Snoozing by Cosine

[TMR]

Yays! It isn't much as demos go but i finally finished my first Amstrad CPC logo/picture/scroller part (running in mode 2 because that seemed like a fun thing to do and using ripped music from Saboteur 2 just because i like that tune) and released it yesterday. It's called Snoozing because i haven't entirely woken up [ahem =-]

It's available to download directly from the Cosine website or via the entry at Pouet (along with a Commodore 128 version that runs on the 640x200 VDC display). Here's a link to a screenshot which looked too big to include in this post when i previewed it and the more "forum-friendly" version that looks pretty naff because all the dithering has been mangled by the scaling:



[Runs away and hides before anyone says nasty things =-]

[tastefulmrship]

Exceptionally nice use of MODE 2. Let's hope people do more MODE 2 demos!

EDIT1: It is, however, humbling to see the number of active C64 groups greeted in this production when compared with the CPC scene! ^_^ C64 has more groups than we have people!

EDIT2: Here's a version of the demo that (hopefully) incorporates T&J's real-hardware fix. (attached)

[TMR]

Exceptionally nice use of MODE 2. Let's hope people do more MODE 2 demos!

i was worryingly pleased with myself when i got the logo working to be honest... =-)

It is, however, humbling to see the number of active C64 groups greeted in this production when compared with the CPC scene! ^_^ C64 has more groups than we have people!

Most of those are C64 groups (and a few are inactive right now) but there's also a few listed from the Spectrum, Atari 8-bit Oric and Plus/4 too.

[TMR]

EDIT2: Here's a version of the demo that (hopefully) incorporates T&J's real-hardware fix. (attached)

Just seen that at Pouet yeah... [sob!]

Edit: i've changed the code so that there's one less NOP before the colour splits - that replicates the offset i'm seeing in the fixed version so fingers crossed it's working - the archive at the Cosine site now contains the revised version as Cosine - Snoozing.dsk and the original as Cosine - Snoozing (WinAPE).dsk [Crosses fingers...]

[tastefulmrship]

Edit: i've changed the code so that there's one less NOP before the colour splits - that replicates the offset i'm seeing in the fixed version so fingers crossed it's working - the archive at the Cosine site now contains the revised version as Cosine - Snoozing.dsk and the original as Cosine - Snoozing (WinAPE).dsk [Crosses fingers...]

Both versions look perfect. Top job!

[Phi2x]

.

[TMR]

Both versions look perfect. Top job!

Thank goodness for that... what with this and the C128 version breaking (again, an emulator issue) i'm now thoroughly sick of this demo and plan to spend the rest of the day sulking. =-)

There's a reason why people recommend to always check the result on real hardware!

Yeah, normally i would but the only way i've got to get things to my 464 is with a tape converter and the lead's snapped so i need to go eBaying... but i honestly thought this thing was simple enough that it couldn't possibly break if the emulator said it was happy... after over fifteen years of coding 8-bit stuff with emulators you'd think i'd know better!

[Phi2x]

.

[TMR]

Colour changes happen 0.5µs late on a real CPC. On WinAPE, it's happening 1µs late. So removing a NOP (1µs) won't do the trick to fix the problem.
As Tom&Jerry said, the only way is to shift the picture by 8 pixels (mode 2).

Crap... okay, used the fixed DSK tastefulmrship posted as Cosine - Snoozing.dsk - thanks everybody for explaining.

[Devilmarkus]

Nice little prod!
Cool usage of MODE 2 here with nice rasters.

You can check it online here:

[Devilmarkus]

BTW.: I tested the snoozing.dsk and the snoozing_winape.dsk.
snoozing.dsk works fine in JavaCPC /- Applet.
But not in CPCBox, so which is the proper dsk now?

Snoozing.dsk in CPCBox:
[attach=2]

[Phi2x]

.

[Devilmarkus]

Did anyone transfer it yet?
I can't... Win7 and no floppy drive...

[MacDeath]

So, Mode2 and split rasters, and  slight scrooling with raster colour change at the bottom.

All with a sweet classical AY chiptune.

Split rasters must be a pain to synchronize i guess, does it renders differently on the various real Hardware (different CRTC, PLUS...) ?

good point is that those split rasters are some sort of colour attributes, yet not really character based sadly.


er... is this technic actually called split raster ?

Snoozing.dsk in CPCBox

yeah, the "split raster colour attributes" are a bit aside...
the bloc us demo suffered such problem on emulators according to some internet sources (videos...)

Perhaps such technic would need to scan for the actual hardware/emulator used and then adapt slightly the code when running the Demo... which can be quite delicate.

The question is: how it looks on a real CPC?

I will try it on my real CPC.

There you are...

Ok, on a CPC6128 (qwerty) from 1985... also old monitor...
This is supposed to be a CRTC0 (according to Wake up demo)

"Attributes" are still slightly on the left.

Donc know how much thought, hard to count the pixels on a real colour CRTM in Mode2.
Looks like less than 1 character perhaps... something like 4 pixels maybe...

Pic related.


IMO this technic is quite hard to use precisely... it may need to be used with designs that may need less precision, just like the way speccy graphic artistes managed to put black zones betweens the coloured zones...

Or you have to get your stuff with ultimate timing precision and take into account the varied Hardwares fluctuations...


Nice try anyway.

Welcome here dear Comodork, now you are also an Amstard.


Could someone test it on a proper Monochrome CRTM ? (I don't have mine anymore, i gave my CPC464+monocolour-monitor to someone...)

[MaV]

The question is: how it looks on a real CPC?

Both versions are off on a CPC 6128 / CRTC 0. Well, I hope the following descriptions make sense.

-) The colouring in the WinApe version is late by one byte. The colour on the left side begins after the upper right angle, exactly at the column where the inner pattern before the letter C begins.

-) The fixed version starts the colouring 1 byte before the Cosine logo (as seen in the fixed version), so it is too early.

So, for a CRTC 0 version one NOP is not fine grained enough to time the colour change, you're either too early or too late. You could of course place the logo to fit one of the two solutions (better base it on the fixed one).

On a CPC6128+ with the fixed version the colour is placed differently again. This time colouring begins exactly at the first column but the couple of pixels at the lower end where the C begins (supposedly) are not coloured.

[Devilmarkus]

Thats exactly what I thought, because I know, CPCBox is very raster accurate.

Good to know...

[MacDeath]

perhaps adding a "1 pixel black border" frame to those coloured buttons may be a good way to give them a bit more contrast/visibility, and would also ease the Colour transition.

there a mockup.


Also, is this an unborn girl ?

I mean, she has no navel /umbilicus/ belly button...

[Phi2x]

.

[Devilmarkus]

Woahhaaaa thats crazy...
We now need cycle exact rastering for each pixel
Are you sure, these nice raster pixels aren't handpainted by some cheap workers from Commodork?
(I know, Amstrad used them while the CPC had been produced)
So, let's look inside the CRT of our beloved CTMs
(And perhaps put some food in)

[MacDeath]

Yeah, that's why most modern Demo try to scan to define what CRTC is used, and modify the code according to this.

Needless to say, it doesn't simplify doing demos.

And it would give some demos effects/sequences availables to only some CRTC.

[TFM]

First impression: Speccy color clash & c64 palette. But it's nice release!

[arnoldemu]

I like this demo a lot and it's a great start on CPC. I like the use of mode 2.. something which doesn't seem to be well explored on CPC.

Keep going... it's good to see new demos on CPC.. it's nicely presented.

I read about the timing issues on pouet. This is something that can be hard to resolve without running on real hardware and then you have to run it on the variants (crtc type 3, type 4 and type 0), because the timing does differ on those.