some fancy ideas for a extended (68k?) CISC-CPU

[meynaf]

Originally Posted by bloodline

It’s not 14bit... it’s going to have a better dynamic range than the normal 8bit, but as soon as you do this two channels are locked together, making you have to do the very mixing Meynaf is so against...

It is 14bit, actually slightly more if using a really good calibration.
And i'm not "so against" -- i'm for it when it's actually useful. It isn't if you simply play CD-like audio, which is two channels.
And please update your beliefs : Paula is true 8-bit PCM, not 1-bit square wave.

[Thorham]

Originally Posted by bloodline

AFAIK paula is essentially using a Sigma-delta DAC technique... so you will need a reconstruction filter regardless

I meant the low pass filter that makes the audio sound so dull. On various Amiga models the cut-off frecuency is too low, on the A1200 it's well over 13khz. Have tested this myself.

Originally Posted by bloodline

It’s not 14bit... it’s going to have a better dynamic range than the normal 8bit

Much better.

Originally Posted by bloodline

as soon as you do this two channels are locked together, making you have to do the very mixing Meynaf is so against...

You don't need mixing for CD style audio. For WAV files you have to do some endian conversion, some bit shifting, and use a calibration lookup table. Not a big deal.

Mixing and resampling isn't a big deal, either. In games that run on a faster CPU, you could easily use CD style music and mix in some fixed rate sound effects. Quite cheap to do.

Originally Posted by bloodline

You aren’t doing 44.1kHz unless you have shutdown the video DMA... that’s not a practical for most situations.

Because of a bug in the hardware, when you use double scan video modes, the maximum audio DMA speed doubles. I use these modes on my A1200. Anyway, I've actually down sampled some CDs to 28khz with SOX on the peecee, and it sounds fine, but yes, you can play back 44khz with DMA.

Originally Posted by bloodline

If you want to use your Amiga as a Dedicated CD player, which sounds almost as good as a real one, all power to you!

It's just fun to have because of how old Paula is.

Originally Posted by bloodline

I prefer to use my Amiga for something it’s good at doing

An A1200 isn't bad at it.

[Thomas Richter]

Originally Posted by Thorham

I'm as nostalgic as a pile of bricks! No, really. I never stopped using my A1200.

For *what* please? For toying around yes. For the average computing business? Hardly.

[Thomas Richter]

Originally Posted by meynaf

It is 14bit, actually slightly more if using a really good calibration.

Check your math, please. And your ears. That's just denial of reality - fanboyism.

Originally Posted by meynaf

And please update your beliefs : Paula is true 8-bit PCM, not 1-bit square wave.

If the frequency of the square wave is high enough, that is pretty much the same thing. But no matter, it is an 8bit DAC, with a precision and calibration of an 8-bit DAC.

[Thorham]

Originally Posted by Thomas Richter

For *what* please? For toying around yes. For the average computing business? Hardly.

Programming. Something which most certainly isn't toying around.

[Thorham]

Originally Posted by Thomas Richter

And your ears. That's just denial of reality - fanboyism.

You obviously just haven't heard it on an A1200.

[bloodline]

Originally Posted by meynaf

You're not trolled, but indeed language barriers don't help. While my English isn't abysmal (in its written form), it's clearly not the right tool.

Your English is great, but I always worry that subtleties are lost when discussing technical matters.

And you think peecees, which (unless using special sound board) are single DMA channel (regardless of number of voices) do better ?
Everyone has single buffer played at any time !
But you can "program" next buffer, if it's what you mean, to seamlessly continue playing.

My point was to try and explain that using the DMA channel as an instrument is very flawed... it was onto a good idea when CPU time and RAM bandwidth were both limited.

With default 68EC020 of A1200 SuperStardust already manages to play multichannel s3m during the menu. So Thorham is right : 68030 !

Then SuperStardust is using CPU mixing, which will introduce latency into the audio path... something which you criticised heavily regarding the presence of!

The ability to compile and run without errors are not my definition of "working".

Ok... have you compiled it?

Oh no. You really seem to have absolutey no idea of how the Amiga plays sounds...
Please let me explain.
The period isn't just the DMA fetch period. It is actually the real speed at which the AUDxDAT (audio data registers) are fed to the DAC. Whenever AUDxDAT is empty (it contains two samples), the DMA refills it.
You can also feed these registers with CPU to remove DMA speed limits.
And the time between two samples is of course constant.

Probably the language barrier issue here. I’m aware how the DMA fetch works... Whatever, my point regarding the quality issues remain valid.

It is of course filtered out, there are no aliasing harmonics.



Your memory doesn't serve
The DACs are true 8-bit, using PWM for volume.

According to this thread I appear to be correct (Ok, it’s PWM rather than PDM, but that’s an implementation issue for a sigma-delta DAC).

http://eab.abime.net/showthread.php?t=63227&page=4

I’m not surprised as I’ve used LP filtered PWM for 8bit audio on micro controllers for years.

Except that the DMA doesn't miss in case of system overload.

If the system is overloaded, and your PC misses a thread scheduling, it’s no different from the Amiga being overloaded and missing an Audio interrupt...

This does not in any manner make it better than the Amiga, far not.
No, my A1200 has always sounded great.

But audio reproduction quality is quantifiable... and even a simple cross platform Audio API like the SDL2 one is better quality and more flexible than the Amiga, which was designed for a different time.

You seem to not understand what i write...

As I said, there might me a language issue

[meynaf]

Originally Posted by Thomas Richter

Check your math, please. And your ears. That's just denial of reality - fanboyism.

My math is right. Yours not. Because D/A are not linear, the number of discrete levels is more than 16384 (we play 2x8 bits = 65536 combinations).
And i have excellent ears for my age.
This isn't fanboyism, it's a reality you just desperately want to deny because of your hatred for the Amiga - and perhaps for my person as well.

[Thomas Richter]

Originally Posted by Bruce Abbott

MS-DOS is an operating system, according to everyone but you. Make up your own definition if you like, but don't expect anyone else to agree with it.

Whatever you say, but CP/M is a joke, and it died out for reasons. It has absolutely nothing to offer, except a command processor and a file management system. Yes, people called that a DOS (with 'D') back then.

Originally Posted by Bruce Abbott

After being introduced by IBM in 1981 as PC DOS, it very quickly became the OS standard for desktop computers, remaining so for the entire period of the Amiga's existence. For something that 'isn't much' its influence was enormous, so much that no other OS managed to dent the hold it had over the market.

I don't deny that it had an enormous impact, but there were very good reasons why MS replaced it by the NT kernel, and why it is completely obsolete today, and nobody uses it today anymore.


As said, CBM never made this step of throwing AmigaOs overboard and replacing it by a serious Os, due to lack of resources.

Originally Posted by Bruce Abbott

But of course you have to dismiss MS-DOS, because if you didn't you would have to admit that Amiga OS was more sophisticated than the industry standard 'serious' OS of the time.

I beg your pardon, we don't need to argue about that. Of course AmigaOs was more sophisticated *back then*. But *back then* is *back then*, and other vendors moved away from such primitive systems. MacOs was also more sophisticated than MS-DOS. It had other advantages, and other drawbacks.

Originally Posted by Bruce Abbott

And that's just games. I don't know about where you come from, but down here in New Zealand virtually all Amigas were bought by people who used their expensive purchase for more than just games. I should know, I owned and operated a computer store selling Amigas from 1991 to 2000.

In Germany, the Amiga was more or less an "updated C64" and typically used as such. Some people bought the big box machines, but there wasn't much serious usage of them except in some niches. TV stations used them occasionally as Amigas had genlock and created a TV-compatible system, but that also died out.

Originally Posted by Bruce Abbott

No, it didn't need to. Most Amiga users were quite happy with how the OS worked - it was hardware performance and compatibility with the 'industry standard' that the Amiga needed to improve. The vast majority of users who switched to PCs did so not because Amiga OS was lacking, but because the latest games, apps, and peripherals were only being produced for PCs. By the time PC's 'needed' a better OS Commodore was long gone and nothing could save the Amiga.

That's all true, but doesn't change a thing that AmigaOs is far far behind, and not a serious operating system by any reasonable modern standard.

Originally Posted by Bruce Abbott

There you go again, redefining the word 'modern' to mean only stuff that is specific to PCs.

Heck, no. Back in time when I had only the Amiga, and no PC after all, there was certainly hardware that was more modern than the Amiga, and operating systems that were more advanced than AmigaOs. Back then, I worked on HP/UX, IRIX and SunOs. Those were rock stable machines.

Originally Posted by Bruce Abbott

Competitive with what? There is more to life than desktop PCs and smartphones.

So what's the market? Serious computing? That happens on workstations, mostly x64 technology, probably some arm in the future. Mobile computing? That's arm, and smartphones. All-purpose computing at home? That's on PCs. Serious gaming? That's on dedicated hardware, mostly build around x64, and PCs. Maker-Scene? That's on arm and raspy.


All of them have operating systems that are far ahead, by multiple decades now.

Originally Posted by Bruce Abbott

More and more people today are finding that 'modern' is actually not that interesting to them.

What's "more and more"? We're talking about a niche here, and that nice is probably the "maker's scene", which goes in the raspy direction. I have one at home, too. Nice powerful tiny gadget for sure. Good operating system on it as well.

Originally Posted by Bruce Abbott

There is no fundamental reason why an Amiga inspired architecture couldn't make inroads into that market, even if just as a 'toy'. Toys are actually big business.

Toys may, yes, indeed. But the Amiga is even behind that.

Originally Posted by Bruce Abbott

Nonsense. Smartphones are not a 'path', they are a mature tech and a dead end. In another 10 years they won't be significantly different from now.

I wouldn't be so sure about that, but we'll see.

Originally Posted by Bruce Abbott

But today I see many people riding ebikes and scooters who would have laughed at the idea a few years ago. I still only see electric cars occasionally (apart from my own one), but the market is expanding rapidly and every major manufacturer has committed to transitioning to electric - even those who only recently considered them a joke. Why? Because some people (the 'trend setters') bought electric vehicles because they were interesting, despite having severe limitations compared to gas cars.

No, I'm sorry. Electric motors have been all around, just that they haven't been used much for mobility due to the lack of competitive battery technology, so one important piece of the puzzle was missing (and still is, to a good degree). The energy density of a battery is small compared to gasoline, and that will still be a problem for a long time.


That's not comparable to AmigaOs which does not satisfy the requirements of today's market. It's not that there is a single component that is probably missing - the overall design is just outdated.

Originally Posted by Bruce Abbott

Think about where you are, and consider that perhaps people here are not so interested in turning their Amigas into 'modern' computers. If they were, they would just get a PC!

That's what is called "retro". It's a simple effect of our generation (which grew up with the machines) now getting into an age where they can afford again to toy with them, and invest money into them. As soon as our generation died out (probably 30 years from now), nothing will be left of the Amiga except a couple of museum pieces, and probably a handful of collectors.

Originally Posted by Bruce Abbott

Think of it more as finding ways to power our bicycles electrically so we can ride them on the same roads as 'modern' cars. And while we might be only be riding for fun, our bikes could also have 'serious' applications.

Unlike cars or bikes, there isn't much AmigaOs has to offer today, that's the problem. It's not a matter of an upgrade or engine change - there are serious design issues that hinder it from operating in a modern environment.


Internet, security, robustness: These are all issues for a machine operating in a potentially hostile environment. As long as the machine remains exotic enough and a niche, nobody would care ("security by obscurity"), but with the design issues it has, it cannot become successful or enter a mass market.

[Thomas Richter]

Originally Posted by meynaf

My math is right. Yours not. Because D/A are not linear, the number of discrete levels is more than 16384 (we play 2x8 bits = 65536 combinations).

Counting theorem. Fix your math.

Originally Posted by meynaf

This isn't fanboyism, it's a reality you just desperately want to deny because of your hatred for the Amiga - and perhaps for my person as well.

I don't hate the Amiga. I'm just not a fanboy. I see it where it is: An old outdated technology that was quite interesting at its time, but that is long obsolete. I still play around with it for good old times, and because I like to, but I'm not blind or death.



The problem is that you deny reality - and that's a problem of a fanboy. You need to get your hands dirty with some new technology to see and understand the difference.

[meynaf]

Originally Posted by bloodline

Your English is great, but I always worry that subtleties are lost when discussing technical matters.

Many things are lost, actually.

Originally Posted by bloodline

My point was to try and explain that using the DMA channel as an instrument is very flawed... it was onto a good idea when CPU time and RAM bandwidth were both limited.

It's not worse than having to resample everything.

Originally Posted by bloodline

Then SuperStardust is using CPU missing, which will introduce latency into the audio path... something which you criticised heavily regarding the presence of!

You don't have finely shaded opinions, apparently.
I criticized the presence of too much latency.

Originally Posted by bloodline

Ok... have you compiled it?

Why would I ? It's useless as a player.

Originally Posted by bloodline

Probably the language barrier issue here. I’m aware how the DMA fetch works... Whatever, my point regarding the quality issues remain valid.

Don't charge language barriers too much. No, you're clearly not aware how the DMA fetch works. DMA fetch and data sent to D/A are two distinct processes, this is what you have to understand, and this totally invalidates your point regarding the quality issues.

Originally Posted by bloodline

According to this thread I appear to be correct (Ok, it’s PWM rather than PDM, but that’s an implementation issue for a sigma-delta DAC).

http://eab.abime.net/showthread.php?t=63227&page=4

I’m not surprised as I’ve used LP filtered PWM for 8bit audio on micro controllers for years.

Then reread that thread.
PWM is only for volume handling, not for 8bit audio generation !
So if you play at max volume there is no PWM involved at all.

Originally Posted by bloodline

If the system is overloaded, and your PC misses a thread scheduling, it’s no different from the Amiga being overloaded and missing an Audio interrupt...

It is very different. Just run tasks on a PC, it's enough. But you can overload an Amiga up to 100% cpu use and it will still not miss an Audio interrupt because it has a very high priority.

Originally Posted by bloodline

But audio reproduction quality is quantifiable...

Not really, no.

Originally Posted by bloodline

and even a simple cross platform Audio API like the SDL2 one is better quality and more flexible than the Amiga, which was designed for a different time.

It is not "better quality", this is a hardware issue, not software.
It is not in any manner more flexible, as everything it can do, the Amiga also can, and more.

Originally Posted by bloodline

As I said, there might me a language issue

I'm starting to really doubt about this

[meynaf]

Originally Posted by Thomas Richter

Counting theorem. Fix your math.

No, fix yours.
But apparently you don't even know how 14-bit replay works. How could you count correctly in these conditions.

Originally Posted by Thomas Richter

I don't hate the Amiga. I'm just not a fanboy. I see it where it is: An old outdated technology that was quite interesting at its time, but that is long obsolete. I still play around with it for good old times, and because I like to, but I'm not blind or death.

I'm not a fanboy either. It was a good machine, and still is. The one and only that gives full freedom and pleasure to code. This is what I value. Give me another machine giving the same, and i'll jump on it (and no, peecees clearly don't apply).

Originally Posted by Thomas Richter

The problem is that you deny reality - and that's a problem of a fanboy.

I don't deny reality. I deny what you think is reality.

Originally Posted by Thomas Richter

You need to get your hands dirty with some new technology to see and understand the difference.

New technology ? Like what ? A PC, that was designed in 1980 ?
Thanks but not thanks. I already have this experience, and multiple times.

[Thomas Richter]

Originally Posted by meynaf

New technology ? Like what ? A PC, that was designed in 1980 ?
Thanks but not thanks. I already have this experience, and multiple times.

There isn't much left of the PC from 1980, really. The strength of this platform was (and currently still is) that its vendors always found a way to upgrade it without breaking compatibility too much. There was always enough critical mass to finance this upgrade path, unlike what happened to the Amiga.


Yes, I dared to critize your beloved machine, but you should really get in touch with reality. There are so many things an Amiga cannot do right, or does not perform right, or has serious problems with. You do not notice, or may choose to ignore, because you're in a mode where you just want to defend your hobby - against me, probably against yourself as well.



That's just all natural, but it doesn't change physical limits, and design limits.



With a 8-bit DAC, calibrated with tolerances for 8bit-reproduction, you cannot reproduce 16-bit sound faithfully. Yes, one can measure distortion precisely. That's actually the easiest exercise.



However, one cannot measure fanboy-ism precisely, that's the problem. You feel attached to your hobby, you like its deficencies - that's psychology, not physics. I cannot defect psychology, but it doesn't matter in the end.



There is no critical mass of fanboys to make an upgrade worthwhile, so it wouldn't happen. You won't make the change.


I contributed my part, and saying that "I hate the Amiga" is just irony. If I would, I wouldn't have done what I did over the last four years. I just have a realistic view on what all this is about; I don't live in a dream world.

[Thomas Richter]

Originally Posted by meynaf

It is not "better quality", this is a hardware issue, not software.
It is not in any manner more flexible, as everything it can do, the Amiga also can, and more.

See, that's exactly the problem. You choose to close your eyes. That's not a solution.

[Thorham]

Originally Posted by Thomas Richter

With a 8-bit DAC, calibrated with tolerances for 8bit-reproduction, you cannot reproduce 16-bit sound faithfully. Yes, one can measure distortion precisely. That's actually the easiest exercise.

The 14bit trick does get you close, though. Decent for CD audio, nice to have for playng mods with more than four channels and also nice for playing four channel mods with DeliTracker's headphone mixer.

[bloodline]

Originally Posted by meynaf

Many things are lost, actually.

Perhaps too much...

It's not worse than having to resample everything.

My point is that PC or Amiga, your audio buffer is being resampled. The DAC operates at a single conversion rate. You just get more options on a modern computers.

You don't have finely shaded opinions, apparently.
I criticized the presence of too much latency.

Well, for professional audio equipment anything less than ~10ms is acceptable. My main audio machines are Macs which have a 3ms round trip (Audio in->DSP-> Audio Out).

The only poor latency API I’m aware of is the now long since redundant Win32 audio API.

Why would I ? It's useless as a player.

Shame, it might have been fun to turn it into something more feature complete, I don’t have the time to work on it myself.

Don't charge language barriers too much. No, you're clearly not aware how the DMA fetch works. DMA fetch and data sent to D/A are two distinct processes, this is what you have to understand, and this totally invalidates your point regarding the quality issues.



Then reread that thread.
PWM is only for volume handling, not for 8bit audio generation !
So if you play at max volume there is no PWM involved at all.

Hmmm... best just to read Toni’s posts... looks like a 3.58Mhz PWM... look at my tracker code, I use that same frequency to derive the playback frequency of each sample!

It is very different. Just run tasks on a PC, it's enough. But you can overload an Amiga up to 100% cpu use and it will still not miss an Audio interrupt because it has a very high priority.

That’s true the Audio interrupts are a high priority, but the tracker code won’t be running in interrupt... once the buffer ends are reached... if the CPU is too busy to setup the next DMA, there will be no sound.

The difference with the Mac and the A500, in my studio, is that my Mac is doing loads of things and can still run nearly a hundred tracks of audio all with DSP plugins, and generating a MIDI clock... while recording the output of my A500 JUST running OctaMED.

It is not "better quality", this is a hardware issue, not software.
It is not in any manner more flexible, as everything it can do, the Amiga also can, and more.

This is what Thomas referred to as Fanboism...

[meynaf]

Originally Posted by Thomas Richter

There isn't much left of the PC from 1980, really. The strength of this platform was (and currently still is) that its vendors always found a way to upgrade it without breaking compatibility too much. There was always enough critical mass to finance this upgrade path, unlike what happened to the Amiga.

There is more left than you think. Like, say, the basic design. It's not as if they changed the cpu family.

Originally Posted by Thomas Richter

Yes, I dared to critize your beloved machine, but you should really get in touch with reality. There are so many things an Amiga cannot do right, or does not perform right, or has serious problems with. You do not notice, or may choose to ignore, because you're in a mode where you just want to defend your hobby - against me, probably against yourself as well.

Nonsense. It's not my "beloved machine". There are so many things PC cannot do right, or does not perform right, or has serious problems with. You do not notice (etc).

Originally Posted by Thomas Richter

That's just all natural, but it doesn't change physical limits, and design limits.

This was a reply to what ?

Originally Posted by Thomas Richter

With a 8-bit DAC, calibrated with tolerances for 8bit-reproduction, you cannot reproduce 16-bit sound faithfully. Yes, one can measure distortion precisely. That's actually the easiest exercise.

I don't need to reproduce 16-bit sound faithfully. I just want something that sounds good, and it does. This is the easiest exercise.

Originally Posted by Thomas Richter

However, one cannot measure fanboy-ism precisely, that's the problem. You feel attached to your hobby, you like its deficencies - that's psychology, not physics. I cannot defect psychology, but it doesn't matter in the end.

I am not fanboy. I stated it and explained it. If "fanboyism" is your only argument, ya better stfu.

Originally Posted by Thomas Richter

There is no critical mass of fanboys to make an upgrade worthwhile, so it wouldn't happen. You won't make the change.

Why would i reply, you're not even able to quote properly.

Originally Posted by Thomas Richter

I contributed my part, and saying that "I hate the Amiga" is just irony. If I would, I wouldn't have done what I did over the last four years. I just have a realistic view on what all this is about; I don't live in a dream world.

What contribution ? The bugs you mentioned in OS 3.9 ? That useless stack cookie ?

Originally Posted by Thomas Richter

See, that's exactly the problem. You choose to close your eyes. That's not a solution.

I see things you don't. You're locked in your small, boring world.
Amiga is a lot more flexible to code on than so-called modern machines, and there is no amount of insisting from your part which will change that.

[meynaf]

Originally Posted by bloodline

Perhaps too much...

That's very possible.

Originally Posted by bloodline

My point is that PC or Amiga, your audio buffer is being resampled. The DAC operates at a single conversion rate. You just get more options on a modern computers.

No, Amiga audio isn't resampled. It's played "as is".
No, you don't have more options on modern computers. You just have fixed frequency output and have to bear with that.

Originally Posted by bloodline

Well, for professional audio equipment anything less than ~10ms is acceptable. My main audio machines are Macs which have a 3ms round trip (Audio in->DSP-> Audio Out).

The latency of Amiga sound (with dma) is around 0.5ms (for low frequencies ; higher freqs make it smaller).

Originally Posted by bloodline

The only poor latency API I’m aware of is the now long since redundant Win32 audio API.

You mean WaveOut ? This is what i used in my audio test.
But, DirectSound is obsolete. I tried to check others, like WASAPI, but it does not look simple and i couldn't find usable sample code.

Originally Posted by bloodline

Shame, it might have been fun to turn it into something more feature complete, I don’t have the time to work on it myself.

It might have been fun indeed, but i think i'll skip that one. I have enough to do on my own.

Originally Posted by bloodline

Hmmm... best just to read Toni’s posts... looks like a 3.58Mhz PWM... look at my tracker code, I use that same frequency to derive the playback frequency of each sample!

Yes, read Toni's posts. If you believe he confirms your claim, quote him here (or link to the post itself).
And looking at your tracker code, what do i see ? 44100.
Once and for all, Amiga sounds is 8-bit d/a, not 1-bit d/a. Even Thomas agreed with that !

Originally Posted by bloodline

That’s true the Audio interrupts are a high priority, but the tracker code won’t be running in interrupt... once the buffer ends are reached... if the CPU is too busy to setup the next DMA, there will be no sound.

Of course yes, the tracker code also runs in an interrupt !
Listen, pal, i've written enough players for Delitracker to know how sound is made, and you're far away from it.

Originally Posted by bloodline

The difference with the Mac and the A500, in my studio, is that my Mac is doing loads of things and can still run nearly a hundred tracks of audio all with DSP plugins, and generating a MIDI clock... while recording the output of my A500 JUST running OctaMED.

But this is just A500. It's night and day in comparison to expanded A1200.

Originally Posted by bloodline

This is what Thomas referred to as Fanboism...

This is what Thomas wrongly referred to as Fanboism...

Frankly, if i were the fan boy you both mention, why the heck would i be currently porting my system framework to Windows ? Are you conscious this is complete nonsense ?

[Thomas Richter]

Originally Posted by meynaf

No, Amiga audio isn't resampled. It's played "as is".
No, you don't have more options on modern computers. You just have fixed frequency output and have to bear with that.

I already told you that this is wrong. Repeating this doesn't make it better.

Originally Posted by meynaf

The latency of Amiga sound (with dma) is around 0.5ms (for low frequencies ; higher freqs make it smaller).

That depends on what else is running on the machine. Not that we haven't had this argument before.

Originally Posted by meynaf

And looking at your tracker code, what do i see ? 44100.

But that's a configuration option. With alsa, you can set the playout frequency, and if you play back through the hw: output, you'll either get what you requested, or an error.

Originally Posted by meynaf

Once and for all, Amiga sounds is 8-bit d/a, not 1-bit d/a. Even Thomas agreed with that !

Yes, but why does that matter? A 1-bit D/A is easier to calibrate and more precise than an analog D/A, all provided the sampling frequency is high enough. That's just signal processing. There is less non-linearity in a 1-bit D/A than in Paula's 8bit DAC.

Originally Posted by meynaf

This is what Thomas wrongly referred to as Fanboism...

Closing your eyes from reality? That's what people do that fear a fair comparison.

Originally Posted by meynaf

Frankly, if i were the fan boy you both mention, why the heck would i be currently porting my system framework to Windows ? Are you conscious this is complete nonsense ?

Well, it is pretty close to nonsense if you think about it. According to your claims, windows cannot do what the Amiga does, so why bother? According to your logic, this is just a "bloat layer" around the windows API, programming the API directly would give you better performance, and would cost less memory. Furthermore, if you want a portable API for multimedia programming, SDL exists already and isn't really too bad.
Last but not least, if you want "more control" over the system, windows is the wrong target to begin with. If you want more control, there is Linux. I don't have a problem with audio latency, and if I want low-latency video playback, I have YUV overlays. If you like computers to toy with, then Raspi would be a better hardware platform as well.



The project contradicts your own claims and goals, so I wonder either why you complain, or why do you start such a project.

[bloodline]

Originally Posted by meynaf

No, Amiga audio isn't resampled. It's played "as is".
No, you don't have more options on modern computers. You just have fixed frequency output and have to bear with that.

You hand waved away my long list of issues with audio reproduction as performed by the Amiga. I’m trying to do you the courtesy of not doing the same to you. But you struggle to see that what you are saying here is irrelevant, particularly if the “fixed output frequency” is greater than double the maximum frequency you need to reproduce. Running the DAC slower doesn’t magically change the output.

The latency of Amiga sound (with dma) is around 0.5ms (for low frequencies ; higher freqs make it smaller).

It could operate with a latency of one nano second, that doesn’t change the fact that it’s 8bit, noisy, half the frequency response of human hearing, and out of tune.

I’m happy to take up to 10ms latency to not have these problems, plus have a massive DSP chain...

You mean WaveOut ? This is what i used in my audio test.
But, DirectSound is obsolete. I tried to check others, like WASAPI, but it does not look simple and i couldn't find usable sample code.

Actually that’s a good question, the last time I coded audio for Windows was over 15 years ago... WaveOut rings a bell, but we used ASIO... I don’t know what is common now. I moved to the Mac platform around 2006, so CoreAudio has been my focus.

Yes, read Toni's posts. If you believe he confirms your claim, quote him here (or link to the post itself).
And looking at your tracker code, what do i see ? 44100.
Once and for all, Amiga sounds is 8-bit d/a, not 1-bit d/a. Even Thomas agreed with that !

Hmmm, I’ve just read the thread again, it seems that no one is particularly sure what the DAC architecture is. I would add that Thomas correctly stated that it doesn’t mater if you have a high frequency PWM or an 8bit resistor ladder DAC, the output will be the same... Though I would expect the PWM to be more accurate.

The most telling post for me is:

http://eab.abime.net/showpost.php?p=845750&postcount=37

Of course yes, the tracker code also runs in an interrupt !
Listen, pal, i've written enough players for Delitracker to know how sound is made, and you're far away from it.

I feel you are being a bit hostile here, but I understand your frustration. I would be better to say I would never run tracker code in an Interrupt. My approach is to do a little in an Interrupt as possible.

But this is just A500. It's night and day in comparison to expanded A1200.

I have several A1200s, one (which used to be my main machine) has a BlizzPPC 250Mhz and BVision in it. But that was overkill just to run a tracker, and far too noisy for my studio.

This is what Thomas wrongly referred to as Fanboism...

Frankly, if i were the fan boy you both mention, why the heck would i be currently porting my system framework to Windows ? Are you conscious this is complete nonsense ?

Ok, you sound like a fanboy, but I’m happy to accept that might be unfair on my part.

I will gladly defend the Amiga for its strengths, but I feel overstating the capabilities and misrepresenting the machine is actually insulting. Worse still, you don’t need to claim other systems are worse, than the Amiga, as that is simply intellectually dishonest, and embarrassing for those of us do accept the Amiga’s shortcomings.