Welcome! Need support, you got it. Or share your ideas and experiences.
Return to Planar Speaker Asylum
68.20.184.198
In Reply to: RE: digital roomcorrection posted by josh358 on January 18, 2011 at 08:35:38
you can set the lower limits on the correction filters by dragging the limit boxes around on the correction window. i don't correct under 200hz automatically. it's too room dependent. just adjust the level for your bass amp and smooth out the room modes.
Follow Ups:
OK, sure, that would solve the problem. Guess you can't correct for group delay though . . .
group delay at those low frequencies would be made apparent by dips and peaks in frequency response. there shouldn't be too many of them so you should be able to get them by hand. the critical frequencies are corrected automatically for group delay. in my filters this is around the crossover frequencies. the critical one being the crossover between the mid panel and the ribbon. i suspect there are several hundred corrections in those frequencies to keep the delay even.
That's what they say on their website (several hundred corrections).
that's pretty funny. i was just guessing. thousands are probably too many.
i understand what they're doing and it works really well. i've taken verification measurements with etf5 and a calibrated microphone, with very repeatable results.
That's impressive, since as I understand it many room correction systems have repeatability problems at higher frequencies.
AkuAnka has a point, Lyngdorf's own speakers are dipoles. I certainly did not expect the room correction to work with 3.6's, but it does to my ears. The system says it is making 20% adjustment only, so it doesn't seem to be working overtime. Also, surely the DEQX system, while seeking to adjust the speakers response only, still works by measuring the in room response. Doesn't the same problem then arise with dipoles? finally I believe the Lyngdorf system is now quite distinct from the TACT one.
DEQX, as I understand it, takes two sets of measurements, a gated nearfield one to derive compensation curves for frequency and phase anomalies in the loudspeaker itself, and one at the listener's position to compensate for room anomalies.
that's exactly how it works. the parametric eq is for "room correction" and the filters are for "speaker correction". once the speakers are corrected, you still have to get the mid and high frequencies to your ears without messing the whole thing up. if you have early reflections there is no "room correction" that can fix that. you might be able to smooth the frequency response a little but it will still be comb filtered and not as realistic.
Yeah, I don't see any way you can compensate for comb filtering above a certain frequency, since it will change as you move your head relative to the room surfaces. So the best course seems to be to limit most room compensation to low frequencies. Even then, the corrections will only be good at one point in the room or will represent a compromise over a larger area.
that is why i like very narrow cuts for room correction. test tones help you find the exact frequency of the problem modes, but music is not test tones. even a bass note played by a musical instrument with the exact same fundamental frequency will sound much closer to "normal" with only that narrow cut.
I don't know about you, but I've always found that moderately broad equalization can do more violence to the sound than no equalization at all. Newer research says that we're much more sensitive to low Q than high Q resonances -- the opposite of what I would have thought -- and I suspect that that's the reason. It seems that equalization has to have a 1/12 octave resolution to be effective.
1/12 octave is the narrowest choice in the deqx parametric eq. that's exactly what i meant by narrow.
Do they also give you manual control over group delay?
if you mean by channel, then yes. you can apply a delay to each channel on the crossover independently. you can't pick a group of frequencies to delay though. i'm not sure why you would want to do that since it does a good job automatically from the speaker measurements.
Well, comb filtering isn't minimum phase, so I was thinking that the room corrections could leave you with group delay.
the comb filtering i'm talking about is at higher frequencies not effected by room dimensions but by early reflections. room correction are at lower frequencies. the time between + and - is greater at these frequencies. there are much bigger problems reproducing these frequencies than group delay.
That's true. At those frequencies, you're lucky if you can just get something that's reasonably flat.
decay of modal frequencies is what muddies it all up. a little bump in frequency response isn't all that bad, but slow decay of that bump sounds terrible. it covers up important harmonics and makes the bass line sound monotone. the best thing to do is a 1/12 octave 6db or better cut centered on the modal frequency. you will still hear the modal frequency but the decay will be 10db down right away and will usually drop off much quicker after that too.
Not to mention that you get pitch shift, a truly bizarre phenomenon. But I'm curious -- are you talking about adjusting the equalizer so that the time-integrated 1/12 octave frequency response is flat, or are you talking about pushing the band down even further because the time smear of the resonance is sonically worse than the resulting frequency response dip?
sometimes a 10 db cut only makes a 3db difference on stacked modes. i wouldn't call it a smear, just slow in room decay.
it's mostly just trial and error. measure, adjust, remeasure, listen to results. i just put the deqx eq window next to the etf5 low frequency measurement window and click away. once the measurement mic is set up and the software configured, it usually doesn't take more than an hour for really good results.
You're giving me a bad case of DEQX envy. Though I'd still like to do this in software if possible, since I can't afford to use DEQX for surround.
the only way to fix the comb filtering is to fix the early reflections. from my experience, anything substantial within 15ms on an impulse response measurement is going to effect music adversely.
I think that may be the most intractable problem in high-end consumer audio: how do you suppress those early reflections in moderately-sized rooms without acoustical treatments that have the wife acceptance factor of a rotten squid?
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: