shoestring

I was late to the party. I didn’t like anything enough pre Pentium but I did start gaining some interested in it a little more over time. For me it was around when Win 95 came out, the system I built was a breath of fresh air and gaming was exciting for me once again with titles such as C&C, Quake and Destruction Derby. I wanted to keep using the Amiga but by that stage my hardware was getting old and worn, maintaining it was no longer affordable and no new software was coming out. The Pentium 120 met my needs

That guy is a machine when it comes to porting stuff. Outstanding work. I just got a stand-alone Vampire V4 yesterday , so I’m looking forward to using the Amiga and checking what I’ve missed out the past 15 years.

Interesting failure mode for a 7805 if that Ingot is indeed using one, they all use those right ? Some of the bad Commodore power supplies use a different regulator which are more prone to over voltage but the 7805 seems to be quite safe to use. I’ve had someone jump in on one of my threads saying they’d never seen that type of failure mode in their 25 years of experience. Was that under load conditions? Either way 6v is too high.

Machine is over 30 years old so I wouldn’t always suspect the PSU but considering it’s an ingot, I’d not be using it in the long term. It’s definitely not uncommon for more chips to fail on your bench one after the other as you’re working on it, I see it constantly with arcade stuff I’ve worked on. My understanding is that FREDDIE is CMOS as opposed to the other chips which are NMOS, so they won’t take as much of a beating and are usually prone to being one of the first chips to go a long with the DRAM. DRAM failures were a huge problem even back in the day when these chips were still brand new. You could find a NOS replacement fairly easily as I did. But with new tech available it would be much easier to get an SRAM replacement or even the Antonia board as everything required is on there. But I’d still be checking with Simius first to confirm that it will run happily without the FREDDIE chip.

In addition to the two missing ram chips, my Antonia board runs happily with the following devices removed from the 600xl. OS ROM BASIC ROM MMU 74LS158N x 2 74LS32N x1 74LS51N x 1 74LS375N x 1

I had a bit of a set back a couple of weeks ago and lost everything in a HD crash, yup.. my own fault. That includes source code for the A8 diagnostic software I wrote and all my arcade related stuff. So I had to start this project virtually from scratch and then use the xex I uploaded here to disassemble/reverse engineer what I did which helped me get back up to speed and on track rather quickly. I now have the Manta's bullets working ( using soft sprites ) in the attract mode with collisions to objects in place ( that includes destruction of parked ships and collisions to the obstacles on it ). Though that didn't take much work because the original game uses soft sprites for the players bullets as well. Uridium.atr Uridiump.xex

Atari 800 with Incognito and SCCC - NTSC Stock 800XL - PAL 600Xl with Antonia - PAL Stock 65XE - PAL

Wow. only $173 shipping to Australia, absolute bargain.

He shared the previous version, 1.2. But I’m sure someone can do better than this. It would be nice with some drums but I couldn’t find a way to make the bass channel sound decent after splitting it.

Jose has trouble accessing Atariage so has emailed me with some additional notes, and to pass this on in the thread.

That's actually quite interesting, I've never thought to try my Targus powered USB hub or just the switching adaptor it runs off ( 5v 4.0a output ) https://au.targus.com/products/usb-3-0-7-port-powered-hub-ach125au

They tend to, some worse than others and some so bad that the screen almost rolls. Their job is to charge so why would they waste $$ for more filtering. I have a PSP-100 charger as a backup, and there is notable difference on the screen between that and my Atari power supply in terms of quality, once you see it.. it's really hard to unsee. Use a proper power supply. Those cheap ass ones can cause house fires.

I sent you an email yesterday with all the ctms for each level, in monograys and 21 rows tall. Check your spam folder if you haven't seen it already.

Convenience and simply overestimated, overlooked the ability of the machine to be able to deal with the number of characters I would really need. But with my new python tool I can clearly see that a 1:1 pixel translation isn’t going to be feasible without resizing the Sprites. Someone mentioned antic E but then it becomes a different game, doesn’t it ? With half the horizontal resolution and only 4 colours for the playfield, maybe more with PMG but then you’d like to prioritise those for the manta. Mode F is better.. everything can be 1:1 but it lacks the sparkle and those huge memory requirements for both modes are prohibitive. Looks unlikely with just 64kb

Yep it doesn't look promising at all, just 6 sprites spawned belonging to foes ( that includes a combo of enemies, bullets & air mines ) pretty much blows the budget. If my calcs from last night are correct I'd need ~115 free per char set to fit it all into 01-Zinc, impossible.. and consider that every level has different requirements. So it's a minimum of 96 or so spare characters just for 6 sprites, not including frames for the enemy explosions, the two glyphs used for stars behind the dreadnought which aren't a big deal but still that's less two. Several characters are also required for the player's bullets, these move 2 character widths at a time and are character aligned. Would be much better off using one of the bitmap modes and writing an original game from scratch but with so many similar shooters, I'd have to think about what it would have to offer before I even go there. The other optionis looking into the feasibility of converting this to a VBXE project.

Notes were taken. Sorry, I’m not ignoring you. I just don’t always have time to respond to every reply in the thread.

We have sprites? Mmmm.... sprites in the border

Great little tool you got there. It gave me a little inspiration to write my own, but it's much less polished and sophisticated for now. The amount of free characters does vary between different levels too, quite significantly too. Character set summary for: zinc.png Total unique shapes generated for chrset 0: 102 Total unique shapes generated for chrset 1: 88 Total unique shapes generated for chrset 2: 99 And then for Level 6 there's quite a big difference. Character set summary for: gold.png Total unique shapes generated for chrset 0: 112 Total unique shapes generated for chrset 1: 101 Total unique shapes generated for chrset 2: 107

Ah of course, well for the enemies here is an example Sprite sheet. Enemies fly higher, so they don't need to cast a shadow, they also do not collide with the player. For each attack wave, only one type of sprite is used ( never a combination ). At most, 5 objects of the same type. So each will become 24x24 px wide or 3x3 chars, that's total of 256 8x8 px characters but maybe less 1 char for the duplicate. You also don't need access to all shapes on the screen since attack waves don't include different combinations of objects, so the required shapes can be copied to an active character set before an attack wave is generated. Then also factor in the animations for all the explosions ( the same frames are used for enemy bullets once they expire and for the air mine once it expires too ) of the enemies too which are quite numerous. There are 10 frames of animation and at 72 bytes for each frame of animation, that's around 720bytes. But you can probably get this down to around ~400 bytes with some redundancy and smarter coding. To throw a spanner into the works, multiple air mines mines can be deployed whilst enemy ships are still visible on the screen , and then of course there are the Manta's bullets and different varieties of enemy bullets as well.

I believe it should be 2 character sets not 3. G2F seems to duplicate most of the characters on the first line and re-arrange some of the others. Not sure if I am importing into G2F incorrectly. Charpad has the correct layout when I import the same .fnt file ( 256 characters ). So that's where the confusion lies. CharPad G2F I had to re-make level 04-Gold from scratch using the original character set as I have no way of keeping the same layout once I import an image, it automatically optimises the character set. I've attached the Charpad project here as well. 07 - Gold.ctm

Thanks, that's a big relief. So essentially all that work has already been done previously. Although for the split screen, only 17 rows of shapes are plotted for the dreadnoughts, there are 2 rows above and below reserved for the stars and the manta never moves beyond the boundaries of the dreadnought. So I'm not sure why the developer included those, I suppose because he could afford to do it. I recall only counting 280 unique tiles last night ( $7800 - $7FFF ), the rest are empty. There's also a character in that range reserved for the player bullets which are actual characters ( not sprites ) , the background character is copied to one of the reserved spaces then the bullet is combined before being copied to the screen. So for the soft sprites, this means they have to be overlayed whenever a bullet travels underneath them, and that's after the bullets have been overlaid a background character I'm assuming popmillo's program doesn't run properly under Win10? You could try running it under VM

Here's the first level done. 14 more to go, let me know if this is what you had in mind.

Firstly, I had to generate all the 15 unique character sets that were derived from A & B by dumping $7800-$7FFF ( attached ) after the game did the changes for me, thanks to trainers and MAME for easy dumping. So the requirements are 3 character sets for each 15 of the levels, each is generated by unpacking from character set A ( which is static, confirmed that it never changes ) then copying the required to character set B before the level starts. Next, since I have a reference for the tiles used in each level, I'll have to recreate the map for the entire game, without all the jpeg artifacts of course. So this will take time. For the [3x3] 24x24px character tiles, how would you like these arranged ? (ie.. vertically or horizontally). I'm not sure how your program works. chrsets.zip

Hmm. I haven’t generated those yet but I can see where this would be useful. I can force the current driver to display each level I want, then dump the screen frame by frame and manually stitch them all together using something like gimp. I wonder if I use something like G2F or Atari font maker to import the characters for each level and then use the screen data ( from $4800 ) to stitch the map by hand each time 40 characters have been shifted horizontally. I could dump the character tiles from mame and export 15 individual .fnt files from $7800 once the game has generated those before each level starts. Though I don’t see any quick method to generate the entire map without manually stitching them together frame by frame. Also mame can do things like view the tile maps and entire screens ( bg, fg , sprites ) generated by them. But I’ll just need to check and see if it supports browsing the different gfx sections for non arcade ( computer ) drivers as well.

For the 20 or so mode lines, here are the total number of shapes used for the dreadnoughts Character set A - 1kb ( from $7800 ) Character set B - 3kb ( Underneath I/O D400-DFFF via bank switching on the c64, I use $5C00 on the Atari for now ) So, 4kb total just to store all the different symbols used for the playfield throughout the game ( not including bonus stage ) The first character set from $7800 is appended/modified using the additional characters stored underneath the I/O ( character set B ). So you end up with a variation of the following depending on which level is built up. ( there are 15 levels or dreadnoughts in the game) $7800-$7FFF 2kb, 256 characters need to be referenced at any one time, as you've mentioned above. Now regarding the missing, or rather incorrect tiles displayed as I mentioned in my previous post. Even after importing the rest of the characters into my project and finally realising a few hours ( doh ) later that I was only going to be able to see the first 128 characters at one time, quickly flipping chbase from $7800 to $7C00 to confirm that the data for the few missing tiles is correct. So I'll definitely need to use an interleaved method similar to what is used in WonderBoy. I've also made some changes to the RMT as well, I think it sounds a little better now. uridium-sketch.v1.3.rmt