audio-talk

And this is the combined plot of all four drivers in the left speaker at the listening position using the x9 self averaged measurements:
For ease of reference this is the previous plot:
The 208 and T90 plots are reasonably different on the face of it.

I thought I'd investigate the variability of the measurements within the averaged reading, here's the SB sub:
This is the 285GMF:
The 208EZ:

And finally the T90A:
I'm not sure there are too many conclusions, other than small distances can count, and my room's probably a bit nasty for sound.

I'm not sure what to try next, but I think I might gaffer tape some cardboard to extend the baffle wings and remeasure to see if it has any affect, and if so what. I'm particularly interested to see if it changes the 285GMF response at all as they're a bit peaky.

I'm open to suggestions. Room treatment will be a consideration at some point I guess.

Ive found since i did the last set of measurements on the fanes that shut curtains made quite abit of difference. the left hand side of the room appears more lively than the right, and that moving the footstool made a big difference, especially to the channel balance

I'm sure you're right Ant - mine's an odd room: concrete floor over the garage, brick walls on 3 sides, stud on the right, crammed full of stuff as it's now my office/play room/storage facility. When I played around with a room mode calculator IIRC there were a few around 110Hz, and multiples of, that appear to be affecting the 285GMF in particular.

It would be interesting to drag it all downstairs to the living room so there was space behind the speakers as well as to the sides and front but it's a hell of a job, I'd need Pickfords.

Funnily enough I did some measurements of the right speaker also - all the previous were with the left speaker only.
This needs a bit of description, and it wasn't easy to show both left and right speakers on the same plot with enough detail - the reddy colour is the right speaker, the green the left.

The left speaker was measured first, then I measured the right starting with the T90A and worked down to the sub. But after the 208 when I started to measure the 285 I got feedback through the mic and had to reduce the measurement signal level a touch to stop it. So it's not a completely like for like comparison of the L and R plots of the SB sub and 285. Whilst the SB sub looks a little better perhaps the 285GMF has a rather horrid peak around 280Hz and another smaller around 130Hz.

Just a quick note that you really shouldn't be getting feedback, and that might be throwing all of your measurements out.

I can't remember which USB audio interface you're using, but anything like "monitor" or "direct monitor" should be off.

That switch feeds the inputs straight to the outputs so that musicians have a low-latency (the signal stays in the analogue domain) monitor feed so they can hear their mic/instrument (through headphones) when they perform.

Chris

Nuts. I wonder when that button was pressed. Only 100 odd measurements to repeat.

It's a Behringer thingy BTW.

It's taken me a little while to summon up the enthusiasm to redo the measurements but here they are. For simplicity I've only used REW's inbuilt 8 sweep averaging rather than do my own and average. And I only measured the Fostex 208Esigma and T90A as a combined "unit".

Here's the left channel: And here's the right channel: The plots suggest a falling response, and perhaps the SB29 subs could be turned down a little, and the 208&T90 up a touch. I don't feel this is required by listening to music though. The 208&T90 response in particular is a consequence of their directionality I suspect, more of which later.

Here's the average of the previous plots for the left and right channels

For a little fun here are plots of the left and right channels, a single measurement taken in front of the middle of my head compared to an 8 average of a centre reading and a further 7 taken on a circle c200mm radius.
This appears to support the theory of the directionality issue of the 208 & T90.

And a final plot of both channels, a single reading compared to a REW 8x average
Thoughts gratefully received. The response is a bit lumpy in parts, I'd guess a reflection of the room and the drivers themselves.

Simon,

When you say "8x average", do you mean you told REW to do multiple sweeps in the "Measurement" pop-up, or did you run the same sweep eight times (ie, clicked "measure" eight times, did 8x single sweeps) and average those?


Assuming the same mic position, I find the difference from 1x sweep to 8x sweeps to be very odd, and would be worthy of further investigation.

Chris

REW did 8 sweeps then gave me the average, for each I positioned the mic centrally then 7 on a circle approx 200mm radius to give a less sensitised reading. I thought this was the thing to do?