Markus
I can't get near a 7.6 ms window in my room and its very large like 20 ft. x 25 ft. x 12 ft. Are you saying that there are NO reflections from any direction within 7.6 ms.? That seems highly inplausible to me in an appartment. What happened to the furniture?
Whats the speaker to mic distance? If thats less than 1 meter then thats an issue with comparing data.
I can't get near a 7.6 ms window in my room and its very large like 20 ft. x 25 ft. x 12 ft. Are you saying that there are NO reflections from any direction within 7.6 ms.? That seems highly inplausible to me in an appartment. What happened to the furniture?
Whats the speaker to mic distance? If thats less than 1 meter then thats an issue with comparing data.
He wrote 1m. Above 1k and for 50deg they follow well, so I think that the WG assembly and CD mounting is as needed.
Speaker to mic distance was 1 m. If I calculated it right then the first reflection has to travel 2.47 m (which is 7 ms) more than the direct sound.
Are you actually windowing the impulse response so that you know that there are no reflections? My numbers are based on actually seing the first reflection in the data. But I measure at 2 meters or more, 1 m is usually considered as too close.
I guess that its not worth worrying about since the discrepancy isn't that large. But you can see that there is no peak in the nearfield at 200-300 Hz. and there is nothing coming from the Comp driver at these frequencies so something is up in the setup.
TYpical acoustic measurement - doesn't add up somewhere!!
I guess that its not worth worrying about since the discrepancy isn't that large. But you can see that there is no peak in the nearfield at 200-300 Hz. and there is nothing coming from the Comp driver at these frequencies so something is up in the setup.
TYpical acoustic measurement - doesn't add up somewhere!!
That's the impulse (and the window) for 30°:
You can clearly see the reflection from the ceiling (1st peak) and from the floor and side walls.
An externally hosted image should be here but it was not working when we last tested it.
You can clearly see the reflection from the ceiling (1st peak) and from the floor and side walls.
I think the differences come from the measurement distances. How do you find the delay time? Well actually thats not that important, although you don't want to be even a few tenths of a ms late in this. Better to start it early than any chance of starting too late.
quick coupla questions on your earlier methodology for setting up the subs earle, if that's ok.
IIRC I think you have mentioned that when you do this yourself, you use bandpass subs. Let's say that in my case I use sealed, but with the dcx I can 'turn them into bandpass' by varying or setting both the HP and LP on them. As can Markus obviously.
What would be the advantages if any of having each sub reproduce a different part of the bass spectrum? It's easy enough to do with the dcx, is it worth doing?
Also you are able to select which input each sub runs off, left, right or sum (is that essentially the same thing as 'mono'??) Is it worthwhile keeping any sub 'on the left' with the left channel, ;any sub on the right' with the right channel, and I dunno, maybe in markus' case where he has a third perhaps run that sum (if it's kinda in the middle I guess). Or if the frequency is low enough (what would that frequency be) run them all 'sum'?
IIRC I think you have mentioned that when you do this yourself, you use bandpass subs. Let's say that in my case I use sealed, but with the dcx I can 'turn them into bandpass' by varying or setting both the HP and LP on them. As can Markus obviously.
What would be the advantages if any of having each sub reproduce a different part of the bass spectrum? It's easy enough to do with the dcx, is it worth doing?
Also you are able to select which input each sub runs off, left, right or sum (is that essentially the same thing as 'mono'??) Is it worthwhile keeping any sub 'on the left' with the left channel, ;any sub on the right' with the right channel, and I dunno, maybe in markus' case where he has a third perhaps run that sum (if it's kinda in the middle I guess). Or if the frequency is low enough (what would that frequency be) run them all 'sum'?
If you just have closed boxes, just use a 2nd order LP, no need to make them bandpass. There is an advantage to an acoustic LP, but not to a higher order electrical one. The idea is a lot of overlap of the sources at LF. Each one operating in a different bandwidth or off of a different channel minimizes this overlap. We did a test of CD sources some years back and found that virtually none of them had stereo LF signals. So just sum the L + R and use that signal for the subs.
Markus
I was refering to the propagation delay from the speaker to the mic. Your signal seems to be right at the edge. It is a prblem to start the "window" too late. You need to get the initial slow rise of the impulse.
I was refering to the propagation delay from the speaker to the mic. Your signal seems to be right at the edge. It is a prblem to start the "window" too late. You need to get the initial slow rise of the impulse.
Earl, the window start is set very exact. You can zoom in the impulse down to 5 digits after the dot (in ms!). Besides that FuzzMeasure allows you to generate a "Minimum Phase Copy" of a impulse response which sets the start of the window automatically at the right position.
So I have to assume that my measurements are correct. Maybe you should measure yours again with 1 m distance so we are sure that everything works as designed?
Best, Markus
P.S. This is what the output of my Receiver looks like:
So I have to assume that my measurements are correct. Maybe you should measure yours again with 1 m distance so we are sure that everything works as designed?
Best, Markus
P.S. This is what the output of my Receiver looks like:
An externally hosted image should be here but it was not working when we last tested it.
Markus
Its not possible for me to remeasure Nathans as I don't have any here at the moment. At any rate the differences are minimal and I have done my procedure for more than five years now and have it pretty well debugged so I am confident in it. Correlation of two different test locations is almost never possible without extensive work.
Measure the impedance and send that to me or post it. If there is anything wrong it will be evident in that. Also the two speakers should have identical impedance curves. I was able to find the faulty driver from the impedance curve.
Impedance is fairly detailed, and extremely accurate so comparing them is easy.
Its not possible for me to remeasure Nathans as I don't have any here at the moment. At any rate the differences are minimal and I have done my procedure for more than five years now and have it pretty well debugged so I am confident in it. Correlation of two different test locations is almost never possible without extensive work.
Measure the impedance and send that to me or post it. If there is anything wrong it will be evident in that. Also the two speakers should have identical impedance curves. I was able to find the faulty driver from the impedance curve.
Impedance is fairly detailed, and extremely accurate so comparing them is easy.
Impedance left speaker (red line):
Right speaker (yellow line):
An externally hosted image should be here but it was not working when we last tested it.
Right speaker (yellow line):
An externally hosted image should be here but it was not working when we last tested it.
Left speaker has slightly lower Q at box resonanance. Either a tiny box leak or slightly different woofer Q in production. There is chance they will even out with playing hours if the latter. Exemplary match for 99 per cent of the audio band non the less.
Markus
Not a very clean impedance measurement. The data is rough enough that you coulldn't really get valid "Q" info about the woofer. The rest looks just as it should.
I'm convince that your speakers are working fine. You might want to try those small mods that I suggested, but they are pretty small changes. They do raise the tweeter output a fraction of a dB, which might make a slightly better balance with the woofer from your data.
Whats the blue line?
Not a very clean impedance measurement. The data is rough enough that you coulldn't really get valid "Q" info about the woofer. The rest looks just as it should.
I'm convince that your speakers are working fine. You might want to try those small mods that I suggested, but they are pretty small changes. They do raise the tweeter output a fraction of a dB, which might make a slightly better balance with the woofer from your data.
Whats the blue line?
Could be a imprecision of the measurement. A speaker is a good microphone and the noise level in my apartment is pretty high (50 - 55 dbspl). Sometimes there's noise louder 60 dbspl when a helicopter is flying along the Hudson activating modes between towers in front of our apartment:
Best, Markus
An externally hosted image should be here but it was not working when we last tested it.
Best, Markus
gedlee said:
I don't think that the evident jitter can change the magnitude of the box resonance IMHO.
- Status
- Not open for further replies.