Problem with Battle Squadron

Eric C · 17 February 2005, 15:13

Hi,

I think I have found a bug within the emulation, with WinUAE 0.9.92.

When I play Battle Squadron, some sprites are not displayed. Those sprites are the number of lives left for player 1 in the upper left corner. The problems is more visible when you must enter your initials... Only the first 2 are shown. The 3rd initial and the [ENTER] character are invisible. See the screenshot... The entered name is CHB (see 11. line in the Bravest Aviators), but it is shown as CH in the corner !

If it is a newly discovered bug, I'll be proud to have found it, because bugs are really rare ;-)
If it's a configuration problem, tell me what to do... But I have the most A500 compatible configuration, and AFAIR, Battle Squadron worked very well on my old A500.

CUL8R

Eric C · 17 February 2005, 16:21

No, I didn't. I go and download it now. Then I'll try Battle Squadron in a few minutes, if WinFellow is simple enough ;-)

Eric C · 17 February 2005, 17:11

That's fun!!!

With WinFellow, my bug disappeared... but some new bugs appeared!!!

Look at the attachment, and compare with the previous one:

1. The left and right scores and missiles [M] are of the same color (left should be blue)
2. The Highest score is truncated !
3. The "2up" is replaced by some "lives" sprites.
4. The 2nd player score is truncated, and some digits are strangely shaped.
5. Some pixels missing in the 1st player's missiles.
6. Some missiles missing for the 2nd player.

Well! Both emulators seem to have problems to handle the display of this game!!!

It seems that there is too many sprites. That was OK for my good old A500, but WinUAE truncates some sprites it cannot manage, and fellow truncates some other sprites.

But now, I'm sure it's an emulator problem, not an ADF one. Thank you for the tip.

Now, Toni knows it is work for him ;-)

Toni Wilen · 17 February 2005, 19:34

Battle Squadron uses copper to write sprite data registers. (non-dma mode) and timing must be 100% cycle exact for sprites to work correctly (both copper and sprite state machine)

Some other examples: Leander (partially) and demo Total Triple Trouble.

Unfortunately sprite state machine is much more complex than hardware reference manual says..

FallenSeraphin · 17 February 2005, 20:44

Disable Picasso96 and work with Direct / Indirect JIT, should fix the corruption... Seriously, im beginning to think that WinFellow tops WinUae by a VERY LARGE margin...

MethodGit · 17 February 2005, 21:14

Quote:

Originally Posted by Toni Wilen

Battle Squadron uses copper to write sprite data registers. (non-dma mode) and timing must be 100% cycle exact for sprites to work correctly (both copper and sprite state machine)

Some other examples: Leander (partially) and demo Total Triple Trouble.

Unfortunately sprite state machine is much more complex than hardware reference manual says..

Could that also be the cause of the problem with the score display in Super Methane Bros (all versions)? I'm not sure if it's meant to show a lot of zeroes, but I do know at least one character in the score always looks dodgy.

TheAmigaMan · 17 February 2005, 21:39

Quote:

Originally Posted by Eric C

Hi,

I think I have found a bug within the emulation, with WinUAE 0.9.92.

When I play Battle Squadron, some sprites are not displayed. Those sprites are the number of lives left for player 1 in the upper left corner. The problems is more visible when you must enter your initials... Only the first 2 are shown. The 3rd initial and the [ENTER] character are invisible. See the screenshot... The entered name is CHB (see 11. line in the Bravest Aviators), but it is shown as CH in the corner !

If it is a newly discovered bug, I'll be proud to have found it, because bugs are really rare ;-)
If it's a configuration problem, tell me what to do... But I have the most A500 compatible configuration, and AFAIR, Battle Squadron worked very well on my old A500.

CUL8R

Why not just buy yourself another Amiga computer?

Eric C · 18 February 2005, 11:52

Toni:

Even with 68000 + more compatible, Match A500 speed, Cycle exact CPU + Blitter, Full collision, disk speed 100%, the problem is still there.
So, I guess that when you write about a timing to be exact, it is more complex than just checking "Cycle exact CPU + Blitter"!? Do you mean it is a known problem?

FallenSeraphin:

I'm emulating an A500. I never play with Picasso and JIT parameters. 68000 +ECS only!

TheAmigaMan:

I don't need to buy one... Mine is still alive... but far from me... One day, I'll go back to my parent's attic and wake it up ;-)

FallenSeraphin · 18 February 2005, 21:12

Dont forget that original (and most after) games worked with OCS, not ECS

... trust me this has fixed some games i couldnt get to work

Eric C · 21 February 2005, 11:27

I hadn't tried OCS yet...

Now, I've tried every combination of OCS, ECS Agnus with Kick 1.2 and 1.3... No change.

Well, in fact, that is a minor bug. I can still play with that bug. My aim was just to know if it is a bug or a mis-configuration. Now, I think that's a bug, and I reported it. I hope it will increase the quality of v1.0 ;-)

Thank you all!

andreas · 21 February 2005, 21:38

Quote:

Originally Posted by Eric C

Now, Toni knows it is work for him ;-)

Not necessarily.
Yet I'm a bit puzzled that both 'M's are of the same color in the 2nd screenshot you posted, whereas they're of a different color in the 1st screenshot on top of the thread.
Some of the problems you've reported might be a problem of the crack you've used. I could even assume that the crack you've used is actually a preview/beta and not the retail/sales version!
Please note that a lot of cracks from a lot of games do not support saving the game or highscore. (e. g. cracked Pinball Dreams, 90% of all versions of Prince of Persia)
Crackers are usually selfish, with a strong urge to eternalize themselves in the images.

Hence, most highscore is either auto-overwritten on startup or protected against overwriting, no matter what you've previously entered.

However, it might work with the original CAPS version perhaps.
(please do not ask questions about this, I have sent you a PM with all info you need.

)

As soon as you have got hold of this version by any imaginable magical way

, I strongly request you to recheck current WinUAE with this version whether some "bugs" (?) have disappeared (or indeed not).

Akira · 21 February 2005, 22:16

Quote:

Originally Posted by FallenSeraphin

im beginning to think that WinFellow tops WinUae by a VERY LARGE margin...

Be serious...

Eric C · 22 February 2005, 10:35

Quote:

Originally Posted by andreas

Yet I'm a bit puzzled that both 'M's are of the same color in the 2nd screenshot you posted, whereas they're of a different color in the 1st screenshot on top of the thread.

The second screenshot is from WinFellow!!! I think I was not really explicit about that.

I had a problem with 2 sprites in the upper left corner with WinUAE. Kodoichi suggested me to try with WinFellow. All other problems shown in the 2nd screenshot come from WinFellow.

In fact, that's why I think of an emulation bug... There is one bug with WinUAE, and with WinFellow, that bug disappeared but others appeared. Probably because they handle sprites differently... but not 100% accurately :-(

FallenSeraphin · 22 February 2005, 16:10

Quote:

Originally Posted by Akira

Be serious...

I have a very slow comp at work in wich WinFellow works a LOT better then WinUAE, so im very serious...

Eric C · 22 February 2005, 16:15

Don't feed the troll ;-)

andreas · 22 February 2005, 17:27

Quote:

Originally Posted by Eric C

The second screenshot is from WinFellow!!! I think I was not really explicit about that.

Not really, it was rather a decent mix-up, making it hard to follow for me. But nevermind

Quote:

I had a problem with 2 sprites in the upper left corner with WinUAE. Kodoichi suggested me to try with WinFellow. All other problems shown in the 2nd screenshot come from WinFellow.

Aaaahh...ok.
Well WinUAE DID show correct colors for me, that's what I was trying to point out. :-)

Quote:

Originally Posted by FallenSeraphin

I have a very slow comp at work in wich WinFellow works a LOT better then WinUAE, so im very serious...

You definitely are.
For instance, there are (still!) economically aggrieved regions in this world where people simply could never afford an Athlon XP machine for financial reasons (the former so-called 'Third World' in particular), so WinFellow - and in general, applications that only demand low resources - will always be a blessing for these people!!
I really like the fact that this gives more people a way to enjoy emulation, because they're not simply locked out!

Shatterhand · 23 February 2005, 02:51

Yes, 2 months ago all I had was a Pentium MMX 233 Mhz. And I only got this new computer because I was REAL lucky

My girlfriend's computer is a Pentium 133Mhz with 96 mb of ram.

I have a friend who has a Pentium MMX 233Mhz with 32 mb of ram

Completely off-topic, just showing that what Andreas says is really the truth

Eric C · 07 March 2005, 14:20

Small update...

I tried CAPS image. The "lives sprites bug" is still there. I guess it is not a bad ADF problem.

Toni Wilen · 07 March 2005, 14:39

Yes, this is chipset emulation problem. Sprite's X-coordinate is updated by copper slightly too late.

This is quite interesting discovery because in theory this should not work on real Amigas either. It seems this is yet another "secret" chipset feature..

Toni Wilen · 07 March 2005, 19:25

I think I found logical reason for this feature.. It may not be 100% right but it seems to work without breaking anything else (famous last words, I know..)

EDIT: This also fixes Bubba'n'Stix background mountain jumping so it can't be totally wrong fix

Fackamato · 09 March 2005, 06:34

Quote:

Originally Posted by Toni Wilen

I think I found logical reason for this feature.. It may not be 100% right but it seems to work without breaking anything else (famous last words, I know..)

EDIT: This also fixes Bubba'n'Stix background mountain jumping so it can't be totally wrong fix

17 February 2005, 15:13	#1
Eric C Registered User Join Date: Feb 2005 Location: Switzerland Posts: 10	Problem with Battle Squadron Hi, I think I have found a bug within the emulation, with WinUAE 0.9.92. When I play Battle Squadron, some sprites are not displayed. Those sprites are the number of lives left for player 1 in the upper left corner. The problems is more visible when you must enter your initials... Only the first 2 are shown. The 3rd initial and the [ENTER] character are invisible. See the screenshot... The entered name is CHB (see 11. line in the Bravest Aviators), but it is shown as CH in the corner ! If it is a newly discovered bug, I'll be proud to have found it, because bugs are really rare ;-) If it's a configuration problem, tell me what to do... But I have the most A500 compatible configuration, and AFAIR, Battle Squadron worked very well on my old A500. CUL8R Attached Thumbnails

17 February 2005, 17:11	#3
Eric C Registered User Join Date: Feb 2005 Location: Switzerland Posts: 10	That's fun!!! With WinFellow, my bug disappeared... but some new bugs appeared!!! Look at the attachment, and compare with the previous one: 1. The left and right scores and missiles [M] are of the same color (left should be blue) 2. The Highest score is truncated ! 3. The "2up" is replaced by some "lives" sprites. 4. The 2nd player score is truncated, and some digits are strangely shaped. 5. Some pixels missing in the 1st player's missiles. 6. Some missiles missing for the 2nd player. Well! Both emulators seem to have problems to handle the display of this game!!! It seems that there is too many sprites. That was OK for my good old A500, but WinUAE truncates some sprites it cannot manage, and fellow truncates some other sprites. But now, I'm sure it's an emulator problem, not an ADF one. Thank you for the tip. Now, Toni knows it is work for him ;-) Attached Thumbnails Last edited by Eric C; 23 February 2005 at 18:09. Reason: Not explicit enough about using WinFellow instead of WinUAE

07 March 2005, 19:25	#20
Toni Wilen WinUAE developer Join Date: Aug 2001 Location: Hämeenlinna/Finland Age: 38 Posts: 11,952	I think I found logical reason for this feature.. It may not be 100% right but it seems to work without breaking anything else (famous last words, I know..) EDIT: This also fixes Bubba'n'Stix background mountain jumping so it can't be totally wrong fix Last edited by Toni Wilen; 08 March 2005 at 15:18.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Battle Squadron special weapon	Camerhil	support.Games	15	13 August 2004 17:56
Req: Battle Squadron intro disk?	vertigo	request.Old Rare Games	7	26 December 2003 11:58
Battle Squadron NTSC	haynor	MarketPlace	3	17 June 2003 08:18
Rare: Battle Squadron 100% (2 disks)	rObOt	request.Old Rare Games	53	23 October 2002 00:29

17 February 2005, 16:21	#2
Eric C Registered User Join Date: Feb 2005 Location: Switzerland Posts: 10	No, I didn't. I go and download it now. Then I'll try Battle Squadron in a few minutes, if WinFellow is simple enough ;-)

17 February 2005, 19:34	#4
Toni Wilen WinUAE developer Join Date: Aug 2001 Location: Hämeenlinna/Finland Age: 38 Posts: 11,952	Battle Squadron uses copper to write sprite data registers. (non-dma mode) and timing must be 100% cycle exact for sprites to work correctly (both copper and sprite state machine) Some other examples: Leander (partially) and demo Total Triple Trouble. Unfortunately sprite state machine is much more complex than hardware reference manual says..

17 February 2005, 20:44	#5
FallenSeraphin Amiga Maniac Join Date: Feb 2005 Location: Portugal Age: 31 Posts: 55	Disable Picasso96 and work with Direct / Indirect JIT, should fix the corruption... Seriously, im beginning to think that WinFellow tops WinUae by a VERY LARGE margin...

18 February 2005, 11:52	#8
Eric C Registered User Join Date: Feb 2005 Location: Switzerland Posts: 10	Toni: Even with 68000 + more compatible, Match A500 speed, Cycle exact CPU + Blitter, Full collision, disk speed 100%, the problem is still there. So, I guess that when you write about a timing to be exact, it is more complex than just checking "Cycle exact CPU + Blitter"!? Do you mean it is a known problem? FallenSeraphin: I'm emulating an A500. I never play with Picasso and JIT parameters. 68000 +ECS only! TheAmigaMan: I don't need to buy one... Mine is still alive... but far from me... One day, I'll go back to my parent's attic and wake it up ;-)

18 February 2005, 21:12	#9
FallenSeraphin Amiga Maniac Join Date: Feb 2005 Location: Portugal Age: 31 Posts: 55	Dont forget that original (and most after) games worked with OCS, not ECS ... trust me this has fixed some games i couldnt get to work

21 February 2005, 11:27	#10
Eric C Registered User Join Date: Feb 2005 Location: Switzerland Posts: 10	I hadn't tried OCS yet... Now, I've tried every combination of OCS, ECS Agnus with Kick 1.2 and 1.3... No change. Well, in fact, that is a minor bug. I can still play with that bug. My aim was just to know if it is a bug or a mis-configuration. Now, I think that's a bug, and I reported it. I hope it will increase the quality of v1.0 ;-) Thank you all!

22 February 2005, 16:15	#15
Eric C Registered User Join Date: Feb 2005 Location: Switzerland Posts: 10	Don't feed the troll ;-)

23 February 2005, 02:51	#17
Shatterhand Junior Member Join Date: Jun 2001 Location: Brazil Age: 30 Posts: 1,637	Yes, 2 months ago all I had was a Pentium MMX 233 Mhz. And I only got this new computer because I was REAL lucky My girlfriend's computer is a Pentium 133Mhz with 96 mb of ram. I have a friend who has a Pentium MMX 233Mhz with 32 mb of ram Completely off-topic, just showing that what Andreas says is really the truth

07 March 2005, 14:20	#18
Eric C Registered User Join Date: Feb 2005 Location: Switzerland Posts: 10	Small update... I tried CAPS image. The "lives sprites bug" is still there. I guess it is not a bad ADF problem.

07 March 2005, 14:39	#19
Toni Wilen WinUAE developer Join Date: Aug 2001 Location: Hämeenlinna/Finland Age: 38 Posts: 11,952	Yes, this is chipset emulation problem. Sprite's X-coordinate is updated by copper slightly too late. This is quite interesting discovery because in theory this should not work on real Amigas either. It seems this is yet another "secret" chipset feature..

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)