ivanreshe

I'm currently trying to work out exactly how long delays ring on for, and ideally make a formula where you input Feedback %, Delay Time, and get the resulting 'ring' time in seconds or m/secs.

I've taken the logical approach of recording the ring time with different feedback and delay times, and worked out the "multiplier" in each case - by dividing delay time by ring time. I then matched up these "multipliers" with the feedback. The result I need would be a formula that can give me the "multiplier" from the feedback % - because then all I need to do is divide the delay time by this "multiplier" which would then give me total ring time for that delay and feedback % - bingo! (sorry if this sound a bit confusing, but it would work).

I made a list of these following results - however, the multipliers decrease exponentially. My A-Level maths days are way behind me, I would really appreciate any advice from someone hot on Max math! I tried playing with linedrive, but didn't really have any luck. Here are my results -

You can test these for yourself if interested - just play a short sample of sound through a delay time of your choice with the feedback % from the above table - divide the delay time by the multiplier and you have the total seconds of ring time from that delayed sound.

Feedback % - Multiplier
* * *
0.1 - 200
0.2 - 142
0.3 - 100
0.4 - 83
0.5 - 67
0.6 - 52
0.7 - 37
0.8 - 24
0.9 - 11

Ivan Reshetilovhttp://www.ivanreshetilov.co.uk


scaling-numbers-into-exponential-output-mathematicians-please-help

in case i understood you right, the "ring time" is nothing

it 1. depends on the input volume and 2. is it a matter of

the human recognition and/or the current speaker volume how

so in other words - taking the input volume into account,

in practice, i think, you should just try out what works

best, for example check log2 against log10 against log50

in case i understood you right, the "ring time" is nothing
what you could calculate.

it 1. depends on the  input volume and 2. is it a matter of
the human recognition and/or the current speaker volume how
"long" it rings.

so in other words - taking the input volume into account,
you would need 2 conversion tables.

in practice, i think, you should just try out what works
best, for example check log2 against log10 against log50
and see what seems to match best.

True - the input volume would be a factor.

However, with the results I have produced thus far, it was done with a very loud sample, and the speakers were turned up slowly as the delay faded, thus giving a general "maximum" value. For the purposes of what I am doing, this would be a great starting point.

Thanks for your suggestions, I'll try them out.

deltime * log( 0.001 ) / log( fb ) = decay time

Looks interesting. However, there seems to be a problem with your delay time eg. putting it to 1000 m/sec gives a delay of something like 150. Unless I'm missing something?

 However, there seems to be a problem with your delay time eg. putting it to 1000 m/sec gives a delay of something like 150. Unless I'm missing something?

"First argument sets max delay time in milliseconds"

in my example this is set to 150. just an arbitrary value you might like to change.

Quote: However, there seems to be a problem with your delay time eg. putting it to 1000 m/sec gives a delay of something like 150. Unless I'm missing something?

take a look at comb~.help: 
"First argument sets max delay time in milliseconds"

in my example this is set to 150. just an arbitrary value you might like to change.


Scaling numbers into exponential output - mathematicians please help!