advertisement


Measurements - Tannoy Stirling HE

H

Hoopsontoast

pfm Member
Finally got round to setting up a measurement system, I had planned to do the Rega R9s but did not get round to it before they went out on loan.

Method:
Mic- MiniDSP UMIK-1 via USB (With Calibration File)
Software - REW on Dell E6220 Laptop
DAC - USB to Phono via Behringer UCA-202
Amplifier - Bryston B60
Speakers - Tannoy Stirling HE

The room:

System by RSdesignUK, on Flickr

First measurements taken at the Listening Position on a sofa just under 3m from the speakers.

Right and Left Plots for the Tannoys at the Listening position. The bass certainly does not sound as pronounced as you would imagine looking at the plots.
Tannoy @ Listening (R + L) by RSdesignUK, on Flickr

Distortion Plot

Tannoy @ Listening (R) Distortion by RSdesignUK, on Flickr

And interestingly, plot showing before and after the foam absorbers, subjectively sounds tighter with them in. (Green -With / Red - Without)
The Group Delay plot is smoother by around 20ms average though.

Tannoy @ Listening (R) With and Without Foam by RSdesignUK, on Flickr

And another peice of interest, with the -3dB/0/+3dB control on the Tannoys, seems to be -2dB/0/+2dB in my room

Tannoy @ Listening (R) with +3, 0 and -3 Treble by RSdesignUK, on Flickr


Any thoughts/advice?
 
The plots are very typical of in-room, and they're not doing anything drastically unexpected. The bass reinforcement is very typical, especially of placement close to a rear wall . You have a trough close to 200Hz which could be bass slap (suckout caused by indirect reflected bass meeting out of phase with direct radiated)...again, typical.

The response above 10KHz seems quite rolled off though. Did you have the mic pointing in between the speakers? This will place the mic off axis so some drop off is expected. It could also be partly due to the room's soft furnishings which typically absorb a fair amount of upper HF compared with say anechoic response, so I wouldn't ever expect in room to be flat...that could sound too bright for most tastes. Saying that, I'd think that an in room measurement might sound a little more balanced between 5 to 10dB down (on axis) at most at the upper end. Yours appears to be closer to 15-20 but your mic isn't measuring on-axis.

There appear to be two peaks one of which might possibly coincide with crossover region and the other with horn resonance (1KHz and 3KHz). The 3KHz one is very likely horn resonance for which there should be a notch filter fitted within the crossover. If so, it might benefit from a slight alteration of values if it sounds as obvious as the plot suggests. The 1KHz peak though is quite significant in the very sensitive mid range area. Not too much can be forensically determined (nor should be) from in-room responses though as there's all sorts of reasons for peaks and troughs, with reflections and cancellations galore.

Might be worth re-doing a few measurements (one speaker at a time) from the listening position with the mic pointed at the speaker. This may give more balanced results; that and try toeing the speakers in so they're on axis at the listening position. Other than that, they are what they are, and as long as you think the sound is good for your tastes, then the measurements really aren't adding anything to that, so I wouldn't stress too much over them. Distortion seems a little high, but again may not be out of the way for the speakers in-room response.

I don't know if you have the ability to gate the responses so you can take the room effects out? This would could be helpful in more accurately determine the 1 and 3KHz peaks. If you don't have that ability, downloading something like Holmimpulse could be helpful, as that can generate step response and gated FR plots.
 
Hoops

Interesting post. Definitely gets one thinking. I look forward to reading more as you learn.

Thanks
 
Not bad results, you'll always get some assymetric reinforcement with that dogleg. If it were me I'd pull speakers forward to flatten the response.
 
Pac1 said:
The plots are very typical of in-room, and they're not doing anything drastically unexpected. The bass reinforcement is very typical, especially of placement close to a rear wall . You have a trough close to 200Hz which could be bass slap (suckout caused by indirect reflected bass meeting out of phase with direct radiated)...again, typical.

The response above 10KHz seems quite rolled off though. Did you have the mic pointing in between the speakers? This will place the mic off axis so some drop off is expected. It could also be partly due to the room's soft furnishings which typically absorb a fair amount of upper HF compared with say anechoic response, so I wouldn't ever expect in room to be flat...that could sound too bright for most tastes. Saying that, I'd think that an in room measurement might sound a little more balanced between 5 to 10dB down (on axis) at most at the upper end. Yours appears to be closer to 15-20 but your mic isn't measuring on-axis.

There appear to be two peaks one of which might possibly coincide with crossover region and the other with horn resonance (1KHz and 3KHz). The 3KHz one is very likely horn resonance for which there should be a notch filter fitted within the crossover. If so, it might benefit from a slight alteration of values if it sounds as obvious as the plot suggests. The 1KHz peak though is quite significant in the very sensitive mid range area. Not too much can be forensically determined (nor should be) from in-room responses though as there's all sorts of reasons for peaks and troughs, with reflections and cancellations galore.

Might be worth re-doing a few measurements (one speaker at a time) from the listening position with the mic pointed at the speaker. This may give more balanced results; that and try toeing the speakers in so they're on axis at the listening position. Other than that, they are what they are, and as long as you think the sound is good for your tastes, then the measurements really aren't adding anything to that, so I wouldn't stress too much over them. Distortion seems a little high, but again may not be out of the way for the speakers in-room response.

I don't know if you have the ability to gate the responses so you can take the room effects out? This would could be helpful in more accurately determine the 1 and 3KHz peaks. If you don't have that ability, downloading something like Holmimpulse could be helpful, as that can generate step response and gated FR plots.
Click to expand...

Thanks for the in depth reply.

I had the mic in the main listening position in between the speakers on the sofa.

They are toed in as a result of my subjective playing around and seeing whats best.

I am thinking of ordering some GIK 244 panels to replace the foam ones, and thought the measurements might help in my choice of which ones to go for or what areas to address.

The Foam panels already help in room with echos.
 
Hoops

What would probably be a more interesting graph would be the waterfall plot to show the room low frequency resonances and how they decay. REW has this functionality. The smoothing you have applied to the supplied graph will somewhat hide the peaks in the bass.

The bass does look a little high compared to the rest of the response, however this will be lifted by the room resonances. Foam absorbers will do very little to tighten bass, but will reduce echoes and make the room sound less live. Excessive treatment may absorb too much high frequency. You really need a good compromise between absorption and difusion.
 
BE718 said:
Hoops

What would probably be a more interesting graph would be the waterfall plot to show the room low frequency resonances and how they decay. REW has this functionality. The smoothing you have applied to the supplied graph will somewhat hide the peaks in the bass.

The bass does look a little high compared to the rest of the response, however this will be lifted by the room resonances. Foam absorbers will do very little to tighten bass, but will reduce echoes and make the room sound less live. Excessive treatment may absorb too much high frequency. You really need a good compromise between absorption and difusion.
Click to expand...

I will post those later, as mentioned, I was hoping that the measurements would be able to influence the selection of GIK products to look at.

For instance the GIK 244 (Range Limiter version) looks like it will address the problem best with the space available, by being most effective in the 80Hz-160Hz range which is roughly my main problem area.

http://gikacoustics.co.uk/product/gik-acoustics-244-bass-trap-flexrange-technology/
RAL-Chart-for-244-Bass-Trap-FlexRange-Technology.jpg


I don't really have the space for the soffit bass absorbers.
 
Don't expect miracles. A few panels cover only a tiny fraction of a room's surface area. Your measured differences with/without foam are probably even within the variability between measurements of the same setup. Foam doesn't do much at all down a few 100Hz.
 
GIK are very helpful and will give you their recommendations if you supply them with full details.

I've just been through this and thought I knew what I needed but according to GIK I was wrong. Finished up with some Freestanding Bass Traps , and they cured the problem better than I had thought possible.
Even on their recommendations I found the best results by fine tuning the position , one advantage of freestanding units.
 
Made quite a usefull indent in that peak by raising the speaker up by 100mm with some food cans. I tried 300mm with some other speakers underneath and not only did it look odd, sounded and measured less well.

With the 100mm 'Stands' the treble from around 500Hz is a lot more linear (apart from the peak around 1kHz).

With 100mm stands (Red) and without (Blue) and both with treble @ -3dB.

Tannoy @ Listening (R) +10cmL & 0cmL NT -3dBT FR by RSdesignUK, on Flickr
 
You must log in or register to reply here.


advertisement


Back
Top