English Amiga Board


Go Back   English Amiga Board > Main > Retrogaming General Discussion

 
 
Thread Tools
Old 10 January 2019, 10:26   #81
dmacon
Registered User

 
Join Date: Nov 2018
Location: Germany
Posts: 33
Quote:
Originally Posted by roondar View Post
B
Am I misreading what you mean?
I'm afraid yes.

I mean, I already wrote that software emulation models have become quite sophisticated.

So I am actually saying: the sound of the SID, or any other synthesis techniques, can be modeled quite well entirely in software, and played back via samples.

As much to the point that, I think even a good listener will be hard-pressed to hear the difference.

However (and maybe that's the confusing aspect), I also have in mind the historical context.

Meaning that, for me, it doesn't make sense to talk about individual audio synthesis chips, without considering the whole system (same reasoning against the "chip xx can play sounds on its own!" argument).

But, within the historical context of our beloved obsolete platforms, that would actually be the appropriate base for a SID<>PAULA comparison.

Because in the end, you can always talk about 14-bit sampling at 3.58 MHz, but you know very well that this method is completely unfeasible on a stock OCS Amiga with 7.09MHz 68000 and 512k Chip.

Anything else is leading to the banal observation that: digital audio streaming & software synthesis has overthrown all other digital HW synthesis techniques. We should all know that.

Quote:
Because to me it's very simple: SID/AY are synth chips that have had better and better sample playback added over the years and everyone agrees this is good and enriches these platforms.
Yes, but of course it has its drawbacks.

On both the 1040 ST and C64, you can't use the CPU much more for anything else.

On the Amiga 500, fixed sample playback is cheap, and it can do a decent approximation of the classic C64 80's style music.

So, as a system, obviously the Amiga 500 provides much more general possibilities. It can cover more musical styles. And combine it with superb graphics.

(Again - this should be obvious to anyone here)

And yet, for it's specific sonic character, the Amiga 500 can not replace the C64 in my opinion.

I even noticed this back as a kid in the 90's, owning an Amiga, but immensely liking the C64 sound.

And in many cases (such cases where the Amiga version of a game was a lame Atari ST conversion), C64 platform games even played better...

However, sound-wise, I do think there is nothing the Atari 1040 ST can offer which an Amiga 500 can not do with similar efforts.

No, I have no interest in ultrasonic square waves.

Quote:
Amiga with Paula has had both better and better sample playback added over the years and more and more innovative software synthesizer options.
Yes, and now, as we are grown-ups, we have digital multi-channel audio interfaces capable of 24-bit resolution at 192kHz.

Quote:
Yet somehow people in this thread seem to think that only one has had meaningful improvements. How is this a fair reading of what happened?
Simple. All improvements run still on a stock C64. You expand the capabilities by expanding the hardware. We already know where this is leading to...

Quote:
Because reading this thread it seems to me that the argument is that adding sample abilities to synth chips = and adding synth abilities to sample chips = . And that feels like a really strange point of view to me.
Nah, please read my posts more carefully.

Amiga Chip music is a different style than C64 chip music.

Notice how I say different, not better or worse.

Quote:
Note I'm not singling out you per se here, you're just the guy who responded to me and your reply was on point for what I mean here
Fair enough, but I think I already made my point clear that anything besides the Atari ST is awesome.

Quote:
True, but then again as I understand it, Paula's modulation technique is not available to the SID.
Yes, but what is your point?

I made exactly this point against Hewitson's argument that Paula is just a DAC.

Quote:
Furthermore, if you really want that sound, there is nothing stopping you from creating a sample that includes said timbre alterations and filters.
I think this is not a suitable solution in the historical context.

Remember, we only have 512K and 7.09MHz.

But again:

-> I like Amiga chip music.

Quote:
Nor are you stopped from creating a software solution (or tracker feature if you like) that allows these dynamic alterations if you can spare the resources to make it happen.
I don't need to create them. They already exist (of course not for my Amiga 500).

Whole commercial music albums are created just only with VSTs.

Quote:
So it's clearly not as simple as saying "Paula can only play back samples, the SID can alter them".
For the specific comparison C64 <> A500: the C64 has a definitive advantage in this field.

The multimode filter provides a big parameter space to shape the sound of the samples.
But again, using samples on the C64 obviously has other big disadvantages, which the Amiga 500 doesn't have.

Again, in the end, what counts is the end result. Do I LIKE what I hear (and hopefully see)?

And much often, I also marvel at the effort and ingeniouity put into a specific solution.

Quote:
I've heard some Amiga mods being played through the C64 SID using a similar player and it does sound nice. I'm not sure it sounds as nice as the real thing, but that could just be me.
It can't sound as nice, since the playback rate using this method is limited to 7.8kHz.

But what do you do then? Yes, put the low-frequency harmonic content on the sample channels using filtering, and generate the high frequency content with the SID voices.

This is how the author of the SID sound examples I posted (LMan) dealt with this issue.

Again: limitations drive creativity. They often lead to very system-specific solutions.

Quote:
On the point of the 44.1KHz playback via SID, surely you know that Paula can do better than one channel 44.1KHz@8 bits?
Sure. And SID can also do better than 44.1kHz with this method.

Do you hear an audible difference between 3.58 MHz and 985kHz sample rate?

Quote:
Not that this was necessarily useful in the real world, merely an observation. Just like how the 985KHz digital elements in the SID don't translate to a useful 985KHz output signal and thus it's a bit of a red herring to discuss them as such.
Nah, with the SID, it is different. I do require this precision, not for the sample playback, but to model the oscillator behaviour correctly, especially the modulation possibilities.

If I emulate the sid sound engine at 44.1kHz rate (as old emulators did), you end up with a horribly aliased/distorted mess when oscillator sync and/or combined waveforms are used.

This also goes for other things, like the pulse-width resolution (with 44.1 khz sample rate, I merely have only ~13 steps left at 3.4kHz oscillator frequency).

Therefore, you definitely want to go for resampling for SID emulation.

Quote:
It may not have been you, but that's not the vibe I got from this thread. It reads much more like "let's bash Paula because it doesn't do synth in hardware" and "let's all talk about how poor samples sound compared to synths".
As I said, I think this discussion makes only sense within the context of the whole system.

I don't need to bother with this ridiculous argument that PAULA can not do sound on it's own when I have a full 16-bit CPU and DMA system with 512k memory behind it.

I think that the Amiga was originally a perfectly balanced system, feature-wise.

Last edited by dmacon; 10 January 2019 at 11:16.
dmacon is offline  
Old 10 January 2019, 10:39   #82
Hewitson
Registered User
Hewitson's Avatar
 
Join Date: Feb 2007
Location: Melbourne, Australia
Age: 37
Posts: 3,406
Quote:
Originally Posted by roondar View Post
The same actually goes for almost all systems that use a synth-style sound chip (including Atari ST, Megadrive, etc).
Actually, the Megadrive is capable of playing back very high quality samples.
Hewitson is offline  
Old 10 January 2019, 12:56   #83
roondar
Registered User

 
Join Date: Jul 2015
Location: The Netherlands
Posts: 1,381
Quote:
Originally Posted by dmacon View Post
However (and maybe that's the confusing aspect), I also have in mind the historical context.

Meaning that, for me, it doesn't make sense to talk about individual audio synthesis chips, without considering the whole system (same reasoning against the "chip xx can play sounds on its own!" argument).

But, within the historical context of our beloved obsolete platforms, that would actually be the appropriate base for a SID<>PAULA comparison.
I'm more than willing to look at it from a historical context, but that does mean my points about Amiga sound being hard to copy on the SID at similar quality levels becomes much more relevant given how big even early modules became (sample wise).

Quote:
Because in the end, you can always talk about 14-bit sampling at 3.58 MHz, but you know very well that this method is completely unfeasible on a stock OCS Amiga with 7.09MHz 68000 and 512k Chip.
I never said it was feasible to do 3.54MHz playback on an A500, I was talking about theoretical abilities. I specifically stated it wasn't feasible, but that higher than standard DMA rates should be possible even on base hardware (and they are).

Also note that 14 bit playback routines have been made for the 7.09MHz Amiga's (admittedly they won't run at 44.1KHz on a 68000). You can find them on Aminet. So clearly even the A500 can play back 14 bit stuff using dedicated players.

And even if you don't want to use specific 14 bit playback software, you can always pre-make a few 14 bit samples (split over two channels as is required) to use with a standard tracker. This will work just fine to create 14 bit output out of an A500 and won't cost any extra CPU time over a standard MOD.

All of this is part of what I mean when I say that people claim incorrectly that no improvements where made to Paula output. Even base model Amiga's can benefit from these extra options.

We could make a little list of new audio stuff that has become available over the years for A500 level machines and it's clearly not going to be empty.

Quote:
And yet, for it's specific sonic character, the Amiga 500 can not replace the C64 in my opinion.
I'd agree. I love the C64 sound.

But this sonic character thing clearly also goes the other way around if we're just using base hardware. Simple to prove really, all I need is a module containing in excess of 64KB of sample data, or one that relies on stereo sound for some of it's effects (or use that modulation feature no one uses ).

This has been my point all along - the C64 can only produce at best a reasonable facsimile of Amiga mod music (assuming the module isn't very big). The reverse is clearly also true - but given a dedicated person carefully sampling and optimising a track specifically to get the same sound I'm still somewhat certain that the Amiga version of a complex SID tune will on average sound closer to the original than the reverse situation (i.e. a complex or big mod converted to the C64). Even on an A500.

As to why I feel this way, well I've been hearing some really high quality mods lately and have been surprised at just how close good mod creators can get to 'copying' a pre-existing tune to the mod format. MOD composing has progressed over the years it seems.

Quote:
Yes, and now, as we are grown-ups, we have digital multi-channel audio interfaces capable of 24-bit resolution at 192kHz.
You're incorrectly assuming all improvements I'm talking about only work on accelerated Amiga's or require extra memory. This is false.

Quote:
Simple. All improvements run still on a stock C64. You expand the capabilities by expanding the hardware. We already know where this is leading to...
I actually had a software package that supported CPU driven 56KHz stereo playback back for my unexpanded A500 in 1992 or 1993 (came with my sampler).

At any rate, many of the improvements made to Paula audio quality also run on a stock A500, including the 14 bit mode (albeit not at 44KHz) and several of the soft-synths. It's simply wrong to say I have to use additional hardware to expand abilities.

Not only that but it's not really true that the SID examples you give are base C64 only (IMHO): all the really cool 'late SID' sample stuff require either modded C64's with custom build digitizer hardware (using designs from around 2008) or PC's to get the source audio material into the C64. To me that's not just using base level hardware or staying in 'historical context', cool though it may be.

Quote:
Yes, but what is your point?

I made exactly this point against Hewitson's argument that Paula is just a DAC.
The point is that even the humble Amiga can do stuff with samples that goes beyond just playing them back and that this neat and sadly oft-unused feature can probably create unique sound expressions that would both be hard to copy on a SID/AY system and be well, uniquely Amiga like.

Quote:
I think this is not a suitable solution in the historical context.

Remember, we only have 512K and 7.09MHz.
This all depends on length of the effect. You can have reasonably long samples even on a base A500. Not minutes long, but several seconds is easily doable. Like I said above, I've been hearing some high quality mods lately and have been impressed by what a dedicated creator can do.

Quote:
But again:

-> I like Amiga chip music.
I'm not doubting that at all. Not even for a moment.

Quote:
I don't need to create them. They already exist (of course not for my Amiga 500).

Whole commercial music albums are created just only with VSTs.
I was of course referring to using something similar (real time or not) on the A500. There are synth based trackers and trackers that also support synth sounds available that run on A500's just fine. I see no real reason why a somewhat similar effect couldn't be done somehow (even if this would include all kinds of limits or require 'rendering' the output samples).

Quote:
For the specific comparison C64 <> A500: the C64 has a definitive advantage in this field.

The multimode filter provides a big parameter space to shape the sound of the samples.
But again, using samples on the C64 obviously has other big disadvantages, which the Amiga 500 doesn't have.

Again, in the end, what counts is the end result. Do I LIKE what I hear (and hopefully see)?

And much often, I also marvel at the effort and ingeniouity put into a specific solution.
I have the same.


I just also have it for Amiga based solutions and feel they're being underappreciated here. I think stuff such as the AHX tracker is really cool, as is PreTracker (I think that's its name anyway). I also feel 14 bit playback is an awesome thing (more so because it actually works on all Amiga's down to the A1000) and still think the various additional-channel-mixing routines out there are very interesting and ingenious stuff.

It's more me disliking how that is swept under the rug as irrelevant or the 'that only works on faster machine' myths in this thread that keeps me replying.

Quote:
It can't sound as nice, since the playback rate using this method is limited to 7.8kHz.

But what do you do then? Yes, put the low-frequency harmonic content on the sample channels using filtering, and generate the high frequency content with the SID voices.

This is how the author of the SID sound examples I posted (LMan) dealt with this issue.

Again: limitations drive creativity. They often lead to very system-specific solutions.
Amiga audio has plenty of limitations. It has plenty of system-specific solutions. Arguably it's heavy reliance on samples give it some limitations the C64 doesn't have to deal with (such as how creating dozens and dozens of SID 'instruments' will barely increase memory load, but doing the same is much harder on the Amiga due to memory pressure). I'd argue that the same driving force for creativity exists, just in a different way.

Quote:
Sure. And SID can also do better than 44.1kHz with this method.

Do you hear an audible difference between 3.58 MHz and 985kHz sample rate?
My point here should've been clear but apparently isn't. I'll say it differently: even a base A500 can play back better sampled audio than the C64 can, period. Both in form (mono vs stereo/1 channel/4 channels mixed vs 4 channels unmixed/8 channels mixed), frequency (this one is more theoretical as the human ear/output quality of the machines will start getting in the way but it's still true), bitrate (8 bits max vs 8 or 14) and in length.

This is not to say that SID playback of samples is not impressive. But it's not at the same level as the Amiga.

Quote:
Nah, with the SID, it is different. I do require this precision, not for the sample playback, but to model the oscillator behaviour correctly, especially the modulation possibilities.

If I emulate the sid sound engine at 44.1kHz rate (as old emulators did), you end up with a horribly aliased/distorted mess when oscillator sync and/or combined waveforms are used.

This also goes for other things, like the pulse-width resolution (with 44.1 khz sample rate, I merely have only ~13 steps left at 3.4kHz oscillator frequency).

Therefore, you definitely want to go for resampling for SID emulation.
My point was not about SID emulation. But I'd agree you need such things for accurate emulation.

Quote:
As I said, I think this discussion makes only sense within the context of the whole system.

I don't need to bother with this ridiculous argument that PAULA can not do sound on it's own when I have a full 16-bit CPU and DMA system with 512k memory behind it.

I think that the Amiga was originally a perfectly balanced system, feature-wise.
I almost fully agree. I still think the designers could've added two or four more voices. The system easily has the bandwidth for it and it would've made a lot of things much easier/nicer.

But for 1985 it was pretty good


Quote:
Originally Posted by Hewitson View Post
Actually, the Megadrive is capable of playing back very high quality samples.
True, but only in the same sense that the A500 can play back 8 channels of audio - i.e. playing back high quality samples on the Megadrive uses a lot of CPU time to do. Megadrive samples are not DMA based, they require software driving them.

But you are correct, it is possible though only rarely done in games.

Last edited by roondar; 10 January 2019 at 17:46. Reason: typo / rephrased something
roondar is offline  
Old 10 January 2019, 13:31   #84
Hewitson
Registered User
Hewitson's Avatar
 
Join Date: Feb 2007
Location: Melbourne, Australia
Age: 37
Posts: 3,406
Quote:
Originally Posted by roondar View Post
True, but only in the same sense that the A500 can play back 8 channels of audio - i.e. playing back high quality samples on the Megadrive uses a lot of CPU time to do. Megadrive samples are not DMA based, they require software driving them.
I realise that the YM2612 in the Megadrive is not capable of DMA transfers. I do not know how much CPU time playing back a high quality sample would use, but see Mega Turrican for an example of reasonable quality samples being played with an absolute shitload of stuff going on.

Good (and bad) quality Megadrive sample playback: [ Show youtube player ]

Last edited by Hewitson; 10 January 2019 at 13:51. Reason: Found a better youtube video.
Hewitson is offline  
Old 10 January 2019, 14:10   #85
drHirudo
Amiga user
drHirudo's Avatar
 
Join Date: Nov 2008
Location: Sofia / Bulgaria
Posts: 214
Big grin

All the people writing about the AY superiority and the SID being able to prevail in capability everywhere, forgot how good the Paula sounds in reality! And in the start of the thread the person was telling about how some of the AY Atari ST tunes sounded better than their Amiga equivalents, which sometimes were even made by different people. Pretty much a matter of taste to the listener, not about technical capability, which the Amiga+Paula beats against the ST+AY (even the 8-bit Oric can play Atari ST tunes).

When I first got my Amiga many people coming from PC, Atari, C64, Oric, MSX, Apple II or any other platform didn't believe that a computer was producing such awesome music and sound effects (sometimes samples). Even the early Sound Blasters didn't sound that good, especially since it was mostly MIDI sounding music for them.

And regarding the SID. Probably not many of you know or care, but the SID had a philosophical and designer's successor - the Ensoniq ES5503 DOC 8-bit wavetable synthesis sound chip, with 32-channels, stereo. It was used on a computer with the 6510's philosophical and designer's successor - the 65816 - all this on the Apple IIGS computer. Anyway, even the Apple IIGS music sounds crap compared to the Amiga's Paula music, but it was mostly because not much talent really gathered on the IIGS and Apple was mostly focusing their efforts on the Mac line, ignoring the good old Apple II line that made them big in the first place.

To me, the AY/YM chip in the Atari ST line was chosen because it was cheap and the de-facto standard sound generating device of the early to mid 1980-ies (unlike the proprietary SID, by Commodore) and it was used in countless machines, computers, consoles, arcade games, sound boards etc - for example Sincair ZX Spectrum, Oric, Apple II sound cards - https://en.wikipedia.org/wiki/Mockingboard.

Here are how good the AY (Mockingboard) tunes can be when used properly even on 8-bit machine:
[ Show youtube player ]

[ Show youtube player ]

But then again, the Atari ST designers wanted to cut some dollars and added only one AY chip, even if the 68000 can happily control two at the same time without much expense .
drHirudo is offline  
Old 10 January 2019, 14:25   #86
roondar
Registered User

 
Join Date: Jul 2015
Location: The Netherlands
Posts: 1,381
Quote:
Originally Posted by Hewitson View Post
I realise that the YM2612 in the Megadrive is not capable of DMA transfers. I do not know how much CPU time playing back a high quality sample would use, but see Mega Turrican for an example of reasonable quality samples being played with an absolute shitload of stuff going on.

Good (and bad) quality Megadrive sample playback: [ Show youtube player ]
I'll confess, some of those samples do sound better than I had expected.

Most however, sounded fairly poor to me. This seems to be par for the system. Sometimes samples are rather nice, mostly they are not.

Edit: I went and looked up Mega Turrican as you gave it as an example and I don't really agree the samples in that game are that good. The power up vocals are clearly much worse than their Amiga counterparts.

Do note that I was not at all saying saying that it is impossible to get good samples out of the MD (and as you just shown the MD did get games with good quality samples in them), just that it requires you to have the raster time & know how to do so properly and as such was fairly rare. Perhaps my initial statement was a bit too harsh, but it's not really a secret that generally the MD did not normally use high quality samples.

Did the MD actually have more than one sample channel?

Last edited by roondar; 10 January 2019 at 14:38. Reason: Typo
roondar is offline  
Old 10 January 2019, 16:14   #87
robinsonb5
Registered User
 
Join Date: Mar 2012
Location: Norfolk, UK
Posts: 626
Quote:
Originally Posted by roondar View Post
but it's not really a secret that generally the MD did not normally use high quality samples.
ROM space will have been a factor here, too.
robinsonb5 is offline  
Old 10 January 2019, 18:12   #88
pandy71
Registered User
 
Join Date: Jun 2010
Location: PL
Posts: 1,690
Quote:
Originally Posted by dmacon View Post
Do you hear an audible difference between 3.58 MHz and 985kHz sample rate?
Well, DSD use 2.8224MHz 1-bit to be High End audio source (120dBFS).
So i estimate that difference between 985kHz and 3.58MHz is around 50dB
Are you able to hear difference between 7 and 16 bits?

Whenever numbers are used in this thread they should be used in proper context. Going further this way we can say that ZX Spectrum can be considered "close to DSD" as it may theoretically use 1 bit speaker with for example software second order delta sigma modulator and with sufficient oversampling (16 times i.e. 700kHz) will be sufficient to deliver 8 bit quality 44.1kHz and as noise shaping can be used this may lead to subjective 10 - 14 bit audio.
And 5-th order delta sigma may deliver 8 bit with only around 5 times oversampling... Numbers without context may mislead...
pandy71 is offline  
Old 10 January 2019, 20:43   #89
frank_b
Registered User
 
Join Date: Jun 2008
Location: Boston USA
Posts: 441
I like SID voice YM stuff on the ST. Tao of cream is an excellent composer.
[ Show youtube player ]
[ Show youtube player ]
Good music is good music.

People have recently been experimenting with chip (YM) + DMA sound on the STe.
The results aren't too bad either http://www.atari-forum.com/download/file.php?id=38144

Last edited by frank_b; 10 January 2019 at 20:54.
frank_b is offline  
Old 10 January 2019, 22:53   #90
touko
Registered User

touko's Avatar
 
Join Date: Dec 2017
Location: france
Posts: 98
Quote:
Originally Posted by Hewitson View Post
I realise that the YM2612 in the Megadrive is not capable of DMA transfers. I do not know how much CPU time playing back a high quality sample would use, but see Mega Turrican for an example of reasonable quality samples being played with an absolute shitload of stuff going on.

Good (and bad) quality Megadrive sample playback: [ Show youtube player ]
The Md use the Z80 CPU for sound, he is here for master system compatibility, and exclusively dedicated to manage the sound in MD mode .

Sample playback can be really good too on MD:
[ Show youtube player ]

Speaking of samples rate, i think the PCengine is the best, you can easily go beyond 100 khz .

Quote:
Actually, the Megadrive is capable of playing back very high quality samples.
8 bits 26khz max .

Last edited by touko; 10 January 2019 at 23:14.
touko is offline  
Old 11 January 2019, 00:02   #91
roondar
Registered User

 
Join Date: Jul 2015
Location: The Netherlands
Posts: 1,381
Quote:
Originally Posted by frank_b View Post
I like SID voice YM stuff on the ST. Tao of cream is an excellent composer.
[ Show youtube player ]
[ Show youtube player ]
Good music is good music.
Very, very true - good music is good music regardless of what playback device is used.

Quote:
People have recently been experimenting with chip (YM) + DMA sound on the STe.
The results aren't too bad either http://www.atari-forum.com/download/file.php?id=38144
IIRC the STe allows for 50KHz sample playback, is that correct?

I'm curious, are there any well known tricks (as in something like 14 bit mode) for STe DMA sound ?
roondar is offline  
Old 11 January 2019, 00:17   #92
Foebane
Registered User
 
Join Date: Sep 2011
Location: Cardiff, UK
Age: 47
Posts: 1,844
Quote:
Originally Posted by roondar View Post
IIRC the STe allows for 50KHz sample playback, is that correct?

I'm curious, are there any well known tricks (as in something like 14 bit mode) for STe DMA sound ?
I don't think so. As far as I know, the STE DMA PCM audio is only two-channel for stereo purposes, and so you can't even play back Amiga mods with it.

However, some STE demos use streaming audio or long repeating samples mixed with the old AY sound chip, and the best example I've ever found of streaming is this, my favourite ST(E) demo:

[ Show youtube player ]

I think you'll agree, it is rather nice in graphics as well.
Foebane is online now  
Old 11 January 2019, 01:37   #93
Estrayk
Registered User

Estrayk's Avatar
 
Join Date: Apr 2015
Location: Spain
Posts: 190
Quote:
Originally Posted by roondar View Post
IIRC the STe allows for 50KHz sample playback, is that correct?
yep. And Amiga modules sounds very good.
[ Show youtube player ]
Estrayk is online now  
Old 11 January 2019, 01:57   #94
Foebane
Registered User
 
Join Date: Sep 2011
Location: Cardiff, UK
Age: 47
Posts: 1,844
Quote:
Originally Posted by Estrayk View Post
yep. And Amiga modules sounds very good.
[ Show youtube player ]
Yeah, when I said that the STE couldn't play back Amiga mods, I meant that it couldn't do so without modification. The STE DMA PCM audio chip is still TWO-channel, isn't it? So doesn't it have to do some processing to mix channels together for Amiga mod playback??
Foebane is online now  
Old 11 January 2019, 02:49   #95
Hewitson
Registered User
Hewitson's Avatar
 
Join Date: Feb 2007
Location: Melbourne, Australia
Age: 37
Posts: 3,406
Quote:
Originally Posted by drHirudo View Post
Here are how good the AY (Mockingboard) tunes can be when used properly even on 8-bit machine:
[ Show youtube player ]
Incredible! Can't stop listening to the cover of Hung Up, love that song.

What's the first song? Sounds very familiar but can't quite put my finger on it.

Quote:
Originally Posted by roondar
Did the MD actually have more than one sample channel?
The YM2612 doesn't. The PSG can be abused to play samples as well though.

Quote:
Originally Posted by touko
The Md use the Z80 CPU for sound, he is here for master system compatibility, and exclusively dedicated to manage the sound in MD mode .
Many games drove the YM directly from the 68000. It actually produces better results.
Hewitson is offline  
Old 11 January 2019, 10:09   #96
frank_b
Registered User
 
Join Date: Jun 2008
Location: Boston USA
Posts: 441
Quote:
Originally Posted by Foebane View Post
Yeah, when I said that the STE couldn't play back Amiga mods, I meant that it couldn't do so without modification. The STE DMA PCM audio chip is still TWO-channel, isn't it? So doesn't it have to do some processing to mix channels together for Amiga mod playback??
With the lance playback routine it usies 50 khz mixing with 20-25% of the CPU on an 8 mhz 68000. There's supposed to be a trick used with the MC and LMC.
That same routine can apparently do 8 voice mixing with 40% CPU time.
The quality of the playback is excellent.

BTW The Falcon can do mod playback with 32 stereo voices using 16 bit mixing at 50 khz. All on the DSP.
frank_b is offline  
Old 11 January 2019, 10:23   #97
donnie
Registered User
 
Join Date: Jul 2010
Location: sthlm
Posts: 192
[ Show youtube player ]
donnie is offline  
Old 11 January 2019, 11:05   #98
touko
Registered User

touko's Avatar
 
Join Date: Dec 2017
Location: france
Posts: 98
Quote:
Many games drove the YM directly from the 68000. It actually produces better results.
For FM musics, yes you can use the 68k, but it's a pure waste, and you cannot play in-game samples with it without having a big impact on performances because you cannot synchronise the sending of sample datas with the DAC as it needs cycles count for that,and you must add this with the slowness of the FM chip.
Plus, the 68k is halted in VBLANK if you use the DMA,really bad for samples playback .

Even for FM, i don't see how the 68k can be more efficient than a dedicated CPU,mainly when the 68k will loose a big chunk of cycles to deal with the slow 2612 .

Last edited by touko; 11 January 2019 at 11:13.
touko is offline  
Old 11 January 2019, 14:33   #99
roondar
Registered User

 
Join Date: Jul 2015
Location: The Netherlands
Posts: 1,381
Quote:
Originally Posted by frank_b View Post
With the lance playback routine it usies 50 khz mixing with 20-25% of the CPU on an 8 mhz 68000. There's supposed to be a trick used with the MC and LMC.
That same routine can apparently do 8 voice mixing with 40% CPU time.
The quality of the playback is excellent.

BTW The Falcon can do mod playback with 32 stereo voices using 16 bit mixing at 50 khz. All on the DSP.
I wonder if you did not have access to 'tricks' for mixing how much time it would take the 68000 to mix two channels at low/medium frequencies (not playing back, just the mixing). Say 8, 11 or 22KHz. If that number is low enough it might be possible to build a usable mixer to use in Amiga games for say playing back SFX on a dedicated channel (or two dedicated channels). Obviously it would still cost resources, but it might be worthwhile any way.

You could even consider a special purpose mixer where all source samples have to be one specific frequency (i.e. the output frequency) to make mixing easier.

Just a thought anyway.
roondar is offline  
Old 11 January 2019, 15:12   #100
frank_b
Registered User
 
Join Date: Jun 2008
Location: Boston USA
Posts: 441
Quote:
Originally Posted by roondar View Post
I wonder if you did not have access to 'tricks' for mixing how much time it would take the 68000 to mix two channels at low/medium frequencies (not playing back, just the mixing). Say 8, 11 or 22KHz. If that number is low enough it might be possible to build a usable mixer to use in Amiga games for say playing back SFX on a dedicated channel (or two dedicated channels). Obviously it would still cost resources, but it might be worthwhile any way.

You could even consider a special purpose mixer where all source samples have to be one specific frequency (i.e. the output frequency) to make mixing easier.

Just a thought anyway.
Would also give you more RAM for storing samples. Only the playback buffer needs to be in chip ram. It's been done in at least one game on the STe. Obsession on the STe uses 4 channel modules for music and two channels for the sound effects.
On the Amiga the music cuts out when an effect is played.

Btw here's the quality you can expect from the Falcon playing modules.
[ Show youtube player ]
[ Show youtube player ]
[ Show youtube player ]

Last edited by frank_b; 11 January 2019 at 15:40.
frank_b is offline  
 


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools

Similar Threads
Thread Thread Starter Forum Replies Last Post
Yamaha CRW4416SX - wanting to sell B14ck W01f New to Emulation or Amiga scene 1 04 December 2016 11:58
Yamaha SCSI (external Drive) ReD MarketPlace 0 05 November 2011 14:25
Amiga soundchip (Paula) Grimbog Amiga scene 7 28 September 2008 19:04
Yamaha CRW4416SX DDNI support.Hardware 20 20 June 2006 21:28
EBAY Amiga / Atari ST / Atari 2600 stuff Eggsplosion MarketPlace 0 09 October 2004 22:01

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +2. The time now is 01:45.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2019, vBulletin Solutions Inc.
Page generated in 0.10585 seconds with 16 queries