot harder than bankswitching. Certainly it's not an amazingly difficult problem but it is not trivial.

 

The lack of R/W line is easily worked around by having separate address for reads and writes. But, no phi2 signal means that you have to use accurately timed circuitry to simulate phi2 to hit the small window where data is valid for writing between cycles. Worse, there are actually two windows you need to hit as some instructions write on the first cycle and some write on the second.

 

Alternatively you can latch the address bus on writes and wait for the address to change, then grab the data and use the latched address to know where to write it. This method requires much more hardware than regular bankswitching.

 

Also, neither of the above RAM methods can be gleaned from simply examining 6507 datasheets. Atari had to do experiments to figure out the timing.

 

The document here shows some of the research done. Although this was done for the 2600 computer add-on, as I understand the research ultimately resulted in the SARA chip:

http://www.atarimuseum.com/archives/pdf/vi...esign_notes.pdf

----------------

 

And again, the cart slot would be an interesting option for more than sample based stuff, but a general consumer level (lower cost) sound/music add-on as well. DMA sound would have been nice, but I'm not sure how realistic that would be since you'd probably need external sample RAM. (maybe it could just add a small buffer and intend the 68k to periodically fill it, sort of like the circular buffer for MAC sound) An FM Synth chip would have been another option with the low-end/cheap option being the YM2413 (MSX add-on, Japanese Master System/Mk.III, etc) but there would have been much better options aside from that. (the YM2612 was the only other model that didn't require an external -serial- DAC, stereo at that, and embedded a direct write 8-bit DAC channel -if FM channel 5 was disabled- so even lacking DMA, you'd have a COVOX-like option built-in without any added board space -and full parallel use of the YM2149) For any such add-on, youd need an external cable to plug into the AV port and mix with the PSG output. (if stereo, probably offer both passthrough mono on a piggyback AV socket and stereo RCA ports or a smaller/cheaper 1/8" headphone jack/line out -you wouldn't be able to mix it into RF though, so if using a TV -as a fair chunk of the lower-end EU userbase probably was, you'd need separate speakers, or heaphones if the module added a built-in volume slider/dial)

 

Atari certainly hadn't been pushing for add-ons, but that doesn't mean a 3rd party couldn't, and the YM2612 would have been cheap enough to fit into the consumer role while also being attractive to the more professional side of things given it was full 4-op Yamaha FM synth (like the YM2151, but only 6 channels -or 5 in DAC mode) on top of built-in covox like functionality. (the DAC mode was limited to ~27 kHz before missed writes, so interleaved/multiplexed mixing would be limited as such -and if you wanted over 13.5 kHz, you couldn't even use it for 2 channels unless you interleaved at different rate -ie one 18 kHz channel and 1 9 kHz) It also adds 2 high res interval timers, but given the lack of an IRQ line on the cart slot, that would be limited to use via polling. (without any added hacking to the cart slot)

 

8 channel trackers on the Amiga used software mixing, just like the ST(E) does it with playing mod files. So more than 4 channel music on an Amiga eats a lot of CPU time as well. Thats why 8 channel music was only used on title screens etc. and not during game play.

It's still DMA though, so eating far less time than software PCM playback on the ST. (even with DMA, it's not just mixing either, but having to scale the samples for pitch bending would be more of an issue -you could even specially format the samples to not exceed 7 bit resolution and thus totally avoid dealing with overflow, or you could limit things to ~14 kHz with interleaved/multiplex mixing but still having to scale for pitch, or you could do a hybrid set-up with some multiplexed/interleaved channels at fixed -or integer divided- sample rates and others using playback rate for pitch -ie 2 "full" channels and 2 more with interleave mixing for fixed rate FSX/drums/etc with a variable number of channels depending on the output: ie 2 ~14 kHz channels per hardware channel, or 3 ~9 kHz, etc, and you could have stereo panning effects too if the 2 hardware channels selected were L and R channels, so possibly 2 "full" channels and 6 ~9 kHz percussion/SFX channels with stereo, or any other of a variety of configurations -and aside from avoiding adding, interleave mixing means all channels have the full 8-bit resolution rather than losing precision from mixing to 8-bit)

 

STe or MAC (or Sound Blaster) type DMA audio was much better than what the ST had to work with, or far better than a plain 8-bit DAC with parallel I/O interface. (support for a pretty high playback rate -like the STe or Tandy DAC with 50/48 kHz- would also facilitate multiplex/interleave mixing as well -hence the 25 kHz 4 channel interleave/multiplex configuration with close approximation of the Amiga's output but requiring software pitch bending/scaling for additional notes beyond the RAM allotted to samples, or doing 6 channels -3L/3R- at ~16.7 kHz)

 

Hell, you *could* do a 3 channel sample player on the stock ST with pitch control without any scaling (or 8-bit to PSG look-up), but 1: you'd still have all the overhead from "brute force" CPU sample playback (CPU sets every single sample), and 2: it would be limited to logaritmic/PSG formatted 4-bit (or 5-bit) samples. (pitch would be controlled via playback rate -which also means you'd have more CPU time at lower pitches- and would be generally weaker than what POKEY allowed with pure linear 4-bit PCM and 4 hardware channels -albeit with much more CPU grunt than the A8) Few players seem to use that method, but it's what the 4 channel sample system soundtracker on the Spectrum uses. (A few games used single YM channels for drums/FX simultaneous with music -Turrican II does for drums)

 

And, of course, the simplest/cheapest hack to at least allow better software driven MOD/PCM would be a small array of DACs and a parallel interface (or a simple IC that hardware mixed several 8-bit channels -accessed via simple write ports- to a higher res output). Still CPU intensive (setting every sample) but you'd avoid the overhead of mixing and scaling for pitch, like the 4x COVOX on the A8 allows. (no need for any buffer RAM or DMA logic, just a simple ASIC or some discrete logic and an array of DACs -or even external resistor ladders on a PCB along with discrete interface logic to attach to the cart slot; the latter could have been a DIY hack)

 

 

 

 

 

 

No Amiga IS unique in having pure DAC and zero waveform generation as sound hardware @ kitty. Also the filter was aggressive and only had 1bit on/off state unlike SID. The only other control was channel volume and sample playback speed.

Nope, that would be the CoCo (or various arcade games), not the Amiga. Paula is a pretty neat piece of hardware, and calling it "just some DACs" makes no sense. It was the DMA logic with signed PCM decoding that mattered, not the DACs themselves. (the logic feeding the DACs is what mattered, no logic and you've got a COVOX set-up or similar hacks via DACs or PSGs on the ST, CoCo, A8, C64, MSX, SMS, Mega Drive, PCE/TG-16 etc -the latter could do 3 10-bit linear PCM channels too, or 6 5-bit channels, but all direct write DACs -or an additional look-up for 10-bit to dual 5-bit- no DMA -though the Supergrafx added a feature to allow buffering into the 32 word wave RAM for PCM playback rather than just using that for waveform programming)

The fact that they also went above and beyond with 4 DMA channels (to the Mac's 2 -1 dedicated to the floppy drive square wave on early models) and also provided fairly high res frequency/playback rate control were also critical features on top of the PCM decoding DMA logic.

 

Remove the DACs from Paula, and it's not much less impressive (less integrated, but the same functionality -integrating the DACs is probably one of the least impressive features of Paula's design in that respect), but remove the DMA logic and you've got a glorified quad covox with hardware volume control and filtering.

 

 

If the ST had 4 8-bit DACs with a bit of logic to provide parallel write ports for each, it would still not even be remotely close to what Paula does in hardware, let alone what you can do in software on top of that. (ie with the same amount of resource it would take for the 68k to drive 4 direct-write DACs, the Amiga could software mix many more channels) And you'd never be using that in-game other than for very limited gameplay with mostly static screens. (thus, you'd still need the PSG unless you wanted no sound in games -or sound more like Double Dragon or such, or much choppier gameplay just to allow some very low quality MOD or only 1 or 2 channel MOD -maybe PWM mixed in, or maybe only PWM, but that's still pretty CPU intensive for the 68k, especially using interrupts -of course, adding hardware scrolling and/or using a faster CPU would free up a ton more CPU time to make bare DACs much more usable in-game to the point of removing the PSG being more realistic . . . actually, some of the interval timers in the ST might have been able to drive square waves on such DACs without CPU assistance, though it would depend how flexible the timers were -ie if they could do what the i8253 does for the PC)

Edited February 2, 2011 by kool kitty89

[+Psionic]

If strictly comparing music, I'd take the Amiga over the ST any day (as with the C-64 over the Atari 8-bits).

 

I'm sure the ST is capable of some decent music, but some games are just an abomination. The music in

is terrible compared to the Amiga version. Even if you compare it to the console ports, it's actually closer to the Master System version than it is to the Genesis/Mega Drive.

[emkay]

If strictly comparing music, I'd take the Amiga over the ST any day (as with the C-64 over the Atari 8-bits).

 

 

A very coarse comparision. Actually, making music with SID is an easy thing, compared to POKEY. But POKEY has it's advantages compared to the SID.

 

The YM is worse in all cases, comparing to PAULA, this is no question.

Less channels, less resolution, less frequency range, less waveforms.... whatever.

 

It was a big mistake, NOT to use POKEY in the ST series. 4 DAC channels and/or multi waveform modulations, up to 4 channels, driven at 8MHz .....

[yerzmyey]

If strictly comparing music, I'd take the Amiga over the ST any day (as with the C-64 over the Atari 8-bits).

 

 

The YM is worse in all cases, comparing to PAULA, this is no question.

 

I question this.

They're made for different purposes. Paula is for digital music mostly (by default: only) and YM/AY is for synthetic music mostly (again - by default: only). So in general, this comparison is as pointless as to compare: what is better - a sheaf-binder or a cucumber.

 

Not everybody like MODs. (I do but some people prefer synthetic music and they even hate digitals).

What annoys me is, that everybody knows that playing MODs takes ST's CPU power but not everybody knows that playin' synthetic music (real chiptunes, not any samples from C64) takes _a lot_ of Amiga's CPU power.

 

Amiga uses so-called "sample synthesis" based on a lot of veeeeery small samples. Hence it can generate various synthetic waveforms BUT it takes a lot of CPU power and actually MUCH more than MODs on ST.

Because I can still make and listen MODs on my 520STfm and I CAN'T make on my Amiga500 chiptunes with the best editor - AHX - because it works only on higher models (like A1200 and so on).

So THIS has to be funny for ST users, then.

 

PS: Of course I can make chiptunes on OctaMED on A500 but AGAIN - it has... NO VOLUMES in chiptune mode. (!!!)

[DracIsBack]

Having said that, there were and still are good and bad examples of music/sound on both both platforms and not having had any Amiga for years I find I now appreciate the ST a lot more than I did back in the day.

 

Back in the day, I had an IBM XT with bleepy PC sound and a CGA graphics card. At school, I had black and white Mac 512s.

 

Both the Amiga and ST blew me away so much with their graphics and sound in comparison at the time that I didn't really pay attention to how they compared except that they seemed more similar than what I was using.

[emkay]

If strictly comparing music, I'd take the Amiga over the ST any day (as with the C-64 over the Atari 8-bits).

 

 

The YM is worse in all cases, comparing to PAULA, this is no question.

 

I question this.

They're made for different purposes. Paula is for digital music mostly (by default: only) and YM/AY is for synthetic music mostly (again - by default: only). So in general, this comparison is as pointless as to compare: what is better - a sheaf-binder or a cucumber.

 

 

It's a clear point. If you say the YM sound is the wanted sound, well it's the best the YM can do, ofcourse.

 

Things get more interesting, when Music has to sound a bit more reliable.

 

 

 

 

 

Not everybody like MODs. (I do but some people prefer synthetic music and they even hate digitals).

What annoys me is, that everybody knows that playing MODs takes ST's CPU power but not everybody knows that playin' synthetic music (real chiptunes, not any samples from C64) takes _a lot_ of Amiga's CPU power.

 

 

There is no real need for making synthetic music. And it is even not impossible with PAULA. MOD File is just a definition, the result is that counts.

Chiptunes .... well .... interesting that they got their fans. Chiptunes are just some very rudimentary Synthesizer tunes. And real Synthesizers can sound marvellous. YM never sounded "marvellous" just interesting. AMIGA can sound "real" and "marvellous" without any sidenoise.

Amiga uses so-called "sample synthesis" based on a lot of veeeeery small samples. Hence it can generate various synthetic waveforms BUT it takes a lot of CPU power and actually MUCH more than MODs on ST.

Because I can still make and listen MODs on my 520STfm and I CAN'T make on my Amiga500 chiptunes with the best editor - AHX - because it works only on higher models (like A1200 and so on).

So THIS has to be funny for ST users, then.

 

MODs on the ST.... lol.... even the A8 can play them. But the replay quality is extremely reduced on a standard ST, even if using full CPU powers.

The YM is also not able to have "various" waveforms without heavy CPU handling. So where is the funny point?

 

 

PS: Of course I can make chiptunes on OctaMED on A500 but AGAIN - it has... NO VOLUMES in chiptune mode. (!!!)

 

Don't judge Hardware by the given Software. Because everyone on the AMIGA wanted to make MOD-like music, Synthesizing software wasn't needed to be created.

[sack-c0s]

 

A very coarse comparision. Actually, making music with SID is an easy thing, compared to POKEY. But POKEY has it's advantages compared to the SID.

 

The YM is worse in all cases, comparing to PAULA, this is no question.

Less channels, less resolution, less frequency range, less waveforms.... whatever.

 

 

I disagree that making SID music is easy - the SID gives you a lot of features, but it ends up being a lot of rope to hang yourself with - you need to use those features well. People also have more of an expectation regarding the end result. I find the pokey slightly less tricky because it has a few less variables to work with - but there are other tradeoffs (normally for me it ends up being 4th channel vs frequency stability).

 

The YM seems to be the easiest chip though because there's less to play with so people seem to accept the constraints, but to be honest although the pokey has quirks and flaws it has a distinctive sound and *personality* - I'd rather have that than something generic. It gets better with sync buzzer sounds but I need to get my head around those.

 

The NES soundchip I don't know what to make of that to be honest - the noise and PCM channels are okay for drums but I can't figure out what the hell they were thinking of with that pointlessly quiet sine channel - you end up with a YM with a channel missing. I would have preferred another square channel instead. At least the duty cycle settings make up for it a little.

[kool kitty89]

It was a big mistake, NOT to use POKEY in the ST series. 4 DAC channels and/or multi waveform modulations, up to 4 channels, driven at 8MHz .....

A lot of those tricks would be poor on the ST, eating up proportionally much more CPU time if using the same interrupt method and would be no better than the AY in many cases.

The tricks to push saw/triangle modulations might not work as such, but you can hack the envelope for that at no CPU cost. (and software envelopes are very CPU light by comparison -on the order of hundreds of interrups per second, if that, rather than thousands for modulations)

 

You have 3 DAC channels to work with on the ST as it is and several interval timers that can be used more or less like POKEY's (higher resolution actually), but the problem is the 68k would lose a fair amount of CPU time doing that and it would be a much bigger trade-off in-game than on the A8. (6502 has extremely fast interrupts -also makes software PCM much more favorable via interrupts)

 

So POKEY's volume modulation modes would only be moderately better than the AY's volume modulation: you get 4-bit linear vs 5-bit logarithmic (usually used as 4-bit logarithmic if single-channel), so you could get 4 rather than 3 channels, but that's also 1/3 more overhead.

POKEY also can't do the high-res DAC hack that the AY can via look-up tables: using all 4 channels at 5-bit resolution and optimized look-up, you can reasonably approximate 8-bit resolution output or a bit better than that even before the granularity gets too nasty. (ie to the point where mixing to 8-bit and dropping sample res is preferable to mixing to higher res -I think 10 bit might be reasonably possible)

 

Of course, the mixed hack is more CPU intensive by far than using separate hardware channels, thus you could manage much higher sample rates doing 3-channel volume modulation PCM than true MOD (or do a lot more other stuff while playing the digital music), similar for 4 4-bit channels as well. (of course, 4-bit samples also take up 1/2 the space of 8-bit ones )

It's not the look-up method that's the real kicker for overhead though, it's the fact that you have to software mix all the channels at a common playback rate and thus resample (scale) samples on the fly for notes. (thus if you pushed a single 8-bit MOD channel on the ST, it would be much less than 1/4 the resource of 4 channels, though a single plain 4-bit channel would be even less still due to lack of table look-up overhead -that's also why streaming PCM on the ST can be played at much higher rates than MOD, though most streaming stuff is pretty low rate due to the size it takes up anyway -might be better off left as 4-bit samples for that matter and double the sample rate)

 

 

 

 

Just like if the ST had 4 8-bit DACs in it. Sure, you'd have better quality software MOD, but you wouldn't have it in-game. (the Lynx's sound chip would have been nice thoughm option for 8-bit DAC/volume modulation on 4 channels plus fairly flaxible chip synth capabilities when CPU resource is scarce)

 

DMA audio and CPU brute-force DAC audio are very different things.

 

 

 

 

If strictly comparing music, I'd take the Amiga over the ST any day (as with the C-64 over the Atari 8-bits).

 

 

The YM is worse in all cases, comparing to PAULA, this is no question.

I question this.

They're made for different purposes. Paula is for digital music mostly (by default: only) and YM/AY is for synthetic music mostly (again - by default: only). So in general, this comparison is as pointless as to compare: what is better - a sheaf-binder or a cucumber.

 

Paula is a different animal entirely really, and it would have been extremely limited in an older machine with limited RAM. (granted there's some pretty impressive 8k MOD out there ) That's one of the reasons you didn't see more samples on 8-bits, not just the CPU resource, but the already constrained memory. (and for demos in titles and such, CPU resource is a non-issue, hence why you saw MODs in a fair amount of ST demo screens but not MOD/sample stuff on 8-bits -other than the occasional low quality fixed pitch percussion/chime/etc sample -like in stormlord on the C64)

 

The AY should be very capable of software modulation for different waveforms (namely pulse), and of course it has the envelope to hack saw and triangle waves at low res.

Even the rather crappy SN76489 would be just as capable at doing modulations like this:

http://battleofthebits.org/arena/Entry/down%20and%20dirty/3657/

 

Though if you want non CPU heavy chip synth that still sounds awesome, there's lots of other options:

http://battleofthebits.org/arena/Entry/Polaracer/3617/

(software envelopes are very CPU light even via interrupts since they're so infrequent compared to modulations/PCM -probably would have seen them a lot on the MSX, Speccy and CPC if they had useful timers available for such -a lot of NES games used them and better ST tunes did for sure . . . or maybe they do have programmable timer interrupts and I'm mistaken -the other option is software timed loops, easier in demos but tough to manage in game from what I understand -the envelope saw/triangle hack would be just as realistic to push on the 8-bits though -or the Intellivision for that matter )

 

It's got some notable advantages over POKEY and even some over SID (stereo, 3 channel plus simultaneous noise, 12-bit frequency resolution over POKEY's 8-bit normally, envelope POKEY lacks -without CPU help, etc).

 

And modulations are nothing specific to POKEY, POKEY's built-in timers makes it possible on the A8 where it otherwise wouldn't be without cycle timed loops (no other timers), but PWM can be done on anything including a beeper. (in fact that's pretty much all you can practically do with a beeper beyond square waves)

The fact that 6502 systems are extremely interrupt friendly is probably why you see the modulation stuff more there (or would have if the A8 had gotten more support back then ), but the ST has plenty of interval timers to push . . . but the overhead is much greater by comparison. (ie you'll eat up a lot of CPU time doing interrupt PWM on the ST, the gap from PWM to software PCM with the 68k is a lot smaller than the 6502 -though the 6502 is also much better suited to interrupt driven PCM as well)

Same problem with Z80 based systems except I think most of those lacked interval timers anyway. (so you'd have to resort to Apple-II type tight coding instead, though THAT would indeed avoid the interrupt overhead problem and make it far more competitive with the 6502 -even the 6502 would be significantly more efficent with tight cycle counted loops, but the gain over interrupts isn't nearly as extreme as some other architectures)

 

 

Not everybody like MODs. (I do but some people prefer synthetic music and they even hate digitals).

What annoys me is, that everybody knows that playing MODs takes ST's CPU power but not everybody knows that playin' synthetic music (real chiptunes, not any samples from C64) takes _a lot_ of Amiga's CPU power.

You mean like true AHX using CPU generated waveforms? (or does AHX cheat with samples?)

 

It would take a lot less CPU resource on the ST to push modulation on the AY. (I wonder why games didn't opt for that fairly often 0you'd have to space some CPU resource, of course, but less than doing PCM in-game, which some did push -mmuch more often too)

Hell, doing Spectrum style AY digital music would be a lot less intensive than MOD too, more so if you used only 1 or 2 channels rather than all 3. (given the overhead from interrupts, single-channel PCM synth probably wouldn't be a massive jump over PWM stuff on the ST, or the same if you dropped the peak sample rate a bit -and with 1 PCM channel you'd have 2 channels+noise left, and that PCM channel could be switched between PCM and chip synth as needed )

 

I wonder if some games would have pushed in-game 3 channel digital synth (or at least 1 or 2 channel) if Atari had offered 12/16 MHz models from the start (or well before the STe at least). They'd probably need chip-only modes to make things run faster optionally. (especially for 8 MHz users)

 

That's something POKEY should definitely be better for though, plain 4-bit DACs rather than the logarithmic stuff, and up to 4 channels. (with really good formatting and preprocessing, the AY probably would be pretty good though, and the samples you use seem to match that -even with the coarse 2dB volume steps on the SN PSG, Sega managed to push out some surprisingly decent 3 kHz samples in Afterburner II on the MD -only game to do that on the MD AFIK, others used the 8-bit DAC mode on the YM2612)

 

Amiga uses so-called "sample synthesis" based on a lot of veeeeery small samples. Hence it can generate various synthetic waveforms BUT it takes a lot of CPU power and actually MUCH more than MODs on ST.

That sounds like wavetable synthesis, or maybe granular synthesis of some sort. (either adding together or sequentially building up waveforms -real wavetable synth mind you, not the misnomer applied to sample synth)

 

Or do you just mean a selection of short samples with extremely tight loop times (no adding or compiling), like the PC Engine/TG-16? That would do it too and is one of Paula's shortcomings vs actual sample synthesizers. (ie the SNES's sound system, Ricoh chip in the Sega CD/FM Towns/etc, etc)

 

That's also something that a 6502 would have been a lot better at. (especially if you're using timer interrupts to set loop times)

 

 

 

A shame Atari didn't use the YM2203 though, that would have been pretty awesome, even on the STe along with the DMA sound. (definitely more of an advantage over the Amiga, especially in the right hands) Fully backwards compatible with the YM2149/AY and similar footprint too, so could have made sense to introduce later on even. (even if the price wasn't an issue, it was new in '85 and may not have been available -NEC probably had top priority for the PC8801 series)

It would have been really interesting to see what the better EU (and probably a few US) composers/programmers came up with compared to what the Japanese managed on the 8801. (not to mention it would have given arcade-like sound hardware -3 channels of the arcade standard YM2151, though the YM2203 itself was occasionally used in the arcade too)

http://www.youtube.com/view_play_list?p=D43F807A6A30062E (a lot of those are the PCE CD tracks, look for the PC88 versions, like

)

 

Of course there's the Megadrive to give some indication of what they'd have managed, but that's got 3 more FM channels and weaker PSG. (or 2 more FM channels, weaker PSG, and 1 software driven 8-bit DAC channel)

Edited February 4, 2011 by kool kitty89

[kool kitty89]

^Disregard the link to the YS PC8801 music . . . that's a later game using the awesome YM2608, not the older YM2203. (3 more FM channels, 6 preprogrammed rhythm/percussion sample synth channels, 1 ADPCM channel and the same PSG -or "SSG" as Yamaha calls it)

[emkay]

Paula is a different animal entirely really, and it would have been extremely limited in an older machine with limited RAM. (granted there's some pretty impressive 8k MOD out there ) That's one of the reasons you didn't see more samples on 8-bits, not just the CPU resource, but the already constrained memory. (and for demos in titles and such, CPU resource is a non-issue, hence why you saw MODs in a fair amount of ST demo screens but not MOD/sample stuff on 8-bits -other than the occasional low quality fixed pitch percussion/chime/etc sample -like in stormlord on the C64)

 

 

 

Have you ever listened to the Softsynth tunes on the A8? The waveforms were calculated via CPU and played writing to all 4 "DAC"s. due to the real 4 channels, no interferences happen there.

 

 

The source of such tune is slightly higher than "normal" POKEY tunes. But the waves are "free" . No need for Megabytes of DMA. Just the CPU could do faster there.