Pierre Alexandre Tremblay

Thanks to Rodrigo Constanzo we have the first video tutorial of how to use and abuse the fruit of the HIRT (the HISS Impulse Response Toolbox). This time round, it is the basic use of capturing an impulse response with some of the tools for MaxMSP.

The paper and all the externals are still available here:

ps for those who prefer youtube, it will soon be up there too!

a-quick-tutorial-video-on-how-to-create-an-impulse-response

This is a great video! I am looking forward to seeing more,

I would also love to see some videos on using these

tools to create new timbres. For example live instruments

This is a great video! I am looking forward to seeing more,
especially room correction.

I would also love to see some videos on using these
tools to create new timbres. For example live instruments
convolved with interesting sounds.


thanks for the props. Now on YouTube if you prefer ;-)

http://www.youtube.com/watch?v=0waUi6qpf-k

and I n00bishly thought [yafr] was all I needed. Excellent work guys.

thanks. Would anyone fancy having the patch posted here?

Yes please I'd love the patch from the video. Great, clear presentation; and great externals.

Definitely more tutorial vids coming. The next one will be speaker/room correction stuff, and a couple down the road will be creative and non-linear convolution stuff.

Trying to figure out how to extract the IR into a stereo wav file to be used in other plugins (like max 4 live Convolution Reverb Pro). Cant figure it out :(

That's what I would be very interested in as well. I'm going to check out the patches now and see what gives. I need a handy IR creation thingy, since buying Altiverb is just out of the question for me right now and you can basically do anything with Max if you try hard enough... Any tips from the smart people, on how to extract a stereo .wav ? :)

http://anechoic.8088.org/mr_hay/wordpress/?p=635

Ok Sidora, I made an OS X .app which I hope does what you want it to. You should be able to just run this and work away without even having to boot Max, and save .AIFF's as you go along. It's merely an adaptation of the Max4Live patch which is included with Live and %99.99999 of the credit should not go to me. But rather the HISS crew at The University of Huddersfield.

Let me know what you think and the download link is towards the bottom of my post .

http://anechoic.8088.org/mr_hay/wordpress/?p=635 

This patch is cool and simple but alas I am a confused whether Irmeasure should be filling the buffer for the references or I have to record separately into the buffers from my microphone? When I run the patch the buffers do not fill. Thanks

This patch is cool and simple but alas I am a  confused whether Irmeasure should be filling the buffer for the references or I have to record separately into the buffers from my microphone? When I run the patch the buffers do not fill. Thanks

I want to create ambisonic IRs for an ambisonic convolution reverb.

The process of using the HISS tools (irmeasure) in real time and then converting the IRs to b format is pretty straight forward = no problem here.

What really is giving me problems is post processing of recorded material. Due to time issues I recorded sweeps through an ambi mic (ch 1,2,3,4), and wanted to do the IR creation in post, BUT... I can easily deconvolve the recorded files with buffconvolve, but this gives me very long files, which do not work well as IRs. So the question would be, which best practice with the HISS toolbox to extract an IR from an already recorded file (same sweep of course = having the reference sweep)??? Any insight would be very very helpful .

thanks, also an "I have no idea" is appreciated :-)

Reviving an ancient topic. 
I want to create ambisonic IRs for an ambisonic convolution reverb.
The process of using the HISS tools (irmeasure) in real time and then converting the IRs to b format is pretty straight forward = no problem here.
What really is giving me problems is post processing of recorded material. Due to time issues I recorded sweeps through an ambi mic (ch 1,2,3,4), and wanted to do the IR creation in post, BUT...  I can easily deconvolve  the recorded files with buffconvolve, but  this gives me very long files, which do not work well as IRs. So the question would be, which best practice with the HISS toolbox to extract an IR from an already recorded file (same sweep of course = having the reference sweep)???  Any insight would be very very helpful .

no other idea than "use a spike instead of a sweep, then all this is easier"

definitely worth trying. I also cropped the deconvolved resulting file to a rough estimate of the room decay time, and with a spike it is a lot easier to measure something like this too. The thing is I would love to be a bit more "scientific", let us say rigorous in the measurement .

of course the convolution with sweeps should end up with a file of the same lenght.

but in my opinion getting the first reflections/phase right is more interesting than a proper frequency representation, so it doesnt really matter that a spike might not give you a spectrum as clean as a sweep, especially with cheaper speakers.

as long as you know how to place stuff and dont catch a policesiren or a muezzin call with your recording all is good.

of course the convolution with sweeps should end up with a file of the same lenght.

but in my opinion getting  the first reflections/phase right is more interesting  than a proper frequency representation, so it doesnt really matter that a spike might not give you a spectrum as clean as a sweep, especially with cheaper speakers.

as long as you know how to place stuff and dont catch a policesiren or a muezzin call with your recording all is good.



It's been a while since I play with this stuff in detail, but I think irextract~ is what you're after, as you can reprocess the recorded files/sweeps with it.

Rodrigo thanks, I tried with irextract~ already, though being the recorded sweep of 45 sec. I keep getting an error message from irextract~ "buffer is not long enough for requested output length" (Yes I have tried changing the extracting buffer destination's length AND I have inputted the exact same parameters I have used in irmeasure ~ ). 

I do get an IR, a very short output file, even if I am expecting a 5sec length, but what also puzzles me is the resulting amplitude of a sfile playing through multiconvolve, of which amplitude I need to multiply by 30 to get a decent signal output; do you suggest normalizing the IR resulting file? With the spike method (as mentioned by Roman too) I am indeed ready to go. btw I am using a quite powerful amplified speaker (meyer sound -> 136 dB spl @ 1m), not as linear as I wanted though, and an Ambeo Mic.

Oh well it will be an interesting Saturday... ahahha.

Rodrigo thanks, I tried with  irextract~  already, though being the recorded sweep of 45 sec. I keep getting an error message from  irextract~  "buffer is not long enough for requested output length"  (Yes I have tried changing the extracting buffer destination's length AND I have inputted the exact same parameters I have used in irmeasure ~  ). 
I do get an IR, a very short output file, even if I am expecting a 5sec length, but what also puzzles me is the resulting amplitude of a sfile playing through multiconvolve, of which amplitude I need to multiply by 30 to get a decent signal output; do you suggest normalizing the IR resulting file?  With the spike method (as mentioned by Roman too) I am indeed ready to go. btw I am using a quite powerful amplified speaker (meyer sound -> 136 dB spl @ 1m), not as linear as I wanted though, and an Ambeo Mic.
I am most definitely missing something. 
Oh well it will be an interesting Saturday... ahahha. 

Thanks for this post. I've download the patch posted by Rodrigo (thanks) but I can't figure out how to actually play and record the sweep. On my machine nothing happens once I've set the sweep settings. I'm puzzled. Any help much appreciated.

Thanks for this post.  I've download the patch posted by Rodrigo (thanks) but I can't figure out how to actually play and record the sweep. On my machine nothing happens once I've set the sweep settings. I'm puzzled.  Any help much appreciated.

Hmm. If you hit that bang in the top left (with the Sweep settings), you should see a bunch of stuff happening and depending on how you have your audio configured, you should hear it too.

HI, im new to using this tool and am having trouble implementing the output of irmeasure. After running the sweep, "measurement done" is turning green but there is no indication in the graph that the new references have been stored (graph is empty). similarly, when click "set references" in the audition portion of the patch after running the sweep, the resulting wet audio is silent. I'm sure this is something simple and any pointers are appreciated.

HI, im new to using this tool and am having trouble implementing the output of irmeasure. After running the sweep, "measurement done" is turning green but there is no indication in the graph that the new references have been stored (graph is empty). similarly, when click  "set references" in the audition portion of the patch after running the sweep,  the resulting wet audio is silent.  I'm sure this is something simple and any pointers are appreciated. 

a quick tutorial video on how to create an impulse response