Hi,
I'm working on some software tool which looks like this:
The idea behind this project is to make the filer design process (crossovers) as easy as dragging control points over frequency response plot.
Currently, there is digital engine, which calculate biquads for given control point F,dB,Q position, and then plots the FR of chain of biquads on screen, as you drag the control point.
There is a feature of importing FR data from file, so you can attach actual frequency response to every "output" block, and then play with it with control points.
There are many missing things, as presets - storage and recall. Project save/open, etc.
No export of data...
I'm working on reversing of DCX2496 presets file format - around 60% of the file is already reversed by me. Another 40% won't be an issue.
I'd like to add some "target build" devices, so you can "compile" your filter to
- DCX2496
- miniDSP
- RockBox EQ
- Other processors
- C sourcecode generator for different targets (by templates) (?).
- PC based DSP (i'd like to make this engine too, and link it to the StudioX for real-time settings simulation with multichannel soundcard).
And, probably, a set of classic analog circuits (from Linkwitz's site). This will need some more programming, as it
- should be analog simulation (already supported in code for plotting, not for digital filtering)
- some calculation of actual resistor and capacitor values.
Ahh, and i miss the feature of "summing" the signals to single FR plot.
I don't really know where to start with it.
Would sum of electronic transfer functions equal to acoustical sum of same transfer functions?
Hehe, i've already got 8dB increase in dynamic range in my car, jut by entering the EQ preset from the car to this program, drawing a same curve from scratch but with different PEQ points around 0dB, and then entering the new curve points back to car player.
I've tested the new preset, and it sounds the same as the old one, but 8db louder. Yay!
I'm working on some software tool which looks like this:
The idea behind this project is to make the filer design process (crossovers) as easy as dragging control points over frequency response plot.
Currently, there is digital engine, which calculate biquads for given control point F,dB,Q position, and then plots the FR of chain of biquads on screen, as you drag the control point.
There is a feature of importing FR data from file, so you can attach actual frequency response to every "output" block, and then play with it with control points.
There are many missing things, as presets - storage and recall. Project save/open, etc.
No export of data...
I'm working on reversing of DCX2496 presets file format - around 60% of the file is already reversed by me. Another 40% won't be an issue.
I'd like to add some "target build" devices, so you can "compile" your filter to
- DCX2496
- miniDSP
- RockBox EQ
- Other processors
- C sourcecode generator for different targets (by templates) (?).
- PC based DSP (i'd like to make this engine too, and link it to the StudioX for real-time settings simulation with multichannel soundcard).
And, probably, a set of classic analog circuits (from Linkwitz's site). This will need some more programming, as it
- should be analog simulation (already supported in code for plotting, not for digital filtering)
- some calculation of actual resistor and capacitor values.
Ahh, and i miss the feature of "summing" the signals to single FR plot.
I don't really know where to start with it.
Would sum of electronic transfer functions equal to acoustical sum of same transfer functions?
Hehe, i've already got 8dB increase in dynamic range in my car, jut by entering the EQ preset from the car to this program, drawing a same curve from scratch but with different PEQ points around 0dB, and then entering the new curve points back to car player.
I've tested the new preset, and it sounds the same as the old one, but 8db louder. Yay!
Last edited:
I've closed the loop between biquads in my program and biquads in foobar's DSP. Now when you move the control points on screen, you hear the filtered sound thru foobar with a slight delay of foobar's buffer.
Now it's time to add FR measurements involving FFT in realtime, like ARTA/etc, to show actual modified response recorded by microphone - so even deaf could make a speaker sound right
And - summed response. Unfortunately i'm clueless about phase measuremets of speaker drivers - which make the most difference when you summ the driver outputs in the air.
Now it's time to add FR measurements involving FFT in realtime, like ARTA/etc, to show actual modified response recorded by microphone - so even deaf could make a speaker sound right
And - summed response. Unfortunately i'm clueless about phase measuremets of speaker drivers - which make the most difference when you summ the driver outputs in the air.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.