REW and room acoustics - what to the graphs tell you?

The alternative cause is a fault with the drive unit. I discovered Much the same in my system last year,  a distinct increase in distortion at a specific frequency (mine around 600Hz IIRC) - indeed with the frequency sweep I could hear a ‘tizziness’ as it swept past the errant frequency, though I had not noticed with music.  Testing one at a time revealed it was only from one channel. Swapping the speaker cables over on the power amp confirmed it was the speaker not amp or earlier. Being in the midrange units’ range, I swapped the drivers between the speakers, which confirmed it was the mid unit not anything else, and that was easily replaced.

My system had been unintentionally subjected to a very loud switch-on thump some weeks previously, so it is possible that had caused the damage.

Yes, exactly why I asked for the separate traces for each individual channel.
As I said previously "As drive units tend to be assembled by hand, faults are very often associated with individual units rather than entire batches.".

(Although I also want the single traces to eliminate the possibility speaker / speaker interaction causing some of the peaks at LF and to positively identify the speaker phasing).

I shall also try different speakers, out of interest - the only other ones I have that could be interesting are IBL and ACT SCM 10 and SCM 7

OK, here are some more scans, showing both, left and right.  Assuming I can get the insert picture thing working...

OK, that failed.  I'm not sure what I'm doing wrong.  I'm putting this on IMGUR and clicking on the mountain and pasting in the BBCODE from IMGUR.

[url=https://imgur.com/U092M8c...2M8c.jpg[/img][/url]

Oh well, that was using the link.  Still no good.  

[url=https://imgur.com/U092M8c][img]http://i.imgur.com/U092M8c.jpg[/img][/url]

You need the direct link, not bbcode or anything like that.

This is by taking the bit between img and /img tags in your last attempt - that's the bit you need to paste into the 'mountain' box here on the forum, nothing else...

Figured it out!  Sorry, that's a lot of graphs...

Rather concerned at the drop off after 10k.  I listened to a sine wave going up at that sort of frequency, and found that I couldn't here much above 12 or 13k, and thought "well, at my age what do you expect?" but I'm wondering whether that drop is partly responsible...

Hooray, we have pics in our time! 

I feel so proud!

I just tried with the IBLs.  Interesting.  I don't get that 1.7k to 2k rise in harmonic distortion.  I get a very similar drop in volume from 10K.  I get more of a drop at 70Hz with the IBLs in the SPL.  Rather lower volume generally from about 180Hz and down - not surprisingly - and a generally somewhat lower volume from about 2k and up.

I also tried turning off the amplifier (NAP500) to see what happens with the 50Hz signature.  It's there, quite prominently.  So it isn't from the power amp.  I'm not aware of a 50Hz sound particularly.  I might try to trace that and see what is causing it when I get a chance.

My guess is that the mains hum is from the PC that's running REW...

Very " busy" looking graphs

But what are they supposed to look like and why?

The only thing that stands out to my untrained eye is the distortion around the 3k mark...but wouldn't know what to make of that anyway

Yes - that and the uptick at about 10k.  No idea what causes that.

First, did you check the phasing of the speakers as IB suggested?

In respect of the distortion traces...

Although the distortion peak is of rather larger amplitude on the left than the on right it is in a very similar frequency range.  However the peak in the right channel is more spread out; in fact they actually look to have a similar amount of total energy in the peaks.  Interestingly the frequency of 1.7kHz-1.8kHz is about equal to the longitudinal resonant frequency of the cylindrical sub-enclosure used for the BMR, so the rather 'odd' response at that frequency may well be caused by the 'tuning' of this rather unusual speaker design.

You can ignore the fall off of response above 10k, it's the same on both channels, so it's probably a design feature (listening to speakers that are flat here is often quite tiring in the long term).  You can also probably ignore the rise in distortion at 9kHz on the right, that's too high to be important and could be something in the environment getting excited (even something as simple as a loose screw!).

I'll get back to the response below 400Hz later when I have more time.

Tried some sound treatments in the room - foam pieces in various places (pro-coustix Ultraflex Studio pack).  Some differences to the plots, but nothing spectacularly different.  I will need to experiment with placement - but there are so many corners in the room that it is hard to guess where would be most effective.  There are glass-fronted bookcases on the wall that's behind the speakers (one each side, to the outside of the speakers) and windows.  I put foam over as much of the glass as possible.  Not sure how much of a difference that made to the sound I experienced.  The plots improved a little, but no difference to the 1.7KHz part.

Beachcomber posted:

Tried some sound treatments in the room - foam pieces in various places (pro-coustix Ultraflex Studio pack).  Some differences to the plots, but nothing spectacularly different.  I will need to experiment with placement - but there are so many corners in the room that it is hard to guess where would be most effective.  There are glass-fronted bookcases on the wall that's behind the speakers (one each side, to the outside of the speakers) and windows.  I put foam over as much of the glass as possible.  Not sure how much of a difference that made to the sound I experienced.  The plots improved a little, but no difference to the 1.7KHz part.

Unfortunatly foam traps indeed would be unlikely to have any observable effect on the troublesome bottom end, because their absorbance doesn’t go low enough (in fact the manufacturer only says upper bass and mid).The same people do a fibreglass one which is more likely to be effective at low frequencies, though unfortunately they don’t publish any performance figures so it is impossible to compare with other manufacturers’ products other than by the dimensions and declared construction. The other foam traps will probably reduce reflections at the higher end of the spectrum, beneficial if the room is quite reflective, which may be what you are seeing.

The distortion around 1.7kHz probably being from the speakers themselves wouldn’t be reduced by any room treatment.

For lower frequencies neither foam nor any fibre material are going to make a significant difference when placed at a room boundary.  These type of open materials are 'flow mode absorbers', absorbing energy by controlling airflow.  At the room boundaries (and particularly the corners) the airflow is at a minimum so flow mode absorbers have little or no effect (on the other hand pressure is at a maximum, so 'pressure mode absorbers' work best in the corners!).

To reduce bass resonance (anything less than 100Hz) you need to use 'pressure mode absorbers' in the corners.
The three main forms of there are:
Mass Loaded Membrane
Helmholtz Resonators
Damped Membrane

To get a flow mode absorber placed at a room boundary to work even reasonably well at 100Hz (i.e. for 100Hz to be the lower -6dB point of its absorption curve), it would have to be 0.85m thick - i.e. λ/4.

Ah. Well, that's £100.00 I'll not see again.  Never mind.  I'll experiment a bit more, but given this I doubt much will come of it.

I'm worried about the 1.7kHz bit - should I talk to Naim about that?  or is that likely simply to be a characteristic of the S600?

I feel slightly peeved because they describe some of the pieces as bass traps.  Shows how much you can trust marketing names.

I know, when looking for bass traps I initially looked for "less expensive" commercially made units.  I noticed that very few published any information about the actual absorption characteristics of their devices and the few who do showed little or no absorption below about 70Hz and typically maximum absorption in the 120-150Hz range.  120Hz corresponds to a room with a maximum dimension of 1.42m (about the size of a lavatory (and not even a whole bathroom!).  The rest are using similar designs and (given that they don't publish figures) are probably even less effective!  I think these companies are knowingly misleading potential customers.

On the other had there are other companies like GTK that do produce effective products (down to about 50Hz at least), but they are very expensive.

When looking and researching I had the advantage of a decent background in physics and signal handling, so with some research I could identify the systems which were most unlikely to meet the performance expectations suggested by the manufacturers claims - most people don't have this advantage and some of the manufacturers exploit this without limit.

Currently I'm working on some CAD designs for pressure mode DIY units that should be reasonably effective down to about 30Hz with a maximum absorption from about 50Hz to about 250Hz.  The designs aren't yet to the level where I can publish them, but the current system I have is quite effective (12dB reduction of the resonance and 3x reduction of the reverberation time) at 42Hz, and I hope to exceed that performance with the new design.  The trouble is that they're not going to be small - about the size of a two or three large bookcases.

Yes, size can be a problem. When I was looking at GIK their best bass traps weren’t enough, so I was drawn to something by RPG - but they were very expensive and the manufacturer wax  not good at answering searching questions about actual performance (as opposed to marketing hype), and then when they did it seemed unlkely to do the job I wanted. I ended up thinking I would go down the DIY path with mass loaded membrane absorbers, but it would have to wait both for spare time to build and as I have plans to move home in maybe 3 or 4 years I may end up not bothering until After moving (however, if the spare time comes along, I may well play...)

Beachcomber posted:

Ah. Well, that's £100.00 I'll not see again.  Never mind.  I'll experiment a bit more, but given this I doubt much will come of it.

I'm worried about the 1.7kHz bit - should I talk to Naim about that?  or is that likely simply to be a characteristic of the S600?

I suspect it's a characteristic of the S600's BMR and its mounting, but no harm in sending a e-mail to Naim.  If it is a fault they'll probably want to fix it!

The only things outstanding are

The reverberation time of the room is rather too long across the board:
      Above 200Hz it RT60 should be about 0.33s, but the spectrogram and waterfall plots both suggest times about double that.
      Below 200Hz it RT60 can rise, but it should still stay be below 0.6s, again the spectrogram and waterfall plots both suggest times about double that.

At 28Hz, 48Hz and 61Hz, reverberation time is very much too long and there are corresponding peaks in the SPL traces at the same frequencies: This indicates uncontrolled room resonance modes.  As you rightly surmised, the only real cure for this when using full range speakers is to install bass traps.  Moving the location of the speakers can help to some degree by reducing the coupling of the speaker to the acoustic resonance of the room, but the resonance (and reverberation time problem) will still be there.

Lastly there's some acoustic cancellation going on at LF - the nulls in the 'Both Speakers' traces are deeper than either of the 'single speaker' traces.  This could be an out of phase problem, or could be a speaker placement issue.

One final word...

Overall, those results really are pretty good.
Don't worry too much about any of the specific details too much: All taken together, the graphs indicate that your system is generally working well, with the only particular exceptions being the reverberation time, and to a lesser extent the LF speaker to speaker cancellation.