andym00

Awwww, so all my work over the last 2 weeks adding true audio rate FM to SID voices was in vain ? Do you 100% stand by that comment ? If you want to really hear what 2 operator FM stuff can do, go find Oxygenstar on 8bit Collective.. His FM stuff is astonishing, and I mean that from the bottom of my heart.. But what blows my mind totally is that in tunes that embraces the harshness of the YM stuff, and use it like and instrument, instead of trying to be something that it isn't and will never be.. SID's guilty of that as well Some tunes I've heard recently, mixing FM voices, raw square waves, and rough samples are truly amazing.. It's just a shame these have to be produced on tools that are never destined to run on the native hardware, I mean as a stand alone piece of executable music.. People like: Check out stuff like this.. OxygenStar - Look How I Go OxygenStar - Left Out And utilising lovely square wave beauty tat it can be at times.. trash80 - icarus trash80 - missing_you Plus, Mr.Spastic.. I hope there's a very good reason we're not hearing stuff like this from the A8 In C64 land there seems to be some resentment towards lovely square waves And yes, I like Oxygenstar because of his penchant for simple FM synths But how he uses them is mind-boggling.. In the early 90s I spent far too much time writing FM drivers, and the stuff they get out now is fecking amazing to say the least.. But overall I just love that way simple sounds are used to make amazing tunes, rather than trying to bend those basic sounds into something they're not... And yes (according to Atariksi) I'm c64 nut, but SID I find too familiar in that it reminds me too much of the original (subtractive) synths back in the early 80s that I learnt synth programming on but with not enough ooomph to it.. I don't want that Anyway, I'm in the bar and shouldn't be in control of a keyboard connected to the internet right now, so forgive me.. Just must say that FM is horribly underrated, and that (square waves and lots of them!) are incredibly powerful.. Just go look up Walsh Synthesis if you don't believe me

Found this on the 6502 forums where coincidently that exact problem was being discussed.. Last post in the thread.. http://forum.6502.org/viewtopic.php?t=225

I've got a bit lost in this thread with all the Pokey talk and timers, but I thought I'd mention this that I was just reading.. http://members.lycos.co.uk/leeedavison/6502/vic20/isa/index.html It's not of direct relevance really, but it is interesting in that he states that the 6502 in the Vic he was using was missing NMIs, and goes on to explain why.. Just seemed relevant and I didn't realise this was the case, and hadn't seen this mentioned in here.. I'll go away now It is a good read though, hooking up ISA cards to Vic-20s

But wouldn't you need to force a badline to make VIC read the new character line ? And with (probably) 8 sprites eating away 19-21 cycles per lines and you forcing badlines everywhere plus changing screens (and potentially character sets) and setting x-positions I think you're going to run out of cycles pretty sharp like.. I know you can do smaller stuff easily enough by changing character sets, and maybe the way you say would work, but in a game ? Forcing the badlines is going to wreak havoc with your screen as well since it'll scroll the screen up as well when you force the new badline ? No ? I've generally stayed well clear of all this $D011 insanity such as line-crunch etc, so maybe I'm missing something.. Brain is now being consumed by this.. Must work!! Must not think about SFII floors

I thought as much That's the one particular feature of the A8 I'm really envious of, the ability to change the screen address on a scanline basis.. I would have killed for that on the 64.. Damn it, I'm supposed to be working but instead now have a head full of Street Fighter 2 and 6502 code thoughts

Regards Street Fighter.. On the A8, doing the parallax floor bit, ie changing the scroll on every line to somewhere over a 2 screen wide area, I assume that's very easily possible ? Just changing the scroll registers each line I mean ? I do like those kind of floors and it's always bugged me that on the 64 that it's a bloody nightmare to do that kind of thing without burning through all the sprites, so much so that I've never seen anything beyond a space harrier like floor.. I know there's lot of parallax demos and things, but I mean properly in the style of SFII with a true background image, and changing the scrolls every raster-line.. On the A8 is it really just as simple as changing x-scroll on each line and having a wider screen ? Or are there technical reasons why it's not viable? Any demos that show this kind of thing off ? And also, is there any reason why when jumping and the screen scrolls it wouldn't be possible to introduce a height perspective effect in this, by changing screen addresses on each raster line to get the shrinking, stretching of the floor ? Sorry if that's a dumb question, but I've only a fleeting idea of the A8s hardware, and from what I've read and heard, I'm guessing it's pretty easily doable ? Just curious, as after looking at the street fighter on the Amstrad for the first time in ages now my brain is full of thoughts of SFII with big sprites, and that lovely floor

Bugger.. My warm soft spot for it has just dwindled into a damp ember now..Okay, so maybe it is a bit naff then

Whilst I'm not mad about this machine (the colour saturation was right off the scale...great if you wanted to do pictures of nuclear reactor fuel!) I think it was also not coded for properly, most of their games were straight ports from the Timex/Sinclair source code as they shared the Z80. One example is it has a near flawless conversion of Donkey Kong down to the pixel level, but still without decent hardware scrolling/sprites etc it was always tough to do good arcade games on beyond this early 80s era of things like Pacman and Donkey Kong etc in my opinion. When people get to grips with the video controller some pretty cool stuff is possible on it..Prehistoric II is a prime example of what it really can do when in the right hands I've still no idea if this is real or not, but if it is then it dumps on most of the other home conversion from a great height.. It's a vastly under-estimated machine I think.. Of course there's a huge catalogue of duffers for it, but Prehistoric II and Street Fighter II both made me realise that it's a quite potent little machine.. It shares the 6845 VDC (along with the BBC and all PC video cards through from MDA to VGA/MCGA) and can be prodded into doing some very cool tricks, included in which is hardware scrolling and all manner of other 'odd' things If that SF-II is real then my soft spot for this weird little machine just got bigger

With the 64 the only condition I'm aware of that can cause dropped interrupts is if you read the interrupt control register (I think 2 cycles) before an IRQ is triggered which can cause it not to raise a CIA interrupt, be that NMI or IRQ depending upon which CIA you're polling, but this is CIA bug not '6502' bug.. So basically if you ACK the interrupt 2 cycles before it happens, it won't happen from the point of view of triggering an interrupt.. Wish I had something useful to contribute to this thread, but it's been fascinating to watch you guys doing your Sherlock Holmes impressions this weekend I'm dying to know what is is that causes this now, and good luck

This should answer all of your questions http://www.gorenfeld.net/lou/pseudo/

That's great news! I'm dead curious how it fairs against NUFLI stuff with these additions I've been working on some stuff to do this kind of conversion with it in mind to process blocks of bitmap and sprites in double width sprite chunks (for a game level), with variable colour split sizes down to 8 lines for the multicolours and sprite colours, but it looks properly shit in comparison to what MUCSU turns out

You got a link to that ? I'm having a hard time with google search results for searches along the lines of Slot Machine Atari, not surprisingly edit: All I can find are these two, and I'm guessing they're not what you meant, given the mention of 16 Lumas ?

Actually it's only ASL/LSR/ROL/ROR abs,X that take 6 cycles instead of 7, unless crossing a page.. INC/DEC abs,X still always take 7 cycles on the 65c02, same as 6502, regardless of whether it crosses a page boundary or not..

Another source to consider would be the NES version that surfaced a few years ago, interesting given it's very weak sprite abilities and not being so hot in the bitmap manipulation stakes.. An interesting solution, and Andrew Davie in that thread also comments on how they handled the graphics, by splitting charsets on each line, and the charset containing everything that's need for that particular character line of the fighter.. Anyway, food for thought.. Here's the thread.. http://www.atariage.com/forums/index.php?showtopic=87396 And here's the youtube video of it in action.. Looks mightly impressive for the NES actually..

Awww, but Emkays just offered to write Way of the Exploding Fist

Ummm.... But glad you nudged this thread up, I was looking for this particular one about software sprites but wasn't getting anywhere finding it..

Works okay in WinVice here but does flicker, though the flickering is never going to vanish, unless the emulation is properly synced to the monitor display rate.. This kind of stuff just doesn't look good on emulators at all unless you're at 60hz and running something NTSC based.. Unfortunely this isn't NTSC compatible and hadn't been NTSC fixed yet, if indeed there's even enough time left over from the display to run the music and scroller and handle the loading..But you knew this already of course

Make sure 'True Drive Emulation' is enabled..

Erm, it's the main scroller in the slide show that you & Rybags are talking about ? If so, that one's in the bottom border, so it's just $3fff (or whatever is appropriate) being set each scanline.. The last part scroller, looks to me like just inverted text being scroller over in the center some x-expanded sprites, and then on the left and right edges, 2 x-expanded sprites overlaying on top to maskover the scroller, with maybe some colour ram shenanigans going on as well.. The last part does work on vice, just seems to take a while to sort itself out for some reason... There's a lovely loading screen if you hit space to skip the pictures As regards SID differences, I know of the differences in capacitors, but the frequency response of the 8580 is markedly different to that of the 6581, I mean way off, as is the resonance.. Of course there's differences between the same models as well, but that's not really stopped people using the filters I'm just saying that the real difference is between the 6581 and 8580, and not just regards sample playback, which is easily cured by ensuring DC offsets are up in the chip close to something resembling the 6581.. Plus there's the differences in combined waveforms as well

You can't but you can detect which SID revision you're running on very easily buy setting oscillator 3 to its maximum frequency.. setting the testbit to lock the oscillator, then select sawtooth and gate off and then read immediately with the next instruction (lda abs) from the oscillator output.. They start at slightly different sample values, 3 on the 6581, 2 on the 8580.. Easy peasy

Highest priority sprite is 0 Lowest is 7.. So you'd want to put sprite 0 behind the background, sprite 1 in front, and (if memory serves me right) it'll display foreground data where sprite 1 overlaps sprite 0.. If I recall right it resolves sprites to sprite priority before sprite to foreground priority.. It's decided to draw the masked sprite, but then oops, it's behind the foreground, ergo (it appears) to cut a hole in sprite Logically, the hardware simply takes the pixel for the first sprite it finds that has NZ data regardless of the foreground/background setting then resolves it against the screen pixel based upon the foreground/background setting for that sprite.. It's a standard way of doing things on a fair few consoles for exactly this type of effect.. I'll give it a go when I get home.. I just grabbed winvice, and trying the keyboard on my laptop is acting very weird when typing in WinVice..

Okay, apologies for bringing this up again, but it's just been pointed out to me, that using the timers means you lose audio voices ? You kept that very very quiet in your sales pitch.. You want a 15Khz timer ? You lose one voice.. You want your 1.79Mhz timer ? You lose two voices.. Hmmm, interesting.. So you can have your cycle-exact interrupts, as long as.. 1. You're in bitmap mode, or have rolled your own bad-line free character mode.. 2. You're prepared to lose 2 voices.. 3. You WSYNC every line 4. (IF) you use DLIs, they use the same cycles each time.. OR You do exactly as the c64 does and de-jitter the interrupt manually and regain the use of the CPU during the frame when not servicing interrupts.. Suddenly life doesn't seem so bad on the 64 with regards to cycle-exact interrupts, in comparison to the trade-offs you glossed over..

I'm very familiar with the WSYNC stuff from my many years spent on the 2600 and also on the 7800 with DLIs and co That's all I wanted from him, rather than the incomplete pictures that have been painted by him.. And yeah, in bitmap modes you have way more freedom, of that I'm very envious and so are my interrupts What was interesting to just find out was that on A8 you must write to WSYNC before cycle 100 ?? Does this mean that you lose the remaining 14 additional cycles, or would they have been consumed by other memory accesses anyway ?

No, you read, and properly this time and don't go mis-quoting me and making out I said things I didn't.. I didn't dispute that DLIs or WSYNCs work in all modes.. You've just made that up.. I didn't say there were unstable/undeterminisitic/subjective/random/pink/alien-generated 'BA signals' in any mode... You just made that up again.. And of course you can switch modes where you want the cycles back, I didn't say you couldn't.. Ummm and for that matter, so can the 64.. Want a character screen with no badlines ? No problem, just abort the DMA on every 8th scanline, viola, no badlines and a display too boot[1].. The fact is that you're not portraying a clear picture here at all, willy-nilly deciding to leave out details of critical importance.. As for your question of "software algorithm is better than hardware that's already perfectly in sync with video beam", it seems like you have similar restrictions to the 64 the real difference being just a faster processor to do shit with when it comes to monkeying around with registers on those scanlines.. [1] Whereas Atariski would choose to leave it at that, there is the issue that doing that means the character matrix is repeated from the first fetch, as is the colour ram, so you have to change character sets every 8th line, but it does what it says on the tin.. A display with no badlines, and a display with stable/deterministic/objective/non-random/blue/hardware-generated 'BA signals'...

During the first char line you have interleaved char fetches and 1st line gfx fetches which shut down all CPU activity during the active part of the line. The remaining cycles are in the off-screen area (porches). So a very tiny, and invisible window That's all I was trying to get him to come clean with.. His claim that: is not entirely truthful, again.. It should really say like the c64, we have to write off the idea of doing any cycle exact interrupts on a badline.. Instead of making it sound like it can interrupt anywhere, anytime with exact cycle accuracy.. And I guess this also rather blows the idea of doing any of the much talked horizontal colour changes, mode changes, sprite plexing and other fancy shenangians in character modes since it'll look a bit poor not being able to do anything on every 8th line.. I'm sure you could work some design into it, but I'm just being baffled with all this talk of what you can do, but then it turns out it's not quite as Atariski, Emkay and Co. make out or there's some fairly large restrictions that someones deliberately forgotten to mention..