- Joined
- May 6, 2021
- Messages
- 125
- Likes
- 581
As to on the site, don't know, but you can do it.
Unless one uses a measurement system that measures the time of flight (like TDS a process invented by the first person to address loudspeaker acoustic phase, the late dick Heyser), what one see's is the speakers phase shift combined with the time delay involved in the "time of flight".
As opposed to the phase shift involved in the "time of flight" it is the loudspeakers acoustic phase that indicates what the loudspeaker is doing.
Dick's way of describing acoustic phase was that it's "as if" the source were moving forward and back in it's location in time. If one has flat amplitude, one also needs flat phase around zero degrees in order to reproduce the input signal's wave shape.
How to do it?
A real danger with this is it's possible to make it worse as the only things you can fix are the things coming from the loudspeaker drivers.
Take a step ladder or other similar platform, put the speaker on it (and best if this is outside or even better a very large room). Put the measurement mic say a meter away on axis.
Use a gated measurement system like ARTA, establish a decent but not too loud level.
It is important to average a number of measurements, this kind of measurement has a weakness or two.
While every measurement looks real, these are susceptible to noise at the lower frequency end AND there is a limit to the valid data based on how long the window was in time. As a result, (and just like astrophotography cameras) each time you double the number of measurements / exposures you average, you cut that "noise" in half.
With ARTA, there is a yellow bar at the bottom that shows the point where below that, you can't trust the data. Generally one would look at the impulse measurement and looking at the ETC view, it's easy to see where the first reflection is so go back to the impulse and put the begining of the window right before the impulse and the end of the window right before the first reflection.
A result is a subwoofers response essentially can't be measured in a room with the sequence or MLS type impulse measurement systems and in that case, should be measured on the ground outside with the mic on a ground boundary. The TEF / TDS process using frequency window gating instead of time gating is better suited to this part if available
So once you have a measurement you need to assign the time in order to remove the phase wrap that the time of flight caused. In a normal multi way speaker, it may be hard to find where that is because of crossover phase shift.
Rule 1 low pass filters and driver response cause the most phase shift (time delay) and so adjust the time delay in the display so that the bass phase response below crossover is more or less flat where the upper part of the flat bass response is.
Save that measurement as a text file and then load it into Re-Phase.
RePhase can generate the impulse response you need for FIR correction and is like a nearly unlimited equalizer AND unlike normal EQ, one can adjust the phase independently of amplitude.
Once you have a curve you like (and have looked up what kind of FIR file you need), save that file and load it into your FIR filter.
Now a caution, the more closely the source resembles / measured like a single source in space (like a coax driver) the higher up you will be safe correcting the response etc.
You can still correct high up with conventional speakers with separate drivers but because of larger driver to driver spacing's (and so the distance form each source) the correction only applies to where the mic was so is generally not a good idea to go very high.. This is often the case at concerts where the sources are far apart and where it sounds good at the mixing board but most everywhere else stinks because making it better there, makes it worse elsewhere..
In 2014 I was interested to see what FIR filters could do and I bought an open DRC from mini-dsp to play with.
Here was what i got fooling around in my living room using it to correct the loudspeaker. Now this loudspeaker (the first two i made in 2005) didn't have phase shift from a crossover to begin with but the before and after is clearly different but did not sound as different as the curves.
Since then, i have used FIR filters on some of our recording studio stuff at work and some more "home" oriented like some prototypes I made last spring that Erin heard this summer.
Personally I think this stuff makes an audible difference when the other distractions are minimized.
Best,
Tom
Unless one uses a measurement system that measures the time of flight (like TDS a process invented by the first person to address loudspeaker acoustic phase, the late dick Heyser), what one see's is the speakers phase shift combined with the time delay involved in the "time of flight".
As opposed to the phase shift involved in the "time of flight" it is the loudspeakers acoustic phase that indicates what the loudspeaker is doing.
Dick's way of describing acoustic phase was that it's "as if" the source were moving forward and back in it's location in time. If one has flat amplitude, one also needs flat phase around zero degrees in order to reproduce the input signal's wave shape.
How to do it?
A real danger with this is it's possible to make it worse as the only things you can fix are the things coming from the loudspeaker drivers.
Take a step ladder or other similar platform, put the speaker on it (and best if this is outside or even better a very large room). Put the measurement mic say a meter away on axis.
Use a gated measurement system like ARTA, establish a decent but not too loud level.
It is important to average a number of measurements, this kind of measurement has a weakness or two.
While every measurement looks real, these are susceptible to noise at the lower frequency end AND there is a limit to the valid data based on how long the window was in time. As a result, (and just like astrophotography cameras) each time you double the number of measurements / exposures you average, you cut that "noise" in half.
With ARTA, there is a yellow bar at the bottom that shows the point where below that, you can't trust the data. Generally one would look at the impulse measurement and looking at the ETC view, it's easy to see where the first reflection is so go back to the impulse and put the begining of the window right before the impulse and the end of the window right before the first reflection.
A result is a subwoofers response essentially can't be measured in a room with the sequence or MLS type impulse measurement systems and in that case, should be measured on the ground outside with the mic on a ground boundary. The TEF / TDS process using frequency window gating instead of time gating is better suited to this part if available
So once you have a measurement you need to assign the time in order to remove the phase wrap that the time of flight caused. In a normal multi way speaker, it may be hard to find where that is because of crossover phase shift.
Rule 1 low pass filters and driver response cause the most phase shift (time delay) and so adjust the time delay in the display so that the bass phase response below crossover is more or less flat where the upper part of the flat bass response is.
Save that measurement as a text file and then load it into Re-Phase.
RePhase can generate the impulse response you need for FIR correction and is like a nearly unlimited equalizer AND unlike normal EQ, one can adjust the phase independently of amplitude.
Once you have a curve you like (and have looked up what kind of FIR file you need), save that file and load it into your FIR filter.
Now a caution, the more closely the source resembles / measured like a single source in space (like a coax driver) the higher up you will be safe correcting the response etc.
You can still correct high up with conventional speakers with separate drivers but because of larger driver to driver spacing's (and so the distance form each source) the correction only applies to where the mic was so is generally not a good idea to go very high.. This is often the case at concerts where the sources are far apart and where it sounds good at the mixing board but most everywhere else stinks because making it better there, makes it worse elsewhere..
In 2014 I was interested to see what FIR filters could do and I bought an open DRC from mini-dsp to play with.
Here was what i got fooling around in my living room using it to correct the loudspeaker. Now this loudspeaker (the first two i made in 2005) didn't have phase shift from a crossover to begin with but the before and after is clearly different but did not sound as different as the curves.
Since then, i have used FIR filters on some of our recording studio stuff at work and some more "home" oriented like some prototypes I made last spring that Erin heard this summer.
Personally I think this stuff makes an audible difference when the other distractions are minimized.
Best,
Tom
