Possible U-235 bug?

[JagChris]

If anyone has a minute I would like to see if they can recreate this behavior on their end so I can eliminate the problem isnt just on my end.

 

I have been playing 'Shoot Em Up' and ran across a strange sound anomoly. It doesnt happen all the time but I have been able to reproduce it twice.

 

Play Shoot Em Up and just hold the trigger down using the firehose method to kill everything. Even if you die just start again with the trigger down. Sometimes after a while it sounds like it is making the same shooting sound twice, and it comes off as thsi really profound RAK A TAK TAK sound. When it first happened I thought someone was outside with power tools or my fridge was freaking out. After a few minutes it will quit. Sometimes I can reproduce it and sometimes I can't. When I can get it to exhibit this odd behavior it happens around level 7-8 or so starting from level one so it takes a few minutes. Sometimes I can hear what seems to be a buildup to it in the sound as I go through the levels.

[+DrTypo]

I didn't manage to recreate the problem on my PAL system.

However I don't think it's a U23se bug. I think "Shoot'em up" isn't using it properly.

 

U235se proceeds commands lists.

Each time the game wants to play a sample, it issues a command list to play a single sample.

When testing the sound engine, I noticed that when the sound engine doesn't like when commands lists are issued too fast (I guess there must be a polling interval).

I think the proper way would be to enqueue all the "play sample" commands in a single command list and flush it every tick or two.

 

On a NTSC, it is more likely to have commands issued too fast than a on PAL system (60fps vs 50fps). This could be the explanation of the behavior you noticed.

 

I may fix "Shoot'em up" sometime but it's a not priority right now for me.

[LinkoVitch]

hi (Yep I still live )

 

It could be as you described DrTypo. The way the SE works is it checks to see if there is a list being pointed at and goes to play it, it then zero's out the pointer so to know it has completed the list. The pointer is checked whenever it's got some free time, so quite a lot given the DSP is spinning away at 25MHz

 

It could be that you are essentially restarting the same sample over and over again. I haven't touched my jag in must be 2 months so I haven't listened or tried this