One thing I should try is doing an autocorrelation on the IR. That may be a method to automatically find the first reflection. Actually an autocorrelation on the data squared would probably be better as it would remove any inversions (although I think that wall reflections are always in phase).
The problem with that is that you'd need much longer IRs to capture the full reflections, no room decays in 170ms and probably not even in half a second.
So, better to chop the partial reflection and model that part with a Reverb room block?
Yep, surely more efficient, easier to achieve and gives you more freedom on the room tone (if room is baked into the IR you can't change it afterwards)
That’s one thing I really dug about the UA Ox Box was how the room sound could sound so natural when added to taste.
What @FractalAudio said.
If I am following you, are you saying you might write a function to identify early reflections in an IR and then perhaps expose this to users in the Cab Block with a button to automatically trim IRs to remove reflections?
with all the craze around farfield/longer impulses I have honestly been trying my hardest to like them but secretly inside -they just do nothing for me. Thought there might be something wrong with me (ok, ok, I know - there's plenty else wrong w me). I'm admittedly a bit relieved to hear someone suggest it's ok not to like them. Still not going to admit that on the gear page!
But even if they are, its still just a moment in time captured and presented as an eq curve. If you wait a split secound to sample or move the mic even slightly, you get a different curve. It seems to me that a more acurate curve would be some degree of averaging. Who hasn't run an eq after or even in parallel to a cab block. Im agreeing with you in terms of artifacts in the ir. I don't think i want them. I am further stating that a given ir for a cabinet may be better or worse than another one specifically because they are such specific and narrow slices of the eq curve. What would happen if you took a curve and smoothed to 1/3 octave or more? Warning...just thinking out loud here.
Yes i did. Im also saying, or humbly suggesting actually, that we dont need all that detail in an ir. That detail does give an ir its specif sound, and there is a place for that. But i am wondering if you would get a better basic ir by averaging out some of that detail.What @FractalAudio said.
Also, a speaker will excite the room at any volume level. You just made a case for reflection-free IRs.
Why should there be a thousand different unique irs for a particular cab? Couldn't there be a specific eq curve that represents a particular cab? For example a specific speaker has a specific eq curve. Things can alter it like mic position, room, ect. It seems that we would want all of that out of the equation so that we could ad room, mic, ect, as we desire. Isnt that the ultimate goal?
That's what the Smoothing parameter is for.
Considering the size of most studios and the resulting reflection timing, are there any examples of close-mic’ed IRs in the Fractal library that you’d still consider using in ultrares?
Here's the cool thing, though: If you have an accurate impulse response, you can recreate the EQ curve precisely. Likewise, if you have an accurate frequency response, you can recreate an accurate impulse response.
If you wait a split second before sampling the response, you're ignoring part of the response. That's why you don't do that.