@Gunnar

You may have misunderstood what I was trying to say. So to rephrase it:


A fully featured tracker, like OctaMED SoundStudio, capable of directly controlling all 16 of your 16 Paula channels while simultaneously controlling a bunch of external MIDI hardware, would be hard to beat. If the 16 channels were also separately available as mono analogue outputs for routing through external gear, you'd go from hard to beat to virtually impossible to beat.

If your SAGA Paula channels have all the same audio quirks (nonlinear DAC, aliasing, 3.5MHz PWM digital volume, nonlinear frequency response etc) as the original Paula when playing 8-bit samples, you go from virtually impossible to beat to an electronic musician's wet dream.

Add programmable resonant filters to each channel and you go from musician's wet dream to so good it's declared criminal and subject to a UN resolution to block proliferation.


Hopefully this clarifies it for you.

Last edited by Karlos on 29-Feb-2024 at 12:28 PM.
Last edited by Karlos on 29-Feb-2024 at 12:26 PM.

_________________
Doing stupid things for fun...

@Karlos

Quote:

A fully featured tracker, like OctaMED SoundStudio, capable of directly controlling all 16 of your 16 Paula channels while simultaneously

OK thanks, I think now understand what you asked for.


Milkytracker does supports the 16 Amiga DMA audio HW channel directly.
With Milky you can make Amiga Mods using the HW - today.
But the combination with extra Midi channels - I think there is no option for this.

@Gunnar

In fact, if you want inspiration for a spin-off product:

Put such a Paula implementation into a standalone unit, fully controllable via MIDI (5-pin DIN and USB), with SD storage on board for saving sample data.

A standard rack or desktop module with multiple channels of freely pannable Paula on the main audio bus, with individual output sockets on the back for external processing and programmable filters (even if implemented digitally before the DAC) would be a hit.

What musicians look for in audio hardware behaviour is rarely the same is what consumers look for. For games, 16 channels of 16 bit pannable sound is clearly good for sound effects. It's also good for basic modules of course. There is a point though, when less is more - a lot of the charm of Amiga modules was the 4 channel limitation and the creativity that resulted in.

Once you progress beyond basic modules as the music distribution format, where you want to master a track for actual distribution as a stream format, you may go right back to basics, using Paula like any other electronic instrument in your studio.

Last edited by Karlos on 29-Feb-2024 at 12:43 PM.

_________________
Doing stupid things for fun...

@Karlos

Quote:

Karlos wrote: @Gunnar In fact, if you want inspiration for a spin-off product: Put such a Paula implementation into a standalone unit, fully controllable via MIDI (5-pin DIN and USB), with SD storage on board for saving sample data.

100%

Side note: My guys at semblie.com in Bosnia can produce a small batch to test the market.

_________________
All the way, with 68k

Is it done yet? :)

_________________
Doing stupid things for fun...

@Hammer

Quote:

A1200's laptop 2.5-inch hard disk wasn't cost-competitive.

Against what? It used common PC laptop drives. However, as I've pointed out, there usually was an Amiga tax on common Amiga targeted HDD.

But the IDE against the bigger SCSI drives was interesting. As some SCSI drives were better value which is strange as IDE was usually cheaper and more common. However it is comparing laptop against desktop sizes. But what I wonder is how the laptop IDE prices compared to usual PC shop prices?

@cdimauro

Quote:

I agree on this. In fact I had to fit a 3.5" hard disk (several of them were suitable: both fitting and without requiring a new power supply) on my A1200. And like me did A LOT of people that didn't want to pay almost DOUBLE the price for the stupid decision of Commodore engineers.

Almost double for a desktop IDE drive against laptop IDE? Although IDE was clearly inferior to SCSI, especially before UDMA arrived, I think it was a good idea to include on board. As it allowed commodity drives to be plugged in internally direct on board. It was limited to PIO 0 though.

I didn't see 2.5 inch as an issue in the day. It may have been because drives came down by the time I bought my A1200. It came with a 40MB drive and I bought a 500MB Quantum shortly after. I can understand the size as it has limited space. But, after the clunky A590 with 20MB that cost a lot more, I thought an internal laptop IDE was neat!

In retrospect, perhaps a 3.5 inch desktop drive would have been better in the A1200 and 2.5 inch laptop drive on the, ahem, A600. However, the A4000 did have the full desktop case, while the A1200 was more compact. It could be hacked in. But my A1200 was made of rare precious metal so I decided not to do that.

@Hypex

My friends A1200 had a underbelly exposed 3.5" HD balanced on the shielding for years with the case gaping open! His Dad ran power from the Floppy lines and hacked together a 40pin to 44pin adapter! It worked stably for decades but it made the case unsightly! It was also noisy but cheap and an off the shelf part!

_________________
"Art challenges technology. Technology inspires the art."
John Lasseter, Co-Founder of Pixar Animation Studios

@Hypex

Quote:

Hypex wrote: @cdimauro Quote: I agree on this. In fact I had to fit a 3.5" hard disk (several of them were suitable: both fitting and without requiring a new power supply) on my A1200. And like me did A LOT of people that didn't want to pay almost DOUBLE the price for the stupid decision of Commodore engineers. Almost double for a desktop IDE drive against laptop IDE?

Yes.
Quote:

Although IDE was clearly inferior to SCSI, especially before UDMA arrived, I think it was a good idea to include on board. As it allowed commodity drives to be plugged in internally direct on board.

Absolutely. "Computer for the masses, not for the classes".
Quote:

It was limited to PIO 0 though.

Eh! Wasn't the Amiga the land of DMAs everywhere? Who created the EIDE interface seems to have forgotten it...
Quote:

I didn't see 2.5 inch as an issue in the day. It may have been because drives came down by the time I bought my A1200. It came with a 40MB drive and I bought a 500MB Quantum shortly after. I can understand the size as it has limited space.

Well, the problem was the cost / size. That's the exactly the reason why many people mounted a 3.5" hard drive on their A1200s, so making it competitive with PCs. BTW, I've bought a 540MB Quantum short after the 1200, because a HD was really needed at the time, even for Amigas.
Quote:

But, after the clunky A590 with 20MB that cost a lot more, I thought an internal laptop IDE was neat!

Absolutely! I really enjoyed it: completely changed the way to use a computer. Floppies? Stone Age technology...
Quote:

In retrospect, perhaps a 3.5 inch desktop drive would have been better in the A1200 and 2.5 inch laptop drive on the, ahem, A600. However, the A4000 did have the full desktop case, while the A1200 was more compact. It could be hacked in. But my A1200 was made of rare precious metal so I decided not to do that.

Up to you. The A1200 had already enough space for fitting plenty of 3.5" which had a power consumption compatible with its power supply. It was nice, really!

@Hypex

Quote:

Hypex wrote: @Hammer Quote: A1200's laptop 2.5-inch hard disk wasn't cost-competitive. Against what? It used common PC laptop drives. However, as I've pointed out, there usually was an Amiga tax on common Amiga targeted HDD. But the IDE against the bigger SCSI drives was interesting. As some SCSI drives were better value which is strange as IDE was usually cheaper and more common. However it is comparing laptop against desktop sizes. But what I wonder is how the laptop IDE prices compared to usual PC shop prices?

For MB per dollar, the common laptop IDE drives have a higher cost when compared to desktop IDE drives.

Last edited by Hammer on 05-Mar-2024 at 05:33 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@Hammer

Quote:

With AmigaOS 4.1 FE with RTG, I run Deluxe Music 2 on the Workbench screen. The problem is when the play button is clicked.

That's what gave me an issue running it normally on a custom screen. Or gave me a freeze. I was running it off a mounted ADF. If that makes any difference. So not off HDD.

Quote:

Fully functional emulated Paula on AHI would be nice. I assume CIA functions are fully functional on AmigaOS 4.1 FE/Amiga NG.

Well would be but an audio.device routing itself through AHI was what we got. CIA would be fully functional on classic version. For NG it's removed and last I checked left mouse emulation was also removed.

Quote:

With RTG on fast 68K or PPC, I usually run my Amiga apps on Workbench.

Given screen dragging doesn't work as good as it used too, with screens behind missing last I used it, RTG to WB looks like a good idea.

@Gunnar

Quote:

16 Amiga/Paula DMA audio channels Each of them can be selected to be stereo positioned controlled "pan" with two volume to be between Left/RIght Each channel can play has either 8bit, 16bit, or 2x16bit stereo samples. Each has own volume own sample rate.

Is there an AHI driver yet? The Paula DMA driver is looking a bit old. Arne is looking more ripped.

@Hypex

Yes, there is AHI, much to Gunnar’s dismay, as he prefers asm coders that bang the hardware directly.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@Hypex

Quote:

Is there an AHI driver yet?

Yes, of course. We provide AHI drivers.

@Hypex

Quote:

Well would be but an audio.device routing itself through AHI was what we got. CIA would be fully functional on classic version. For NG it's removed and last I checked left mouse emulation was also removed.

AmigaOS 4 has 3rd party CIAgent software which is a CIA emulator. http://os4depot.net/index.php?function=showfile&file=emulation/misc/ciagent.lha
This function should be included out of the box.

Draco has two CIA chips.

Amithlon has emulated CIA functions.

PowerPC 60x on classic Amigas are dead-end route since Phase 5's PowerPC accelerators are not being manufactured nor is it open-sourced for uncontrolled clones. The strength of PiStorm and PiStorm32 Lite is the clone hardware and low-cost PRi 3A/3B/4B/CM4 via the economies of scale.

Phase 5 didn't design a generic barebone PPC small board module like RPi and a separate gateway board specific to an Amiga model. The generic PPC small board module should be able to operate on its own via Linux.

RPi didn't copy PC's ATX standard and innovate its barebone and low-cost standard.

Last edited by Hammer on 06-Mar-2024 at 01:56 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@Gunnar

Quote:

Gunnar wrote: @Karlos Quote: A MIDI capable sequencer application that can play the 16 channel while simultaneously controlling said outboard hardware would be a very compelling system. Don't you have all this on Amiga already? The Amiga program "HORNY" can do "infinite" number of Midi channels Is this not perfect for this?

The "HORNY" name for a midi sequencer is not good. Type in "Amiga HORNY" on Google search and it returns major porn sites.

Amiga is a girlfriend in Spanish.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@cdimauro

Quote:

Not yet. If you use ffmpeg to encode a video with the latest HEVC/H.265 codec you can see from its output console that it's still using the MMX extension. Just to give a notable example.

That's interesting being the latest. I would have expected it to use the latest extensions. It then doesn't need too and perhaps MMX could be better suited though I don't know that.

Quote:

x87 is also used in several softwares. You can also use in WinUAE, for example.

x87 is almost as entrenched as x86 itself so it would be hard to simply displace it. Sure, standards and CPU designs change, but x86[-x64] usually implies x87 support as well. Well a modern variant of x87 now days.

Quote:

Indeed. They only doubled the number of "vector" registers with VSX, by reusing the floating point ones and extending them to 128 bits.

Oh not. Not reusing FPU. Well, at least it's not an SPE.

From a technical point of view, PPC or POWER, does differ from AMD64 by having at least 32 registers for general, FPU and vector. So, it does hedge its bets, so to speak, on a maximal register count. 32x128 bits makes 4096 bits. 16x256 bits matches that. So, in full width alone, PPC was ahead until AVX256 matched it.

AVX512 surpasses that. 16x512 bits is 8192 bits. A VSX with 64x128 bits would be able to match that with 8192 bits again. Unfortunately, reusing the FPU registers, just looks like an old x86 FPU hack. Looking desperate there.

Now, any full width VMX or VSX can do that matches or beat x86 is some areas, is rather moot if it's unable to work on that desired width divided into smaller vectors using one operation. Of course, RISC does need to split operations into greater codes. But doing what looks like loop unrolling one one CPU when another can do it all in one, won't look as efficient.

Quote:

But there's a new vector ISA proposal, Simple-V: https://libre-soc.org/openpower/sv/executive_summary/

Oh no. More extensions.

A main issue is compatibility. It looks like they ensured that. That's good for POWER but I won't expect it to be in a new AmigaOne any time soon.

Quote:

Intel is still at 512 bits (AVX-512). The prefix encoding allows to specify another vector length, so 1024 is definitely possible, but if Intel's engineers aren't stupid they should used for a vector extension instead of another packed/SIMD one. The actual packed/SIMD can be left as it is.

More complexity. Not that it matters any more. As over the next few years they will manage to advance even more features. So 16384 here we may come.

@agami

Quote:

I purchased my A1200HD from Maxwell's in Abbotsford in December of 1992 for $1,199 AUD. What I was able to do with that machine during 1993 and well into 1994, other systems could only do at a higher price point, for both hardware and software.

I also bought my A1200HD from them. With a 40MB HDD. But, I bought a demo model, that I recall only cost me $900.

However it must have been late and closer to 1995. I recall a family member buying an early Windows 95 machine and told I bought the wrong one. It was nice to see all the games in 3d, 256 colours, and see the future of Road Rash. But, I didn't want to buy a games machine, that would waste my time. I wanted to buy an Amiga and continue using apps I had been using and projects I wax working on. Ironically, I was told I bought a games machine. An Amiga as a games machine in 1995 is ridiculous, since the PC had taken over by then. I never saw Road Rash 3d on the Amiga or Test Drive 3. I never saw the Breed on PC either but no one missed it.

@Karlos

Quote:

A fully featured tracker, like OctaMED SoundStudio, capable of directly controlling all 16 of your 16 Paula channels while simultaneously controlling a bunch of external MIDI hardware, would be hard to beat. If the 16 channels were also separately available as mono analogue outputs for routing through external gear, you'd go from hard to beat to virtually impossible to beat.

In this case it would be another tracker. Since MED does all the mixing itself and then outputs it as a single audio stream. The only way it uses hardware is directly programming Paula. This may change in the upcoming version which can use AHI. But, it would need options for using AHI to mix it. If it has an option to use AHI module features, which is the proper way even if it locks out other access, then it can make use of AHI mixing. With an AHI hardware mixer driver it could then drive a 16 channel Paula.

However, if not, then a substitute MIDI device would be interesting. That assigned 16 MIDI channels to hardware channels. Working as a sampler.

Quote:

If your SAGA Paula channels have all the same audio quirks (nonlinear DAC, aliasing, 3.5MHz PWM digital volume, nonlinear frequency response etc) as the original Paula when playing 8-bit samples, you go from virtually impossible to beat to an electronic musician's wet dream. Add programmable resonant filters to each channel and you

IIRC it uses a 24 bit digital mixer so may not replicate the Paula crunch exactly. This has also been levelled at UAE where it usually uses a simpler mixer over accurately simulating Paula. Given Paula would be a mlx of digital and analogue components output through an analogue line output, and "Arne" is implemented in an FPGA, there's only so much you can do to get close to the original sound. But Gunnar can fill us in. I heard Test Drive on a Vampire a few years back and the siren wasn't right but it may be resolved by now.

Last edited by Hypex on 10-Mar-2024 at 03:43 PM.

@Hypex

This is why I said "like OctaMED". Obviously not OctaMED itself, since obviously it doesn't. But it's a moot point. Musically, we want Paula's characteristic sound and quirks. Without those, well, I can just use a DAW or even a hardware sampler.

_________________
Doing stupid things for fun...