Home Theater Forum and Systems banner

Asio Convolution Plug-In

1849 Views 7 Replies 3 Participants Last post by  AudiocRaver
Hey guys,

its my first post, so please be kind :)

I have a question, but first Ill give a short overview of the starting point.

Like almost anybody i have two speakers in a room, and want to listen to music from my computer. I set up the stereo triangle on the best (measured) position, tried to work with early reflections points, room modes and the overall reverberation time with absorption and flutter echoes with diffusion. I dont have neighbors anymore, because of all the measurements i did ;-)

This is the point where i start to think about processing the audio signal. Kick me when Im wrong, I think convolution is better then equalising, because its possible to work on time domain as well as on frequency domain.

With the software rePhase i build an IR that corrects just some small things in phase and frequency. I can use the IR in a audio player oder in my DAW. The music now sounds different, and again i would like to measure the signal.

And here is the question:
Is a convolution software or plugin available, that let me use an IR for the complete audio out signal? So it would be possible to measure the effect of the IR with REW?

I hope its understandable. Thanks a lot
Ingobingo
1 - 8 of 8 Posts
You will need a VST rack, like LiveProfessor by Audiostrom, and a convolution plugin, of which there are many possibilities. AudioProfessor has both a free version and a paid version ($85), the free one might do everything you need. There is a free convolution plugin called convolver which is about 8 years old but still very much in use and very popular. There are lots of paid for convolution plugins, including Convo Boy, by Elevayta, which is inexpensive ($30) and low latency, there is also a $1 version. Elevayta also has a $1 VST rack for use only with the Pro paid versions of their plugins.
Thanks AudiocRaver, LiveProfessor works well as VST rack. Not yet with REW, but maybe multi-client mode on the sound device will fix it. Ill try soon.

The Convolver plugin works well with the Foobar Player, and allows to quick A-B the original signal with the signal "through" the Impulse response. Other free Convolution plugins may be NadIR, KeFIR or LeCab2.

Ive tried the free DRCDesigner software from Alan Jordan to measure the setup and create an Impulse response. Here again i cant see what the IR is doing.

The Impulse responses created with rePhase can be easily imported in REW, and they seem to do what is expected.

Has anyone experience in this direction?
If I understand your question correctly, One way to see the result of an impulse in a convolution engine would be to use a recorded REW sweep in fubar, play it back through the convo convolver with the impulse, into REW for analysis. It can be done, but it gets a little tricky getting all of that to happen and work well together in one computer. I have done it, but it can take awhile to get set up and working right, & I have basically stopped trying to do that and just use two computers to accomplish it, then you do not need the recorded sweep, you just use REW as the source and analyzer in one computer and foobar with the Converter in the other.

Does that answer your question?
I read the thread to make sure I understand what you are after. REW will do the signal analysis that you need, I believe, but you have to feed it the right kind of signal, meaning its own measurement sweep. That is the only source it understands property.

Get the latest release of REW and with the waveform generator record the sweep to be used for analysis. To do this properly, set your sweep parameters, rate , beginning and ending frequencies, and run a sweep with the settings so the signal generator receives those values, and pull up the signal generator and use the record function to record that sweep. It's beginning and ending values will not be exactly the same as this sweep you just used, that is the way that rew needs to work.

Now use your DAW to add several little timed blips before the sweep starts so you will know ahead of time when to begin the analysis function in REW. Run that recorded sweep through the convolver playing it with foobar, and into REW and it should do what you want.
See less See more
Happy New Year!

Yes, that is what i want to do, to measure what an IR is doing.

The setup with two computers, like you wrote, will be the best choice to get a good result. Maybe another way is to set the designed IR as calibration file in REW, make a measurement and then remove the cal file.

I did some tests and messed around with rePhase and the REW EQ Section. REW is a comfortable tool to get filters for a specific target, rePhase allows to build filters without touching the phase response. Both tools let me export IR as wave files. On the other hand IRs can be imported in REW.

Which brings me to the question: what can be filtered with an finite impulse response? To get the best starting point it may be helpful to get an average of maybe 5-8 measurements around the listening position.
Then i can reduce some overall peaks. I guess holes cant be eq. (where the group delay is high?)

Is it possible to correct the decay time for room modes, to get say 100Hz from 0.6 seconds to 0.4?
See less See more
You could measure that if you had JRiver ($50, has a 30 day free trial). It has an WDM driver and you can route REW's test sweeps through it and it has a convolution engine which would be applied if using the WDM driver.
Happy New Year!

Yes, that is what i want to do, to measure what an IR is doing.

The setup with two computers, like you wrote, will be the best choice to get a good result. Maybe another way is to set the designed IR as calibration file in REW, make a measurement and then remove the cal file.

I did some tests and messed around with rePhase and the REW EQ Section. REW is a comfortable tool to get filters for a specific target, rePhase allows to build filters without touching the phase response. Both tools let me export IR as wave files. On the other hand IRs can be imported in REW.

Which brings me to the question: what can be filtered with an finite impulse response? To get the best starting point it may be helpful to get an average of maybe 5-8 measurements around the listening position.
Then i can reduce some overall peaks. I guess holes cant be eq. (where the group delay is high?)

Is it possible to correct the decay time for room modes, to get say 100Hz from 0.6 seconds to 0.4?
FIR filters are not so great with room modes, where you want sharp filters with deep cuts sometimes. Pre-ringing tends to be worse with
-steep filter slopes
-large amounts of boost or cut
-at lower frequencies
which is exactly what you are doing with room modes.

Smaller cuts and minimal boosts at mid & high frequencies, with lesser slopes, tend to not be a problem in this regard.

You are correct about not trying to fill any deep, sharp holes, that must be done acoustically.
1 - 8 of 8 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top