Hey POS,
Could a FIR file be built that has coefficients built to regular impulse centering,
and also have coefficients that apply only to certain frequencies at a time later than the frequencies initially went thru?
I bet that made no sense at all...LoL..
Here's the example/goal....
Say a CD driver/horn has a 2000Hz reflection head back into the horn throat caused from discontinuity in the horn mouth.
That reflection back to the throat then rebounds back out the horn in delayed fashion ( vs the ongoing signal).
Is there a way that the 2000Hz reflection could be be nulled with a small out of polarity 2000Hz 'injection' at the appropriate time?
Could a FIR file be built that has coefficients built to regular impulse centering,
and also have coefficients that apply only to certain frequencies at a time later than the frequencies initially went thru?
I bet that made no sense at all...LoL..
Here's the example/goal....
Say a CD driver/horn has a 2000Hz reflection head back into the horn throat caused from discontinuity in the horn mouth.
That reflection back to the throat then rebounds back out the horn in delayed fashion ( vs the ongoing signal).
Is there a way that the 2000Hz reflection could be be nulled with a small out of polarity 2000Hz 'injection' at the appropriate time?
The "FFT length" parameter in rephase is only used for internal calculations and mainly impacts result visualization apparent resolution, and iterative optimization behavior (to a point). What matters most is the number of taps, and this is what Thierry describes.
You can leave FFT length at its default setting and not worry about it.
I should probably just hide this parameter in future versions... 😉
OK so if I generated two filters where the only difference was the "FFT-length", both files would be identical?
//
//
Hi Mark
Never tried this myself, but from the top of my head that should/could be doable by mixing (with sox, audacity, etc.) the initial FIR with a second one generated from identical settings but with an added delay, inverted polarity, and EQ/gain to mimic what really goes back to (and off!) the throat...
Good luck measuring and identifying the proper parameters though 😀
Last edited:
This sounds the same as the virtual bass array approach that I have seen (and have implemented myself) using acourate, details in Virtual Bass Array | Digital Room Correction
I would be slightly surprised if this is feasible/useful at 2kHz though
Good luck needed !! Lol
But hey, this paper just linked by brinkman in another thread gives me hope 🙂 http://mariobon.com/Articoli_storici/Horns_measurements_ETF2010d.pdf
Thanks POS
@3ll3d00d....cool, must study your post
Very interesting! Is this your website and system?
That makes me want to experiment with all this!One more feature on the todo list for rephase 1.5.0 I guess 😀
You can think of addressing what is reflected from the mouth back to the throat and remitted from there, but you cannot address what is directly diffracted /emitted from the mouth (or any discontinuity in the profile for that matter) to the listener, because this will change from one position to another.
Unfortunately the latter is probably more audible than the former.
Hi Mark:
This has come up in some of the Synergy horn threads but just at concept level. Read the EAW whitepaper on Gunness filtering and you will see some of the devilish details as well as the potential. If you go that way, I think you will end up with an array of reflection coefficients (in the frequency domain) instead of a single constant applied with a time delay.
Hi nc535,
Yep, thx. I should visit the synergy threads...maybe a good next foray....
I've been trying to dissect the papers and patents Dave G has on the Fulcrum website, on and off, for some time....with not so much success Lol.
I think I also better stop to learn trace arithmetic....particularly how to put FIR files together.
Apart from the REW help files on trace arithmetic, anybody know of some good primers??