When playing around with the ConvoBoy, FreEQ Boy and other Convolution Plugins I came to the idea, that maybe the process could be reversed. Now, I dont know if you could do that technically - but it would be cool if you could "remove" the room information of a recorded audio file (instead of applying it).
There are so many people out there (in home or semi-pro studios) that record under quite bad room-situations: a lot of unwanted reflections, unwanted frequency changes by room characteristics. Also a lot of microphones are used that attenuate (or cut) certain frequencies.
Now if there was a plugin that has an built in noise-generator (or whatever) to produce an impulse, and I would use the microphone I ususally make my recordings with to record that impulse back into my DAW - would it be possible to use this impulse response to reverse the convolution process (and additionally maybe something similar to FreEQ to clean up the frequency cuts / attenuations) to REMOVE all room / reverb / space / frequency changes that happen on the signal path?
I mean could it remove the reflections of my room, the special frequency characteristics of my room, the special frequency characteristics of my microphone and of my AD units and cables to get a real clean signal out of my very "influenced" recording??
That would be somehow awesome...but impossible maybe...
Greetings
Codex
Idea for a quite unique Plugin: DryBoy
-
- KVRist
- 243 posts since 17 Sep, 2006
-
- KVRian
- 951 posts since 11 Jan, 2004 from Netherlands
Technically speaking, and unsurprisingly, the process is called de-convolution. This is the process of removing instrumental imperfections by knowing the transfer function of the system.
As a start, if you know the frequency response for different recording positions in your room, you could dial the inverse of this into FreEq Boy and correct the frequency amplitudes.
Reverb is more difficult because it requires to match the timing of the de-convolution with the input signal - this is very difficult.
As a start, if you know the frequency response for different recording positions in your room, you could dial the inverse of this into FreEq Boy and correct the frequency amplitudes.
Reverb is more difficult because it requires to match the timing of the de-convolution with the input signal - this is very difficult.
-
- KVRist
- Topic Starter
- 243 posts since 17 Sep, 2006
Ok, I see. But how can I get a exact frequency response of my recording situation? As far as I know usually the EQ has some sort of noise generator to send out an impulse that is recorded back through the recording path.
From then on FreEQ already has everything that we need:
- Using the impulses eq curve to merge it onto the eq curve of the recorded noise. Then save the resulting EQ curve and apply it on future recordings...
Would that work? However I would need some sort of useful impulse that has all frequencies...?
Greetings
Codex
From then on FreEQ already has everything that we need:
- Using the impulses eq curve to merge it onto the eq curve of the recorded noise. Then save the resulting EQ curve and apply it on future recordings...
Would that work? However I would need some sort of useful impulse that has all frequencies...?
Greetings
Codex
-
- KVRist
- 176 posts since 17 Aug, 2006
Many have tried this but with not much result so far. The reason is that an untreated room is ringing in its fundamentals, and also there is comb filtering due to the reflections of the walls.
If you have a problem that comes from a wave that is reflected from the wall and causes a dip in your frequency response then you can simply not cure this by increasing the volume at this frequency. Because you do not only amplify the original wave but also the returning wave from the wall, so with other words the room fights back and is always much stronger than the EQ.
However if you have a well-treated room already then EQing can help to smooth out the rest.
If you have a problem that comes from a wave that is reflected from the wall and causes a dip in your frequency response then you can simply not cure this by increasing the volume at this frequency. Because you do not only amplify the original wave but also the returning wave from the wall, so with other words the room fights back and is always much stronger than the EQ.
However if you have a well-treated room already then EQing can help to smooth out the rest.
Maybe doing jokes without the use of smilies ...
-
- KVRAF
- 10815 posts since 26 Nov, 2004 from UK
-
- KVRAF
- 9590 posts since 17 Sep, 2002 from Gothenburg Sweden
I'd say this is as close to impossible as makes no difference but after seeing that Melodyne Direct Note Access i don't know what to believe in anymore.
-
Music Engineer Music Engineer https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=15959 - KVRAF
- 4379 posts since 8 Mar, 2004 from Berlin, Germany
the impulse- or frequency-response of your room does not only depend on the room itself but also (critically) on the positioning of the microphone and the sound source. in order to compensate that via deconvolution, you would have to exactly(!) reproduce these as they were in the setting where you measured the impulse response - which is not practical at all. to get rid of the comb filtering effects introduced by the nodes of stading waves, you could use a boundary microphone:Alphacodex wrote: I mean could it remove the reflections of my room, the special frequency characteristics of my room...
http://en.wikipedia.org/wiki/PZM_(microphone)
...to get rid of reverberation, a very close distance between mic and sound-source is desirable.
the linear effects (the filtering part) of those things on the other hand should be compensatable via deconvolution, the special frequency characteristics of my microphone and of my AD units and cables to get a real clean signal out of my very "influenced" recording??