My philosophy is to never adapt a speaker to a specific room. As I mostly start building new ones as soon as I finished the previous, they never stay in my main listening room very long.
I didn't ask about adaptation to the room, but I was simply curious about how the speakers behaved in the given place. In my opinion, living room bass is the result of approximately 60% speaker, 40% room response, especially if the boxes are as close to the walls as we can see in your picture. Ultimately, if you want good/balanced bass sound, you must adapt the speakers to the room. Or maybe you use a room correction DSP for such a purpose? Even the best anechoic speaker bass response get trashed if there is big null(s) or peak(s) in the room at the listening position.
Last edited:
You’re right, to get the best sound a speaker should be adepted to the room. Also, the room should be treated and the placement of speakers and listening position should be carefully considered. This is all very true.
However, unfortunately my wife won’t allow me to put larger(ish) speakers in the living room. I have to make do with a small room on the 2nd floor. So not many options regarding speaker (or seat) placement. :-/
However, unfortunately my wife won’t allow me to put larger(ish) speakers in the living room. I have to make do with a small room on the 2nd floor. So not many options regarding speaker (or seat) placement. :-/
If you tune higher, like say 40 Hz or more, you might get a nice bass peak. Could you try making a port half as long as the current one? Could be interesting... A flat response on paper does not guarantee a proper sounding room response, especially in the bass region.
Last edited:
Why the difference between REW and VituixCad in summed port+woofer near-field response? I don't have REW, so I can't see the individual port and woofer near-field measurements, but summed near-field woofer+port measurements in VituixCad looks much more logical. Anyway, I suspect Thiele -Small parameters in Dayton data sheet are wrong - you should measure them.
The braces are fighting the quarter-wave a bit (BTW they would be much more effective if the braces were vertical), and the off-set vent will further reduce it, but you have an ML-TL so the BR tunign eill likely be too high, but that should give a bump in the bottom before roll-off but without modelinghard to know.
Are you sure your box is not properly sealed?
dave
I could try to shorten the port even further, but going from 18.2 to 13.5cm the dip only deepened (and shifted to a slightly higher frequency). According to WinISD, this shifts the BR tuning frequency from 31 to 37Hz. I doubt going higher will solve the issue. Blue: 31Hz tuning (18.2cm port), red: 37Hz (13.5cm port).
Given an ML-TL the vent should likely be longer not shorter.
But you should model it to know.
dave
I'd still try an outdoor 1m groundplane measurement to see if the dip is caused by boundaries / room effect.
The lowpass section together with the impedance adds about 2dB at about 100Hz. I’d try to equalize the second impedance peak with a parallel notch. Did you do a SPL measurement of the port?
I think I solved my conundrum. The dip at 45Hz looked to me like the port and driver are out of phase, which got me thinking. For farfield measurements I use a loopback for timing reference, but for nearfield I set REW to 'no timing reference'. So this time I went to the 'Impulse' window in REW and clicked 'Set t=0 at IR start' button for both the driver and port measurements, and then summed both again. This time the plot looks a lot more like what I would expect. The blue line is with timing/phase correction, the red line without:
I'd either lengthen and shorten the port and see what it does (lay some pencils in it to lower the frequency), or better still, take an impedance sweep.
I already have an impedance sweep at a port length of 18.2cm (as calculated by WinISD). The tuning frequency is nicely at the calculated 31Hz. As the port is slightly tapered, it's hard to predict the actual tuning frequency if I shorten it significantly. But I can ofcourse always check it by doing an impedance measurement.
I'll play around with different port lengths to see if better results can be achieved.
I'll play around with different port lengths to see if better results can be achieved.