Forums

slogra wrote

I'm sorry, i meant the orange footsoldier instead of the yellow one. I would use aa0000 instead of aa5500 on that one. It's a matter of taste though.

Ahhh, okay. Yeah, that looks way better.

slogra wrote

I still kinda like the previous dark purple one with blue. He looks really badass in those colors :p. It's a nice original color combo for a change.

As long as it's not jarring, still on the fence with it tbh.

slogra wrote

I think the chain of the nunchucks should be shorter.
Check out this one for example, the chain is very short if you even see it at all:
https://www.spriters-resource.com/game_boy_advance/teenagemutantninjaturtleskona...

I'll play around with it. I was trying to get closer to the original arcade version when I first created things.

[quote="xfixium"]

cireza wrote

xfixium wrote

cireza wrote

Hmm, I did previously check these values. As I thought that may be the case. However, the outlines are the same four all 4 turtles. At least with the captures I got from running it through Mensen.

You are right indeed. Sorry for the wrong information here, I thought they were all different.

Having two set of colors will indeed match the NES version.

Not a biggie, it's been awhile since I did that investigation, I wasn't sure myself.

Update wise, this is the current game I'm working on, at least for the minute XD. I've been testing states, and just overall performance of the game. Trying to optimize some pieces. SDCC can be tough to judge on the asm output in some scenarios. I have about 120 lines left of active display after processing some of the more major pieces. Not gonna lie, it's going to be tight. Some more progress:

https://www.pyxosoft.com/projects/sms_tmnt/videos/tmnt_0010.mp4

I was looking the video again and comparing with original arcade and NES. Seems that flames in below of screen are minus intense (small) tban other 2 versions. In arcade and NES the flames occuped great portion of screen. You will keep maintain the actual size of flames?

segarule wrote

I was looking the video again and comparing with original arcade and NES. Seems that flames in below of screen are minus intense (small) tban other 2 versions. In arcade and NES the flames occuped great portion of screen. You will keep maintain the actual size of flames?

Probably not. The n*s and arcade versions have bigger vertical resolutions. I was more concerned about keeping the same playable area intact. There are also some other reasons behind the limitation, although, I could possibly overcome those issues if necessary.

tmnt_0002.png (16.77 KB)

old compare when I first started this. The n*s version had to be cropped top and bottom to make it fit in a similar fashion

Really looks great so far, man this is a tall task to re-develop the entire game.

Not sure if Louis is working on other songs, but I also gave the music a shot. Starting with the shortest song in the game :)

Let me know if you think it needs some tweaking.
Feel free to use it (in fact, i'd be honored).

Turtles 2 - The Arcade Game - Character Select (SMS).vgm (9.14 KB)

And another short tune

Turtles 2 - The Arcade Game - Scene 1 Intro (SMS).vgm (10.38 KB)

Oh damn those sound amazing! Much better than anything I would've been able to come up with O_O

xfixium wrote

segarule wrote

I was looking the video again and comparing with original arcade and NES. Seems that flames in below of screen are minus intense (small) tban other 2 versions. In arcade and NES the flames occuped great portion of screen. You will keep maintain the actual size of flames?

Probably not. The n*s and arcade versions have bigger vertical resolutions. I was more concerned about keeping the same playable area intact. There are also some other reasons behind the limitation, although, I could possibly overcome those issues if necessary.

I feel you there. Currently working on a port of Final Fantasy I and the vertical squish has been gnarly to work around since all of the UI elements are designed to fit in a 256x240 screen space, including Japanese characters. I squished it all, but now a Japanese translation would be much harder to make.

Best of luck on this though!

I'm working on the level 2 music.
I've also slightly updated the other 2 songs.

Comments or ideas are welcome.

turtles2-level2_wip1.vgm (36.38 KB)

turtles2-scene1_wip2.vgm (10.38 KB)

turtles2-selectcharacter_wip2.vgm (9.14 KB)

slogra wrote

I'm working on the level 2 music.
I've also slightly updated the other 2 songs.

Comments or ideas are welcome.

Very good! Let´s see how you will go with Technodrome track.

Do you mean the boss theme, or another track?

This song is amazing. Changing the instruments and rithms

slogra wrote

I'm working on the level 2 music.
I've also slightly updated the other 2 songs.

Comments or ideas are welcome.

I wanted to comment on this much earlier, but I didn't get the time to sit down and really listen to these tracks. They sound phenomenal so far!

Thanks!

I just realized that the level 2 song sounds much better an octave down, which was an easy fix. What do you think?

turtles2-level2_wip2.vgm (36.49 KB)

slogra wrote

Thanks!

I just realized that the level 2 song sounds much better an octave down, which was an easy fix. What do you think?

Oh yeah, big difference. Much easier on the ears, frequency wise. Digging what you're doing with the bass channel.

Keep up the amazing work!

Your dev update and the FM track from Louis The Sega Nerd inspired me to make this:

You can find the vgm in the video description.

slogra wrote

Keep up the amazing work!

Your dev update and the FM track from Louis The Sega Nerd inspired me to make this:

You can find the vgm in the video description.

Man, you are a real genius. I don't usually write here, but everything you do is amazing. Congratulations!

slogra wrote

Keep up the amazing work!

Your dev update and the FM track from Louis The Sega Nerd inspired me to make this:

You can find the vgm in the video description.

Amazing stuff, truly talented! I placed your two tracks in from prior, if you don't mind.

WIP:
https://pyxosoft.com/projects/sms_tmnt/videos/tmnt_0011.mp4

there are some minor glitches. But it basically finishes up phase 1 of development.

xfixium wrote

Amazing stuff, truly talented! I placed your two tracks in from prior, if you don't mind.

WIP:
https://pyxosoft.com/projects/sms_tmnt/videos/tmnt_0011.mp4

there are some minor glitches. But it basically finishes up phase 1 of development.

I don't mind at all :). Will you use the new tune at the intro before and during the title screen?

And many thanks for another video update! I don't see any glitches except the little glitch at the top of the screen when scrolling down. Otherwise it looks perfect to me. And those voice samples are really the cherry on top.

Also I had an idea about the sprite limit. Would it be possible to combine normal sprites and background sprites? Use the background sprites for less active enemies that are not moving or attacking.
I guess it's not easy to make it work, but it may help get the flickering down.

slogra wrote

I don't mind at all :). Will you use the new tune at the intro before and during the title screen?

If you don't mind?

slogra wrote

And many thanks for another video update! I don't see any glitches except the little glitch at the top of the screen when scrolling down.

Yeah thatz be one of them. Plus some odd sprite glitches where sprites are at the wrong coordinates for a split second.
(In the Fire!! scene, when the turtles jump off the roof) Not sure what the deal is there, as the spritez stuff is done at the correct times, via DevKit's new metasprite feature. (Which is quite a nice feature)

slogra wrote

Also I had an idea about the sprite limit. Would it be possible to combine normal sprites and background sprites? Use the background sprites for less active enemies that are not moving or attacking.
I guess it's not easy to make it work, but it may help get the flickering down.

I will try to reduce where I can. For Castlevania I made the candles BG tiles instead of sprites. Saves a little flicker. Hopefully there will be chances to reduce sprite use. I know the N*s game uses a separate sprite for the turtle's weapon, due to palette constraints. Maybe the MS version will fair a little better, since it doesn't have that issue.

xfixium wrote

Plus some odd sprite glitches where sprites are at the wrong coordinates for a split second.
(In the Fire!! scene, when the turtles jump off the roof) Not sure what the deal is there, as the spritez stuff is done at the correct times, via DevKit's new metasprite feature. (Which is quite a nice feature)

If you advance frame by frame in an emulator, you could probably tell what's going on. Hard to see in the video.

sverx wrote

If you advance frame by frame in an emulator, you could probably tell what's going on. Hard to see in the video.

Aye, I had done this a bit ago. It's very odd. I can't even explain it. And it happens on every SMS game I've ever dev'd for. My plan was to create a base project displaying the glitch, and present it, so that more experienced eyes can take a look at it. It happens on Emulicious, Kega, and on real hardware. The glitch only seems to happen if the sprite coordinates differ from one frame to another.

This is a small gif showing the issue in Castlevania (This was just easier for me to show):

https://pyxosoft.com/projects/sms_castlevania/castlevania.gif

In the vid it shows Simon's meta sprite for his second attack frame, but for a fraction of a second, it retains the last frame's meta sprite coordinates. Which differ in Simon's top half between those two frames.

And when hovering over those sprites in Emulicious, it looks like the graphic itself is out of position, but reports the sprite coordinates are in the correct position.

Output from SMS Tile Studio for DevKit SMS:

// classic_simon_attack_right
unsigned char const classic_simon_attack_right_00[] = { -8, 0, 0, 0, 0, 2, 0, 16, 4, 8, 16, 6, METASPRITE_END };
unsigned char const classic_simon_attack_right_01[] = { 0, 0, 8, 8, 0, 10, 0, 16, 12, 8, 16, 14, METASPRITE_END };
unsigned char const classic_simon_attack_right_02[] = { 0, 0, 16, 8, 0, 18, 0, 16, 20, 8, 16, 22, METASPRITE_END };
unsigned char const *classic_simon_attack_right_frames[] = { classic_simon_attack_right_00, classic_simon_attack_right_01, classic_simon_attack_right_02 };

SMS Tile Studio:
https://pyxosoft.com/projects/sms_castlevania/castlevania_001.png
https://pyxosoft.com/projects/sms_castlevania/castlevania_002.png

this is very weird, and I suspect the only chance for this to happen would be if you update the SAT midframe

as you by chance calling SMS_copySpritestoSAT before SMS_waitForVBlank?

sverx wrote

this is very weird, and I suspect the only chance for this to happen would be if you update the SAT midframe

as you by chance calling SMS_copySpritestoSAT() before SMS_waitForVBlank()?

I am calling it (More precisely, the unsafe version of it) right after SMS_waitForVBlank(). Now, I do call SMS_initSprites(); right after that copySpritestoSAT call. Could that be affecting things?

xfixium wrote

Now, I do call SMS_initSprites(); right after that copySpritestoSAT call. Could that be affecting things?

It shouldn't, as it simply resets the sprite counter, but you've already 'flushed' the sprite list to VRAM when that happens...

So I have no further ideas. If you wish you can privately send me a small test ROM and I can investigate...

sverx wrote

It shouldn't, as it simply resets the sprite counter, but you've already 'flushed' the sprite list to VRAM when that happens...

So I have no further ideas. If you wish you can privately send me a small test ROM and I can investigate...

I figured as much. It makes no sense to me. I'll put together some source/binary.

also, there's a bit in a certain VDP register that makes all the sprites appear 8 pixels to the left, so I'm wondering if it's possible you're writing there by mistake... are you using line interrupts?

sverx wrote

also, there's a bit in a certain VDP register that makes all the sprites appear 8 pixels to the left, so I'm wondering if it's possible you're writing there by mistake... are you using line interrupts?

I use them in Castlevania, but not during game play. Instead, opting to use the lock horizontal bit VDP feature (Which, maybe is an IRQ under the hood), since the game is only using 2 rows.

I believe I found the issue. The missing component to this, that I forgot to disclose, is that I'm streaming in tiles for Simon. It's a timing issue between UNSAFE_SMS_VRAMmemcpy128 calls and where I'm calling UNSAFE_SMS_copySpritestoSAT. There's a frame where the tiles update via streaming, then active screen occurs, and then the sprites update. Giving it that glitch for just a slight moment.

xfixium wrote

use the lock horizontal bit VDP feature (Which, maybe is an IRQ under the hood)

no, it's not, it's an hardware feature of the VDP, the CPU doesn't get any interrupt

xfixium wrote

I believe I found the issue. The missing component to this, that I forgot to disclose, is that I'm streaming in tiles for Simon. It's a timing issue between UNSAFE_SMS_VRAMmemcpy128 calls and where I'm calling UNSAFE_SMS_copySpritestoSAT. There's a frame where the tiles update via streaming, then active screen occurs, and then the sprites update. Giving it that glitch for just a slight moment.

I see. You should probably update the sprites and Simon tiles in the same vblank, there should be enough time for that.

sverx wrote

no, it's not, it's an hardware feature of the VDP, the CPU doesn't get any interrupt

Very nice

sverx wrote

I see. You should probably update the sprites and Simon tiles in the same vblank, there should be enough time for that.

Definitely doable

Author	Message
xfixium Joined: 18 Jul 2020 Posts: 386	Posted: Tue Jul 25, 2023 12:08 pm
xfixium Joined: 18 Jul 2020 Posts: 386	slogra wrote I'm sorry, i meant the orange footsoldier instead of the yellow one. I would use aa0000 instead of aa5500 on that one. It's a matter of taste though. Ahhh, okay. Yeah, that looks way better. slogra wrote I still kinda like the previous dark purple one with blue. He looks really badass in those colors :p. It's a nice original color combo for a change. As long as it's not jarring, still on the fence with it tbh. slogra wrote I think the chain of the nunchucks should be shorter. Check out this one for example, the chain is very short if you even see it at all: https://www.spriters-resource.com/game_boy_advance/teenagemutantninjaturtleskona... I'll play around with it. I was trying to get closer to the original arcade version when I first created things.

segarule Joined: 23 Jan 2010 Posts: 443	Posted: Wed Jul 26, 2023 10:52 am
segarule Joined: 23 Jan 2010 Posts: 443	[quote="xfixium"] cireza wrote xfixium wrote cireza wrote Hmm, I did previously check these values. As I thought that may be the case. However, the outlines are the same four all 4 turtles. At least with the captures I got from running it through Mensen. You are right indeed. Sorry for the wrong information here, I thought they were all different. Having two set of colors will indeed match the NES version. Not a biggie, it's been awhile since I did that investigation, I wasn't sure myself. Update wise, this is the current game I'm working on, at least for the minute XD. I've been testing states, and just overall performance of the game. Trying to optimize some pieces. SDCC can be tough to judge on the asm output in some scenarios. I have about 120 lines left of active display after processing some of the more major pieces. Not gonna lie, it's going to be tight. Some more progress: https://www.pyxosoft.com/projects/sms_tmnt/videos/tmnt_0010.mp4 I was looking the video again and comparing with original arcade and NES. Seems that flames in below of screen are minus intense (small) tban other 2 versions. In arcade and NES the flames occuped great portion of screen. You will keep maintain the actual size of flames?

xfixium Joined: 18 Jul 2020 Posts: 386	Posted: Wed Jul 26, 2023 12:05 pm
xfixium Joined: 18 Jul 2020 Posts: 386	segarule wrote I was looking the video again and comparing with original arcade and NES. Seems that flames in below of screen are minus intense (small) tban other 2 versions. In arcade and NES the flames occuped great portion of screen. You will keep maintain the actual size of flames? Probably not. The ns and arcade versions have bigger vertical resolutions. I was more concerned about keeping the same playable area intact. There are also some other reasons behind the limitation, although, I could possibly overcome those issues if necessary. tmnt_0002.png (16.77 KB) old compare when I first started this. The ns version had to be cropped top and bottom to make it fit in a similar fashion

Greg2600 Joined: 08 Apr 2011 Posts: 29	Posted: Fri Aug 25, 2023 2:02 am
Greg2600 Joined: 08 Apr 2011 Posts: 29	Really looks great so far, man this is a tall task to re-develop the entire game.

slogra Joined: 09 Jun 2014 Posts: 381	Posted: Sat Aug 26, 2023 9:47 am
slogra Joined: 09 Jun 2014 Posts: 381	Not sure if Louis is working on other songs, but I also gave the music a shot. Starting with the shortest song in the game :) Let me know if you think it needs some tweaking. Feel free to use it (in fact, i'd be honored). Turtles 2 - The Arcade Game - Character Select (SMS).vgm (9.14 KB)

slogra Joined: 09 Jun 2014 Posts: 381	Posted: Sat Aug 26, 2023 10:53 am
slogra Joined: 09 Jun 2014 Posts: 381	And another short tune Turtles 2 - The Arcade Game - Scene 1 Intro (SMS).vgm (10.38 KB)

Louis The SEGA Nerd Joined: 30 Jan 2023 Posts: 55 Location: Manchester, England	Posted: Thu Aug 31, 2023 8:05 pm
	Oh damn those sound amazing! Much better than anything I would've been able to come up with O_O

CerezaSaturn64 Joined: 25 Feb 2023 Posts: 99	Posted: Sat Sep 02, 2023 9:41 pm
CerezaSaturn64 Joined: 25 Feb 2023 Posts: 99	xfixium wrote segarule wrote I was looking the video again and comparing with original arcade and NES. Seems that flames in below of screen are minus intense (small) tban other 2 versions. In arcade and NES the flames occuped great portion of screen. You will keep maintain the actual size of flames? Probably not. The n*s and arcade versions have bigger vertical resolutions. I was more concerned about keeping the same playable area intact. There are also some other reasons behind the limitation, although, I could possibly overcome those issues if necessary. I feel you there. Currently working on a port of Final Fantasy I and the vertical squish has been gnarly to work around since all of the UI elements are designed to fit in a 256x240 screen space, including Japanese characters. I squished it all, but now a Japanese translation would be much harder to make. Best of luck on this though!

slogra Joined: 09 Jun 2014 Posts: 381	Posted: Wed Sep 06, 2023 9:38 am
slogra Joined: 09 Jun 2014 Posts: 381	I'm working on the level 2 music. I've also slightly updated the other 2 songs. Comments or ideas are welcome. turtles2-level2_wip1.vgm (36.38 KB) turtles2-scene1_wip2.vgm (10.38 KB) turtles2-selectcharacter_wip2.vgm (9.14 KB)

segarule Joined: 23 Jan 2010 Posts: 443	Posted: Wed Sep 06, 2023 10:01 am
segarule Joined: 23 Jan 2010 Posts: 443	slogra wrote I'm working on the level 2 music. I've also slightly updated the other 2 songs. Comments or ideas are welcome. Very good! Let´s see how you will go with Technodrome track.

slogra Joined: 09 Jun 2014 Posts: 381	Posted: Wed Sep 06, 2023 10:29 am
slogra Joined: 09 Jun 2014 Posts: 381	Do you mean the boss theme, or another track?

segarule Joined: 23 Jan 2010 Posts: 443	Posted: Wed Sep 06, 2023 11:39 am
segarule Joined: 23 Jan 2010 Posts: 443	This song is amazing. Changing the instruments and rithms

xfixium Joined: 18 Jul 2020 Posts: 386	Posted: Thu Oct 05, 2023 3:56 am
xfixium Joined: 18 Jul 2020 Posts: 386	slogra wrote I'm working on the level 2 music. I've also slightly updated the other 2 songs. Comments or ideas are welcome. I wanted to comment on this much earlier, but I didn't get the time to sit down and really listen to these tracks. They sound phenomenal so far!

slogra Joined: 09 Jun 2014 Posts: 381	Posted: Thu Oct 05, 2023 7:15 pm
slogra Joined: 09 Jun 2014 Posts: 381	Thanks! I just realized that the level 2 song sounds much better an octave down, which was an easy fix. What do you think? turtles2-level2_wip2.vgm (36.49 KB)

xfixium Joined: 18 Jul 2020 Posts: 386	Posted: Thu Oct 05, 2023 7:22 pm
xfixium Joined: 18 Jul 2020 Posts: 386	slogra wrote Thanks! I just realized that the level 2 song sounds much better an octave down, which was an easy fix. What do you think? Oh yeah, big difference. Much easier on the ears, frequency wise. Digging what you're doing with the bass channel.

slogra Joined: 09 Jun 2014 Posts: 381	Posted: Thu May 02, 2024 9:45 am
slogra Joined: 09 Jun 2014 Posts: 381	Keep up the amazing work! Your dev update and the FM track from Louis The Sega Nerd inspired me to make this: You can find the vgm in the video description.

baude Joined: 15 Dec 2015 Posts: 17	Posted: Thu May 02, 2024 2:52 pm
baude Joined: 15 Dec 2015 Posts: 17	slogra wrote Keep up the amazing work! Your dev update and the FM track from Louis The Sega Nerd inspired me to make this: You can find the vgm in the video description. Man, you are a real genius. I don't usually write here, but everything you do is amazing. Congratulations!

xfixium Joined: 18 Jul 2020 Posts: 386	Posted: Thu May 02, 2024 8:12 pm
xfixium Joined: 18 Jul 2020 Posts: 386	slogra wrote Keep up the amazing work! Your dev update and the FM track from Louis The Sega Nerd inspired me to make this: You can find the vgm in the video description. Amazing stuff, truly talented! I placed your two tracks in from prior, if you don't mind. WIP: https://pyxosoft.com/projects/sms_tmnt/videos/tmnt_0011.mp4 there are some minor glitches. But it basically finishes up phase 1 of development.

slogra Joined: 09 Jun 2014 Posts: 381	Posted: Yesterday at 2:37 am
slogra Joined: 09 Jun 2014 Posts: 381	xfixium wrote Amazing stuff, truly talented! I placed your two tracks in from prior, if you don't mind. WIP: https://pyxosoft.com/projects/sms_tmnt/videos/tmnt_0011.mp4 there are some minor glitches. But it basically finishes up phase 1 of development. I don't mind at all :). Will you use the new tune at the intro before and during the title screen? And many thanks for another video update! I don't see any glitches except the little glitch at the top of the screen when scrolling down. Otherwise it looks perfect to me. And those voice samples are really the cherry on top. Also I had an idea about the sprite limit. Would it be possible to combine normal sprites and background sprites? Use the background sprites for less active enemies that are not moving or attacking. I guess it's not easy to make it work, but it may help get the flickering down.

xfixium Joined: 18 Jul 2020 Posts: 386	Posted: Yesterday at 4:34 am
xfixium Joined: 18 Jul 2020 Posts: 386	slogra wrote I don't mind at all :). Will you use the new tune at the intro before and during the title screen? If you don't mind? slogra wrote And many thanks for another video update! I don't see any glitches except the little glitch at the top of the screen when scrolling down. Yeah thatz be one of them. Plus some odd sprite glitches where sprites are at the wrong coordinates for a split second. (In the Fire!! scene, when the turtles jump off the roof) Not sure what the deal is there, as the spritez stuff is done at the correct times, via DevKit's new metasprite feature. (Which is quite a nice feature) slogra wrote Also I had an idea about the sprite limit. Would it be possible to combine normal sprites and background sprites? Use the background sprites for less active enemies that are not moving or attacking. I guess it's not easy to make it work, but it may help get the flickering down. I will try to reduce where I can. For Castlevania I made the candles BG tiles instead of sprites. Saves a little flicker. Hopefully there will be chances to reduce sprite use. I know the N*s game uses a separate sprite for the turtle's weapon, due to palette constraints. Maybe the MS version will fair a little better, since it doesn't have that issue.

sverx Joined: 05 Sep 2013 Posts: 3841 Location: Stockholm, Sweden	Posted: Yesterday at 9:19 am
	xfixium wrote Plus some odd sprite glitches where sprites are at the wrong coordinates for a split second. (In the Fire!! scene, when the turtles jump off the roof) Not sure what the deal is there, as the spritez stuff is done at the correct times, via DevKit's new metasprite feature. (Which is quite a nice feature) If you advance frame by frame in an emulator, you could probably tell what's going on. Hard to see in the video.

xfixium Joined: 18 Jul 2020 Posts: 386	Posted: Yesterday at 11:54 am
xfixium Joined: 18 Jul 2020 Posts: 386	sverx wrote If you advance frame by frame in an emulator, you could probably tell what's going on. Hard to see in the video. Aye, I had done this a bit ago. It's very odd. I can't even explain it. And it happens on every SMS game I've ever dev'd for. My plan was to create a base project displaying the glitch, and present it, so that more experienced eyes can take a look at it. It happens on Emulicious, Kega, and on real hardware. The glitch only seems to happen if the sprite coordinates differ from one frame to another. This is a small gif showing the issue in Castlevania (This was just easier for me to show): https://pyxosoft.com/projects/sms_castlevania/castlevania.gif In the vid it shows Simon's meta sprite for his second attack frame, but for a fraction of a second, it retains the last frame's meta sprite coordinates. Which differ in Simon's top half between those two frames. And when hovering over those sprites in Emulicious, it looks like the graphic itself is out of position, but reports the sprite coordinates are in the correct position. Output from SMS Tile Studio for DevKit SMS: // classic_simon_attack_right unsigned char const classic_simon_attack_right_00[] = { -8, 0, 0, 0, 0, 2, 0, 16, 4, 8, 16, 6, METASPRITE_END }; unsigned char const classic_simon_attack_right_01[] = { 0, 0, 8, 8, 0, 10, 0, 16, 12, 8, 16, 14, METASPRITE_END }; unsigned char const classic_simon_attack_right_02[] = { 0, 0, 16, 8, 0, 18, 0, 16, 20, 8, 16, 22, METASPRITE_END }; unsigned char const *classic_simon_attack_right_frames[] = { classic_simon_attack_right_00, classic_simon_attack_right_01, classic_simon_attack_right_02 }; SMS Tile Studio: https://pyxosoft.com/projects/sms_castlevania/castlevania_001.png https://pyxosoft.com/projects/sms_castlevania/castlevania_002.png

sverx Joined: 05 Sep 2013 Posts: 3841 Location: Stockholm, Sweden	Posted: Yesterday at 12:10 pm
	this is very weird, and I suspect the only chance for this to happen would be if you update the SAT midframe as you by chance calling SMS_copySpritestoSAT before SMS_waitForVBlank?

xfixium Joined: 18 Jul 2020 Posts: 386	Posted: Yesterday at 12:24 pm
xfixium Joined: 18 Jul 2020 Posts: 386	sverx wrote this is very weird, and I suspect the only chance for this to happen would be if you update the SAT midframe as you by chance calling SMS_copySpritestoSAT() before SMS_waitForVBlank()? I am calling it (More precisely, the unsafe version of it) right after SMS_waitForVBlank(). Now, I do call SMS_initSprites(); right after that copySpritestoSAT call. Could that be affecting things?

sverx Joined: 05 Sep 2013 Posts: 3841 Location: Stockholm, Sweden	Posted: Yesterday at 12:31 pm
	xfixium wrote Now, I do call SMS_initSprites(); right after that copySpritestoSAT call. Could that be affecting things? It shouldn't, as it simply resets the sprite counter, but you've already 'flushed' the sprite list to VRAM when that happens... So I have no further ideas. If you wish you can privately send me a small test ROM and I can investigate...

xfixium Joined: 18 Jul 2020 Posts: 386	Posted: Yesterday at 12:56 pm
xfixium Joined: 18 Jul 2020 Posts: 386	sverx wrote It shouldn't, as it simply resets the sprite counter, but you've already 'flushed' the sprite list to VRAM when that happens... So I have no further ideas. If you wish you can privately send me a small test ROM and I can investigate... I figured as much. It makes no sense to me. I'll put together some source/binary.

sverx Joined: 05 Sep 2013 Posts: 3841 Location: Stockholm, Sweden	Posted: Yesterday at 1:13 pm
	also, there's a bit in a certain VDP register that makes all the sprites appear 8 pixels to the left, so I'm wondering if it's possible you're writing there by mistake... are you using line interrupts?

xfixium Joined: 18 Jul 2020 Posts: 386	Posted: Yesterday at 3:26 pm
xfixium Joined: 18 Jul 2020 Posts: 386	sverx wrote also, there's a bit in a certain VDP register that makes all the sprites appear 8 pixels to the left, so I'm wondering if it's possible you're writing there by mistake... are you using line interrupts? I use them in Castlevania, but not during game play. Instead, opting to use the lock horizontal bit VDP feature (Which, maybe is an IRQ under the hood), since the game is only using 2 rows. I believe I found the issue. The missing component to this, that I forgot to disclose, is that I'm streaming in tiles for Simon. It's a timing issue between UNSAFE_SMS_VRAMmemcpy128 calls and where I'm calling UNSAFE_SMS_copySpritestoSAT. There's a frame where the tiles update via streaming, then active screen occurs, and then the sprites update. Giving it that glitch for just a slight moment.

sverx Joined: 05 Sep 2013 Posts: 3841 Location: Stockholm, Sweden	Posted: Yesterday at 4:13 pm
	xfixium wrote use the lock horizontal bit VDP feature (Which, maybe is an IRQ under the hood) no, it's not, it's an hardware feature of the VDP, the CPU doesn't get any interrupt xfixium wrote I believe I found the issue. The missing component to this, that I forgot to disclose, is that I'm streaming in tiles for Simon. It's a timing issue between UNSAFE_SMS_VRAMmemcpy128 calls and where I'm calling UNSAFE_SMS_copySpritestoSAT. There's a frame where the tiles update via streaming, then active screen occurs, and then the sprites update. Giving it that glitch for just a slight moment. I see. You should probably update the sprites and Simon tiles in the same vblank, there should be enough time for that.

xfixium Joined: 18 Jul 2020 Posts: 386	Posted: Yesterday at 5:03 pm
xfixium Joined: 18 Jul 2020 Posts: 386	sverx wrote no, it's not, it's an hardware feature of the VDP, the CPU doesn't get any interrupt Very nice sverx wrote I see. You should probably update the sprites and Simon tiles in the same vblank, there should be enough time for that. Definitely doable

Forums

View topic - SMS TMNT Homebrew