One reason for copying the character set data from ROM is to allow you to use existing characters/numbers for things like "Score:00100" and to use other characters for your custom symbols/font. Doing it this way means you don't need to switch between character sets. But, if you plan to customise the entire character set then there's no need to do the initial copy. Really enjoying the vids so far! It's been EDzactly what I've been looking for. I've been following along using the WUSDN editor - so glad to have found it - it's been making my journey back to my childhood much more enjoyable. Keep the videos coming - there's nothing like a good EDucation :)
Love it! Yeah, I will be moving back to copying eventually, since I need to be able to "cherry pick" certain characters, since I'll have more than the charset will hold.
@@MissionEdPossible You're too cool, it was necessary for me to do so, even though I am a seasoned 8bit coder! By the way, in ANTIC 2, you can use the P/M graphics at 4x wide (All sprites 4x wide) and fill the screen, then you can get a 3rd (Even 4th) color by turning on/off playermissle pixels! I'm terrible at explaining it but it's useful when you use ANTIC 2 or 3. Also, ANTIC 3 is 10 lines high instead of 8 like ANTIC 2.
Another thing you can do to put graphics assets in the proper locations without doing the math is use .align size. Since a character set requires 1K boundary for mode 2 or mode 4/5, you use .align $0400, then declare the charset label. MADS puts charset at the next address aligned to a 1K boundary. Then after charset is declared, just put the .bytes defining the character set and MADS assembles the character set images in the memory for the character set. No copying necessary.
Or in an assembler like ca65, you can declare a new segment that's aligned to $0400. That way it doesn't have to pad the file with a bunch of unused bytes to get to that alignment and the linker can put that segment wherever it's convenient.
@@SyntheToonz Maybe that depends on the assembler and the align parameter. I was just reading about it padding the data segment to align data. It makes sense that it would create a new segment for something that needs to be aligned to a page boundary or larger, but it would be ridiculous to make a segment for something that needs to be aligned to a double byte to a few bytes.
@@chitlitlah i figure if the assembler discussed can also cross assemble for the c64, then for that target it would have to pad. Otherwise any decent assembler that knows the Ataris will make load segments.
Hi, i'm working on Linux and trying to do this using Stella as emulator, but its gime me an error when trying to execute the .obx file. You know if some extra configuration is needed? Ty!
Well, first off, Stella is an Atari 2600/VCS emulator, which has a much different architecture. You're going to want an Atari 8-bit emulator. For unix based systems, you'll probably want to try out atari800. Good luck!
I am converting this to the Atari 5200 to learn what you are explaining. Is copying the Character Set to RAM that important? I was going to just set CHBASE to a location on the ROM cartridge and read the Character Set directly from ROM. Or is this an ATARI 5200 specific capability?
It's not important if you're just replacing the characters - If you're going to use the existing ROM character set, there's no need to mess with it at all.
If this: mva chars,x charset,x results in this: lda chars,x sta charset,x lda chars+1,x sta charset+1,x (copying two characters) Then this: loop mva chars,x charset,x inx cpx #16 bne loop Should increment X twice. loop mva chars,x charset,x inx inx cpx #16 bne loop (Same situation for copying bytes to screen memory using Y register.)
Keep going man, this tutorial series is great.
Thanks, will do!
I'm trying to make my first game on an 800 I picked up, and this series has already been invaluable
Great! Glad it's helpful!
One reason for copying the character set data from ROM is to allow you to use existing characters/numbers for things like "Score:00100" and to use other characters for your custom symbols/font. Doing it this way means you don't need to switch between character sets. But, if you plan to customise the entire character set then there's no need to do the initial copy.
Really enjoying the vids so far! It's been EDzactly what I've been looking for. I've been following along using the WUSDN editor - so glad to have found it - it's been making my journey back to my childhood much more enjoyable.
Keep the videos coming - there's nothing like a good EDucation :)
Love it! Yeah, I will be moving back to copying eventually, since I need to be able to "cherry pick" certain characters, since I'll have more than the charset will hold.
So yesterday I fell asleep listening to this playlist and dreamt about assembly language all night!
Love it!
Dude, you are AWESOME. You've earned your 580th sub!
Thanks so much! Very much appreciated!
@@MissionEdPossible You're too cool, it was necessary for me to do so, even though I am a seasoned 8bit coder!
By the way, in ANTIC 2, you can use the P/M graphics at 4x wide (All sprites 4x wide) and fill the screen, then you can get a 3rd (Even 4th) color by turning on/off playermissle pixels! I'm terrible at explaining it but it's useful when you use ANTIC 2 or 3. Also, ANTIC 3 is 10 lines high instead of 8 like ANTIC 2.
This is extremely awesome. Thank you so much for this great series. I'm totally excited about the coming videos.
Episode 4 is up now!
Another thing you can do to put graphics assets in the proper locations without doing the math is use .align size. Since a character set requires 1K boundary for mode 2 or mode 4/5, you use .align $0400, then declare the charset label. MADS puts charset at the next address aligned to a 1K boundary.
Then after charset is declared, just put the .bytes defining the character set and MADS assembles the character set images in the memory for the character set. No copying necessary.
Yep, I knew about the alignment thing, but I figured I'd do it the hard way. :)
Or in an assembler like ca65, you can declare a new segment that's aligned to $0400. That way it doesn't have to pad the file with a bunch of unused bytes to get to that alignment and the linker can put that segment wherever it's convenient.
@@chitlitlah .align does not pad/fill. It will create a load file segment.
@@SyntheToonz Maybe that depends on the assembler and the align parameter. I was just reading about it padding the data segment to align data. It makes sense that it would create a new segment for something that needs to be aligned to a page boundary or larger, but it would be ridiculous to make a segment for something that needs to be aligned to a double byte to a few bytes.
@@chitlitlah i figure if the assembler discussed can also cross assemble for the c64, then for that target it would have to pad. Otherwise any decent assembler that knows the Ataris will make load segments.
More magic!
Another great episode, easy to follow, keep up the good work 😀
Thanks so much!
спасибо, очень полезно.
Всегда пожалуйста!
Hi, i'm working on Linux and trying to do this using Stella as emulator, but its gime me an error when trying to execute the .obx file. You know if some extra configuration is needed? Ty!
Well, first off, Stella is an Atari 2600/VCS emulator, which has a much different architecture. You're going to want an Atari 8-bit emulator. For unix based systems, you'll probably want to try out atari800. Good luck!
I am converting this to the Atari 5200 to learn what you are explaining. Is copying the Character Set to RAM that important? I was going to just set CHBASE to a location on the ROM cartridge and read the Character Set directly from ROM. Or is this an ATARI 5200 specific capability?
It's not important if you're just replacing the characters - If you're going to use the existing ROM character set, there's no need to mess with it at all.
Oh, and if you get it working, definitely share!
If this:
mva chars,x charset,x
results in this:
lda chars,x
sta charset,x
lda chars+1,x
sta charset+1,x
(copying two characters)
Then this:
loop
mva chars,x charset,x
inx
cpx #16
bne loop
Should increment X twice.
loop
mva chars,x charset,x
inx
inx
cpx #16
bne loop
(Same situation for copying bytes to screen memory using Y register.)
So, mwa is the one to copy a word (two characters), mva copies a byte. I *think* I'm using it right -- if not, definitely let me know :)
@@MissionEdPossible oh. I thought it was the same as the one used at the start for SDLSTL/H. Sooo, never mind.
Instead of loading the char set just origin the data to $3C00
Interesting, I'll try that out - thanks!