Like the OP in this thread Decoupling input polling and logic from visual frame rate - #3 by diogocosta
I too am hoping to build a music /audio game where input latency and audio is the main focus. My research so far points to I need to write a little input handler in c or assembly code; Aiming for something like 5k to 10k interrupts per second is my plan.I will use the game to run my audio at live show and also research players I can jam with on tour. And to see the effects of high frequency monitor refresh also.
What does the title of the post have to do with the body?
And why do you need 5-10k checks per second? That’s as low as 0.1ms interval, seems like a complete overkill.
Out of curiosity, I dug around a bit and found a couple interesting facts related to this topic:
- human rhythm perception is in the range of 1-10ms,
- human average reaction time is about 150-250ms (doesn’t really matter that much for rhythm games though)
- popular rhythm games usually treat errors below 30ms as “perfect”,
- hardware input lag below 25ms is considered to be “good” in competitive gaming,
Having all that in mind, checking input every 0.1ms definitely seems like an overkill.
I’m really curious as to why you think that’s what you need, if you could elaborate a little.
Thanks!
4 Likes
Plus a good profesional external sound card has a delay of slightly above 1 ms in best conditions, a few ms on average (going from memory from work’s external recording USB soundcard setup)
Seems to me you’re setting up arbitrary limits that are totally unrealistic for consumer audio setups, as even pro external soundcards have delays numbered in ms, not fractions of ms.
Cheers !
Edit : as wchc said, wth is going on with that post title ?
1 Like
I’m guessing he’s trying to remove any input lag caused by his software. Doing 5,000-10,000 checks per second would accomplish that, but it’s a pretty brute-force solution. I feel like I’d need more information to figure out exactly what he’s trying to do, and what the best approach for him is.