What was wrong with the 464?

[sigh]

I didn't like the "one" joystick port and the keyboard clashing. Playing 2 player games whether it be VS or co-op could turn into frustration very easily. I enjoyed BMX simulator 2 with it's 4 player mode as well as others like Jet Ski simulator, but having 3 guys cramped around around a keyboard, interfering with each others controls, ruined the experience.

I also agree with Macdeath in that it should of shipped with 128Kb from the beginning, whether it be a single chip or 2 separate 64kb chips.

I also felt that some form of built in assembler on the ROM chip would of been nice along with the built in basic.

[MacDeath]

But ok, 24k of vram

CPC can do that to, except you'll have 192x256xmode0 full screen instead of 256x192x16 at 6mhz.

CPC is actually somewhat faster in 128x192xmode0 because 4mhz instead of 6mhz but 50% less video.

Had SAM coupé the same modes as CPC it could then do awesome prowness the CPc couldn't... but no it was to chose between speccy or Atari ST...

would have been fun to be able to port CPC games the way they did on enterprise... but at 6Mhz..

[ralferoo]

Most of the "mistakes" are purely down to the speed at which the CPC was designed, and significantly the rushed change from 6502 to Z80. If they'd done the IO addressing more intelligently (even just swapping high and low address bytes, peripherals could decode the high bits if they wanted) and fixed the pointless missing wait state (rather than regularly every 4th cycle, just WAIT' = every 2nd NOR (RD' NAND WR') would have given the CPU 12% more speed on average). But if the machine worked as it was and it was needed immediately, then it's easier not to fiddle with it...

The screen layout was clearly a conscious decision because it makes it possible to scroll text by character line very quickly. The BBC layout would probably have been slightly more beneficial but it makes graphics drawing a bit less nice. And perhaps that was all designed when the CPC was still 6502 based and they didn't want to appear totally the same as the BBC!

I guess I'm more surprised how popular the CPC became as a games machine, as it was clearly targeting the business/home office user with the monitor which lead to an expensive machine for games use (but of course, more total profit for Sugar!)...

[chinnyhill10]

I also agree with Macdeath in that it should of shipped with 128Kb from the beginning, whether it be a single chip or 2 separate 64kb chips.

I don't get what part of "128k in 1983 was expensive" people don't understand. Sugar wanted to hit a price point. We're not talking a few pence difference but doubling the cost of one of the most expensive components in the machine.


Put it this way, a modern computer with 16GB of RAM costs more than a computer with 8GB of RAM.

[chinnyhill10]

The screen layout was clearly a conscious decision because it makes it possible to scroll text by character line very quickly. The BBC layout would probably have been slightly more beneficial but it makes graphics drawing a bit less nice. And perhaps that was all designed when the CPC was still 6502 based and they didn't want to appear totally the same as the BBC!

I think Roland Perry and co started pretty much from scratch with their design. They were only tied down to the casing and the tape drive. Components that had already been purchased were sold on. The Amstrad Story covers most of this as does the excellent article on The Register a few months back.

[AMSDOS]

I converted this program which relied heavily on BASIC FILL commands, though after converting the program, it wasn't lighting up the keyboard keys which seemed to be the problem.
However, in BASIC 1.0 there was a lag in the sound, maybe because of the extra tasks I was performing to light of the keys perhaps, though my workaround functioned no differently in BASIC 1.1 compared to the original program.
My gut in this case tells me that it's because of the number of Array's in the program and having to setup another one with the appropriate information to light up the appropriate key.
It seems better to clear the array's in BASIC 1.0 because it tends to put a lag on things I've noticed, especially if it's BASIC 1.0 with AMSDOS and one wants to CATalog.

[TFM]

I also agree with Macdeath in that it should of shipped with 128Kb from the beginning, whether it be a single chip or 2 separate 64kb chips.

Loading 128 KB from tape... How long will that take? 

[Novabug]

Smooth scrolling (like the C64) to prevent flick screen, and a more reliable modulator. Other than that, I would change much at all. Speecy ports were not the fault of Amstrad themselves, the pubs put pressure on devs to meet a release date, so they often made shortcuts. Same happened with the GX/Plus cart games (Well, half of them) Shovelware from the CPC due to pressure and, more than likely laziness. At least we got Pang and Robocop 2.

[TFM]

If they'd done the IO addressing more intelligently (even just swapping high and low address bytes, peripherals could decode the high bits if they wanted)

Actually... that's right! On the other hand for a long time there was no external hardware which would have benifited from block I/O operations. Correct me if I'm wrong.

...and fixed the pointless missing wait state (rather than regularly every 4th cycle, just WAIT' = every 2nd NOR (RD' NAND WR') would have given the CPU 12% more speed on average).

That's shocking news!    At least we can calculate in us on CPC. 

[TFM]

Smooth scrolling (like the C64) to prevent flick screen

The CPC can very easy scroll: Up and down pixel-precise (see Mission Genozide) and in X it's two-Mode-0 pixel-precise (Relentless f.e.). And it's easy to code.


If programmers use speccy ports and are to lazy to change scrolling, don't blame the CPC. 

[Zoe Robinson]

Loading 128 KB from tape... How long will that take? 

OMG, I'm having flashbacks to the time 3D Construction Kit had a read error on the last few blocks to load and I had to start again. 

[MacDeath]

weren't most CPC tapes using Spectrum loading routines ?

[AMSDOS]

The AA52 Covertape was a Nightmare because The Untouchables had it's music in the extra 64k and it was loading in a Standard SPEED WRITE 0.


At the time I thought nothing would take longer to load than Roland In Time. 

[Sykobee (Briggsy)]

* The low quality built-in speaker.   * the memory layout of the screen - every 8-lines   * the hardwarescrolling could have been better   * hardwaresprites
* no fast charactermode   * no low-memory screen-modes   * not an option to buy it cheap without a monitor   * the external modulator was low quality   
* quick and dirty spectrumports   * the high sharpedged keyboard

   
• Well the CPC was produced to a price point TBH. The speaker should have been in the monitor, in stereo, as with the pluses.
• The screen memory layout was a side effect of using a cheap late 70's CRTC controller chip modified to do bitmap graphics in a clever, but esoteric manner.
• Yes, if the gate array had used a slightly bigger GAL/PAL thing, then smoother horizontal scrolling (using the scanline delay method, as used by some other 8-bit micros in conjunction with using the border to hide the edges of the delay) could have been achieved easily.
• Hardware sprites were not an option with the initial design, not without custom hardware.
• This would have required extra work to cope with switching the CRTC from the hacked up bitmap mode into a real character mode. I also don't think you'd have got a fast 80 column mode out of it, as character modes require two reads per character per line, and the CPC had 128 bytes per scanline.
• Again, a limitation of the CRTC and how it was clocked. A half-clock option would have enabled a 64-byte scanline mode with 8KB screen I guess - 2 colour MODE 1 and 4 colour MODE 0 in effect.
• At least CPC owners could use it whilst the main TV in the house was being watched by someone else. But yes, I agree, an option for houses with second TVs. But that modulator ...
• Can't do much about this sadly, lazy programming is lazy programming.
• I liked the keyboard.

More colours - again, a bigger GAL/PAL would have allowed this. Still, they found the space to put in colour swapping hardware :S I'd have settled for 64-colours, 2-bits per R,G,B (the PC1512 did get this palette option I think). Or maybe the Enterprise 64's 256 colours - RGB332.

An 80KB CPC480 would have been neat, that extra 16KB could have allowed more games to have free double buffering and large graphic sets, for example.

Like most 8-bitters, the CPC was a collection of compromises designed to meet a price point. It did well, and it had a real personality.

[Sykobee (Briggsy)]

And vomit colours.  And colours per sprite limits.

The sad thing is that the C64 reads 4 bits of colour ram per character, and uses it to change 1 colour in that character.
Imagine if that had been used to select 1 of 16 four-colour palettes for that character. Alongside a bit more flexibility on colour range.


If you look at an annotated VIC-2 die shot, you can see why. It's 80% sprite hardware and sprite buffers and character/colour buffers.

[AMSDOS]

An 80KB CPC480 would have been neat, that extra 16KB could have allowed more games to have free double buffering and large graphic sets, for example.

Would a CPC472 be better to some extent?

[Sykobee (Briggsy)]

Would a CPC472 be better to some extent?

TBH given the paging system the CPC has is 16KB in size, I doubt it. Even if the RAM was actually wired in.
With hindsight you can see that RAM prices drop in 1984 to be a lot cheaper, and thus an 80KB CPC would have been viable by the time it was available. Hence the 6128 in '85, the 8256, 1512, etc.


So in terms of this topic - the 464's design should have had passive 80/128KB ram capacity designed in in case of a RAM price drop.


And Amstrad should have released the Plus range a year earlier, and based it around the PC1512/PC1640 hardware or similar, with a custom ASIC for the colours and sound, and stuck in the CPC+ casing.

[sigh]

Loading 128 KB from tape... How long will that take? 

30 seconds

My 128kb (Datel 64kb expansion) 464 loaded most games into memory barring a few, but I cant remember how long it takes.

So in terms of this topic - the 464's design should have had passive 80/128KB ram capacity designed in in case of a RAM price drop.

Apart from the MSX 1 released in 1983 in which some came shipped with 128kb, where there any other 128kb home computing machines, before 1985?

[TFM]

Would a CPC472 be better to some extent?

Sure if the RAM would be accessible ;-)

[CraigsBar]

Sure if the RAM would be accessible ;-)

or even connected to, let me think, anything

[MacDeath]

Apart from the MSX 1 released in 1983 in which some came shipped with 128kb, where there any other 128kb home computing machines, before 1985?

Sinclair QL, Machintosh128...
Really... CPC480 would have changed a real lot of things on the paper.
A CPC6160 as well.

[AMSDOS]

Sure if the RAM would be accessible ;-)

So does that mean it doesn't have 72Kb, or it does, but it needs a custom program to access the 8Kb. I know someone had to do this to access the 8Kb in a Multiface 2, it was interesting, though obviously more involved compared to punching out some OUT statements.

[AMSDOS]

Sinclair QL, Machintosh128...
Really... CPC480 would have changed a real lot of things on the paper.
A CPC6160 as well.

Useborne Guide To Expanding Your Micro which was published in 1984 has this:


Apple IIC had 128k
Apple IIE 64K Standard Expandable to 128k
Memotech MTX-500 32k Standard Expandable to 512k


A Sinclair QL could be upgraded to 640k


Later Apple IIE machines came with 128k as standard, not sure when this happened though.

[TotO]

So does that mean it doesn't have 72Kb, or it does, but it needs a custom program to access the 8Kb.

The inside of the CPC-472 was identical to the original CPC-464 apart from a small add-on card fixed to the original ROM socket and supporting the original ROM, a 4164 RAM chip and a standard logic gates chip. The RAM and logic chips weren't conected at all. Printed circuit lines didn't reach them! This add-on board was only a fake to avoid taxes.

[Gryzor]

I quite liked the keyboard; it was ace for its ear. Sure, a bit high, but blew many other machines out of the water...

As for memory, of course, the price point has been mentioned many times in this thread and is pretty simple to grasp. AMSTRAD did know RAM prices were going to drop, hence their product plan for a 664 and a 6128. They wanted the 464 to be the cheapest solution available - instead of adding more RAM when its price fell they preferred to lower the price and stick with their market entry point.

The speaker was tiny, of course, but when it was being designed it was good enough. Instead of a higher-spec speaker I'd have preferred dual speakers (the first time I heard the machine through headphones -Solomon's Key, I think, or Ramparts- I was blown away. Stereo!!!).

As for the modulator: it wasn't so bad back then, compared to the other computers. And don't forget you could get a cheap SCART cable and have perfect picture (no sound of course)!