With cardiod alignment reportedly being less sensitive to placement, I would have thought the cabs would be closer to the front wall to gain more lf spl.
Regarding latency, although I haven't tried it LatencyMon (Resplendence Software - Free Downloads) seems to be able to monitor PC latency.
Would be interested to hear how the FIR filter performs if you do implement it. Particularly the amount of taps used vs the impact it has on the system (both good and bad).
Regarding latency, although I haven't tried it LatencyMon (Resplendence Software - Free Downloads) seems to be able to monitor PC latency.
Would be interested to hear how the FIR filter performs if you do implement it. Particularly the amount of taps used vs the impact it has on the system (both good and bad).
LF roll-off is an accepted trade-off with this alignment, though it's not quite as bad as a dipole I don't think. I like to use flanking woofers below 300hz to smooth the response in room. I've yet to see a narrower pattern making any difference in the modal region but I suppose that could be room dependent.
The problem with being closer to the front wall is these are still pretty wide dispersion below 1khz. With my big tv up front and heavy toe-in front wall reflections are an issue. Also, I'm not sure how being close to a boundary would affect the operation of my vents. Some say there is no effect some say it should be at least 1/4 wl away from boundaries. I'm not yet sure how to verify this with measurement. When it gets nice I'll get some outdoor groundplane measurements and use my garage as a boundary.
I've used FIR a bit in the past but the problem is to get sufficient lf resolution the number of taps goes way up and then latency is a real issue with video. For music of course it's not a problem and it's easily turned off.
The problem with being closer to the front wall is these are still pretty wide dispersion below 1khz. With my big tv up front and heavy toe-in front wall reflections are an issue. Also, I'm not sure how being close to a boundary would affect the operation of my vents. Some say there is no effect some say it should be at least 1/4 wl away from boundaries. I'm not yet sure how to verify this with measurement. When it gets nice I'll get some outdoor groundplane measurements and use my garage as a boundary.
I've used FIR a bit in the past but the problem is to get sufficient lf resolution the number of taps goes way up and then latency is a real issue with video. For music of course it's not a problem and it's easily turned off.
Okay cool. Cheers for clarifying your current set-up decisions.
I understand about the expected loss in LF effiency with cardiods. I had been wondering how illuminated your front wall was given the positioning of the cardiod nulls with cabs toed-in (and was thinking that maybe that the cardiod could be less effective at mitigating front-wall reflections in their band width due to the toe-in).
I think I saw in some of your previous experiments that you managed to produce a supercardiod pattern. Assuming you are still seeking to reduce front wall illumination, could a supercardiod pattern offer increased benefit over a cardiod pattern (I'm thinking that nulls towards the rear flanks of the cabs could reduce the intensity of first reflections off the front wall - conjecture on my part)?
Looking forward to seeing the results of your investigation into the impact of a boundary's proximity to vent performance.
I understand about the expected loss in LF effiency with cardiods. I had been wondering how illuminated your front wall was given the positioning of the cardiod nulls with cabs toed-in (and was thinking that maybe that the cardiod could be less effective at mitigating front-wall reflections in their band width due to the toe-in).
I think I saw in some of your previous experiments that you managed to produce a supercardiod pattern. Assuming you are still seeking to reduce front wall illumination, could a supercardiod pattern offer increased benefit over a cardiod pattern (I'm thinking that nulls towards the rear flanks of the cabs could reduce the intensity of first reflections off the front wall - conjecture on my part)?
Looking forward to seeing the results of your investigation into the impact of a boundary's proximity to vent performance.
I've made some changes here so I thought I'd update the thread. I'm running as a true 3-way now: the wg crossing at 1khz, the mid "kinda cardioid" TD15 running from 200-1khz, and a sealed TD15 on the bottom running up to 200hz. I have some ideas on how to improve the pattern on the acoustic resistance cab but haven't actually done anything there yet, and I have open cell foam plugs cut to fit the wg.
Today thought I'd see if I could get a linear phase response - inspired by xrk's thread: http://www.diyaudio.com/forums/multi-way/277691-s-harsch-xo.html, the phase response I got with my Synergy: http://www.diyaudio.com/forums/multi-way/283068-mini-synergy-horn-experiment.html, and Geddes' claim that his NS-15 has a linear phase response. Sorry, I forgot to save the individual driver curves, and what I show here is with the speaker setup in the middle of the room with a long gate and a freq dependent window. The cd/wg has a BW4 hp at 1khz, the mid TD is a 2nd order Bessel bandpass, and the sealed TD on the bottom has a 200hz BW4 lp.
Getting the phase response was fairly easy.........maintaining a smooth polar response was not. As it sits (sorry I didn't save the data) there's a bit of ripple around the xo region as you go off axis but it's only +/-1dB or so. Still, there's work to be done. The phase response holds up well to about 60° of axis. I'm not sure what to make of the phase wrap at 120hz. That's evident with just the TD mid cab running with no hp. I thought maybe a reflection but it's also there when the speaker is in it's usual spot as measured from the lp.
I haven't really done any listening yet.........listening to sweeps for a couple hours is pretty fatiguing and then I'm really in tune with how live my room is and pretty much anything sounds bad at that point
Today thought I'd see if I could get a linear phase response - inspired by xrk's thread: http://www.diyaudio.com/forums/multi-way/277691-s-harsch-xo.html, the phase response I got with my Synergy: http://www.diyaudio.com/forums/multi-way/283068-mini-synergy-horn-experiment.html, and Geddes' claim that his NS-15 has a linear phase response. Sorry, I forgot to save the individual driver curves, and what I show here is with the speaker setup in the middle of the room with a long gate and a freq dependent window. The cd/wg has a BW4 hp at 1khz, the mid TD is a 2nd order Bessel bandpass, and the sealed TD on the bottom has a 200hz BW4 lp.
Getting the phase response was fairly easy.........maintaining a smooth polar response was not. As it sits (sorry I didn't save the data) there's a bit of ripple around the xo region as you go off axis but it's only +/-1dB or so. Still, there's work to be done. The phase response holds up well to about 60° of axis. I'm not sure what to make of the phase wrap at 120hz. That's evident with just the TD mid cab running with no hp. I thought maybe a reflection but it's also there when the speaker is in it's usual spot as measured from the lp.
I haven't really done any listening yet.........listening to sweeps for a couple hours is pretty fatiguing and then I'm really in tune with how live my room is and pretty much anything sounds bad at that point
Attachments
Phase wrap is just where you're going below -180Hz which is the same as +180Hz so it gets plotted up there. I.E., it's just where the phase response goes through 180. You can take all the phase curve for frequencies above there and move them down 360 degrees for a still correct unwrapped phase curve without the apparent (but not actual) discontinuity..
Nate,
Nice work. When you get a chance, please show the acoustic xo plot with all 3 traces and impulse and step response. Your phase variation is smooth but a lot. Your vertical axis has pretty big steps. A Harsch XO properly done is fairly flat with a circa 55 deg excursion at the xo frequency. The tell tale that the delay is close is when the combined response is a bit lower than the BW24 response of the woofer. The step response will be triangular like but sometimes has a slight rise in front of the sharp rise. When it is dialed in the percussive jump out at you. Instruments sound real like they are in the room.
Nice work. When you get a chance, please show the acoustic xo plot with all 3 traces and impulse and step response. Your phase variation is smooth but a lot. Your vertical axis has pretty big steps. A Harsch XO properly done is fairly flat with a circa 55 deg excursion at the xo frequency. The tell tale that the delay is close is when the combined response is a bit lower than the BW24 response of the woofer. The step response will be triangular like but sometimes has a slight rise in front of the sharp rise. When it is dialed in the percussive jump out at you. Instruments sound real like they are in the room.
Ahh thanks Bill that makes sense.
X - yeah it still needs some work. The tweeter is actually out of phase to get there so something isn't quite right yet. I need to go back thru your thread and see if there's anything I missed.
Edit: the lumpy fr from 400-1khz is the room coming in the measurement.
X - yeah it still needs some work. The tweeter is actually out of phase to get there so something isn't quite right yet. I need to go back thru your thread and see if there's anything I missed.
Edit: the lumpy fr from 400-1khz is the room coming in the measurement.
that's a fairly large fdw, you could reduce that a fair bit and still get usable data (though the response is fairly smooth anyway so it might not change much).
No, it doesn't change enough to make an appreciable difference.
Hopefully I'll have time with the long weekend coming up to fine tune this.
Sorry I wasn't clear - I only tried down to a 10 cycle window. A 4ms gate showed a little more ripple around the xo than the usual LR xo, and of course that kind of resolution won't show much at 400hz regarding the peak there. I have outdoor data on this woofer cab and it looks nothing like that.
)- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.