Roman Thilenius

there is something weird going on in my old doppler effect.

say if you limit the distance parameter to 0-100 meters, shouldnt it be enough to have a tapping buffer of 292 milliseconds size, no matter where a movement starts from?


there is something weird going on in my old doppler effect.

say if you limit the distance parameter to 0-100 meters, shouldnt it be enough to have a tapping buffer of 292 milliseconds size, no matter where a movement starts from?




there is something weird going on in my old doppler effect.

say if you limit the distance parameter to 0-100 meters, shouldnt it be enough

doppler-effect-theory

Hi Roman. It interests me: the delay depends of how much the object approaches. From 2000 to 1000, 292 ms. Not so easy for me to think that: a vibrato or a pitchbend is the same narrative thing. A derive in some ms can make move a sound in a soundscape. With some eq and level work, for my ears, some tiny doppler changes much the illusion of movement in a panoramic, without to change the musical message.

Hi Roman. It interests me: the delay depends of how much the object approaches. From 2000 to 1000, 292 ms. Not so easy for me to think that: a vibrato or a pitchbend is the same narrative thing. A derive in some ms can make move a sound in a soundscape. With some eq and level work, for my ears, some tiny doppler changes much the illusion of movement in a panoramic, without to change the musical message.
best to you
michel z


Hi, Roman.
Assuming 344 m/s for the speed of sound:
1000m / 343m/s = 2.907 s (2907 ms)

the point is that i have a feeling that it sonically doesnt work - it almost tricked me into thinking that i´d need double the space in order to be able to perform moves in both directions. but no. :)

i am interpolating the incoming modulation signal (0-100 meters MSP signal) with rampsmooth and i do the same for the absorbtion filtering and the overall magnitude. for slow moves it sounds as it should, for faster moves it stops working. 

that interpolation is highly suspicious, too, but i first wanted to make sure that this buffer size thing is okay.

btw, i have forgotten to add the minimum tap delay / vector size to the 292 ms, but that should not matter much.


the point is that i have a feeling that it sonically doesnt work - it almost tricked me into thinking that i´d need double the space in order to be able to perform moves in both directions. but no. :)

i am interpolating the incoming modulation signal (0-100 meters MSP signal) with rampsmooth and i do the same for the absorbtion filtering and the overall magnitude. for slow moves it sounds as it should, for faster moves it stops working. 

that interpolation is highly suspicious, too, but i first wanted to make sure that this buffer size thing is okay.

btw, i have forgotten to add the minimum tap delay / vector size to the 292 ms, but that should not matter much.


"Not so easy for me to think that: a vibrato or a pitchbend is the same narrative thing"

yeah, one might think so, but in fact ist not.

if you make a vibrato effect - and the input is a track read from disk sample by sample - or live microphone input - you´d just start to bend down first (i.e. before you go up) and all is good. you dont need to introduce an overall delay of max pitch bend / 2 in order to synchronize the output with the input.

now in a doppler, if the starting position is 700 meters, the input is already delayed likewise, as this is part of the effect anyway. the room - literally - to shorten the delay is already there.

but i wonder if i should maybe limit the modulation speed to the speed of sound? :D that could be clever.


"Not so easy for me to think that: a vibrato or a pitchbend is the same narrative thing"

yeah, one might think so, but in fact ist not.

if you make a vibrato effect - and the input is a track read from disk sample by sample - or live microphone input - you´d just start to bend down first (i.e. before you go up) and all is good. you dont need to introduce an overall delay of max pitch bend / 2 in order to synchronize the output with the input.

now in a doppler, if the starting position is 700 meters, the input is already delayed likewise, as this is part of the effect anyway. the room - literally - to shorten the delay is already there.

but i wonder if i should maybe limit the modulation speed to the speed of sound? :D that could be clever.


Roman, 1000 meters equals 2907 ms, not 292 ms.

Yes, the direction is irrelevant, only the distance between source and listener is needed.

I think that for filtering, a low pass filter such as onepole~ could work better than rampsmooth~.

Roman, 1000 meters equals 2907 ms, not 292 ms.
Yes, the direction is irrelevant, only the distance between source and listener is needed.
I think that for filtering, a low pass filter such as onepole~ could work better than rampsmooth~.

yeah that was a typo; i´ve corrected the start post.

btw, i have made a no-doppler version of the same thing (using simple vdp windowing) and that is yet another case where the filtering and the amplitude go out of sync with the theoretical distance if you modulate it faster.


yeah that was a typo; i´ve corrected the start post.

btw, i have made a no-doppler version of the same thing (using simple vdp windowing) and that is yet another case where the filtering and the amplitude go out of sync with the theoretical distance if you modulate it faster.


fun to think objects with a speed over Mach 1 and their associated bang

thinking to a real stage, with a frontal stereo system, a little delay can be used to let the apparent sound source in real place behind the face. In my mind, it's a frontal scenography in 2d, something like a stereo file...

oups clearly: 1 km =~3 s...
fun to think objects with a speed over Mach 1 and their associated bang
thinking to a real stage, with a frontal stereo system, a little delay can be used to let the apparent sound source in real place behind the face. In my mind, it's a frontal scenography in 2d, something like a stereo file...


doppler effect - theory