I ended last weekend’s experiments intrigued by the post by @alandbush regarding my 4.75kHz peak, and was most interested in the fact that reading the description of what the ‘presence’ frequency range can do to your perception of sound was pretty much a perfect description of why the SOtM kit seamed to irritate me in some areas. I also had some issues using REW last week via the SOtM or mR, whist I got perfectly consistent results, REW would only work if I generated the WAV file in 16bit (it would produce a graph using 24bit, but it was clearly erroneous), another REW issue I had was that although results were consistent through AIR, the mR and SOtM streaming options, the bass end of the curve was massively dipped versus using REW to play the sweep in real time. So although I made measurements that proved the SOtM and mR otions were tonally very similar, I did not have much confidence in the overall results versus the conventional way of using REW, which produces a nice curve and shows the in room bass reinforcement you might expect. I experimented with this again today, with the same problematic results, so I remain a little baffled. Anyway, I gave up on this, because I do not want to spend hours messing about with REW issues.
Reverting to using REW using the conventional way and running the measurement sweeps through AIR indicated a similar peak at about 4.6kHz, this looked similar to last week’s measurements, albeit the peak is at about 4.75kHz.
I tried left / right checks, these show a less pronounced peak at about 4.6kHz, with the next dip at about 5.7kHz. Interestingly, the curves for left and right are very similar. I am not an expert in room acoustics, but it does look like this might be a case of left and right peaks reinforcing each other.
With other experimenting I found moving the measurement microphone back 1m or so gives me a dip at 4.6kHz, but invokes a new peak at 5.1kHz. This peak is perhaps less pronounced, but not by much. After a few runs it was clear that the peak is dependent of microphone location, much as you would expect. It was also clear that there is perhaps anther peak at about 3kHz. Sometimes I will listen to music when sat at a computer desk at the back of the room, this is 2 or 3 metres back from the listening position and offset left to right, this position shows a 3kHz as the most pronounced peak, with 4.75kHz peak almost eliminated.
OK, so I move the measurement microphone and the REW curves change a bit, this is no surprise, in fact, it is exactly what you would expect. This did leave the question as to why I now have a 4.75kHz peak in the listening position and earlier REW reading do not show this. What has changed in the room? Nothing of significance. One thing I could think that I had done is move the right hand Blade to get better access to the rack, this was when I was installing the new kit and making a couple of other changes in the rack. The speaker was then moved back to the same location as indicated by dents in the carpet caused by the speaker’s feet. Did the right hand speaker go back exactly where it was? Maybe not, they are on IsoAccousics feet, not spikes, so the location could be fractionally different. As an experiment, I tried reducing the toe-in of the right hand speaker by an absolute fraction, within the tolerance of the carpet indentation, if you like. I could not see this would make much difference, but it did, the 4.6kHz peak was eliminated. If I am honest, I am slightly staggered by this result. I had changed the toe-in of the right hand Blade by an amount that I would consider to be utterly insignificant, certainly nothing to worry about, but this had taken 4dB off the 4.6kHz peak.
With my new but not really changed that much at all speaker location, I tried the sweep at the desk top location, and the 3kHz peak I mentioned before was no longer prominent either.
These results were almost beyond belief, at this point I was genuinely gobsmacked. EDIT - That's because these results were wrong.
To double check this, I reran the REW checks using the slightly trickier method via the SOtM kit and compared to last week’s curve produced in the same way. This still shows a peak at about 4.75kHz, but it is reduced by a little under 3dB, and it indicates that I have gained a peak at maybe 8.4kHz. So this does show that how you use REW can have a big influence on results. Either way, the 4.75kHz peak is indicated as reduced, that speaker nudge has done something. Intriguingly, that 2 or 3dB number keeps reoccurring, either in graphs or in terms of how much I like to turn the treble setting down.
At this point, all I had done was produce some pretty graphs, and although slightly gobsmacked by the graphs, I had little confidence that the system would sound that much different.
OK, REW was saying that my system should sound different, but I was having difficulty believing that rotating one speaker by something like 1 degree would do much.
Next to some listening, I lined up the same playlist of tracks that had alerted me to the treble issue in the first place and sat down for a listen. In summary, a vast improvement. There were a couple of tracks that I think first alerted me to the ‘harsh sounding SOtM’ issue, these sounded fine now, in fact, they sounded excellent. I then tried a couple of tracks that have never sounded that good, they were still a little problematic, but as an experiment I tried using Roon’s parametric EQ to dial out my ‘new’ 8kHz peak, one of those not very sounding tracks suddenly sounded as good as I have ever heard it. This was a pleasing result. With three treble sensitive tracks I re-tried my ‘blind’ treble adjustment test, the results were 0.0, 0.0 & +0.5. Curiously, in subjective terms, I would still say that the mR sounds darker than the SOtM kit, and yet I can pretty much nail the blind treble test on 0.0dB.
Genuinely fascinating. I would now say that the fundamental issue is solved, and I can now see scope for further investigation and improvement. I have been hunting blind with this one, but now I can clearly see my foe and so can go on the attack. (and how cool is Roon? You can adjust the parametric curve on an iPad, sat in the listening position, in real time. Very, very, cool)
In general terms, the rig is now sounding as good as it ever has, plus I can now see easy scope for further optimisation. As one example, I think my room is generally a good one for audio, it is rectangular / symmetrical, a good size (14’ by 20’), carpeted, rear wall broken up, furnished and full of junk. One key negative issue could be ‘first reflection points’, on the left side I have nice floor to ceiling curtains to soften things up a bit, on the right side a solid wall (breeze block and dry wall). Definitely not ideal, definitely not symmetrical, from an audio perspective at least. Plus I have learned that my adjusting the speaker toe-in by basically nothing much can make the difference between a sometimes annoying and fatiguing listen, to audio joy. Definite scope for optimisation.
This does leave one mystery though, why did the 4.75kHz ‘presence’ issue bother me with the SOtM kit, and not with AIR or the mR. Indeed, I watched a couple of movies over the Easter break, with the Blades on front speaker duty fed (analogue) from a Yamaha AV surround decoder. It sounded spectacular. So why did this issue only hit me with the SOtM kit in place?
Considering AIR, the mR and SOtM kit show almost identical frequency curves, a logical conclusion would be that the SOtM kit does actually sound harsher in the presence region. Everyone says the SOtM kit is more transparent, so this is counter-intuitive, but it is logical. Time for another experiment. I set up a very strange Roon parametric EQ, to eliminate bass, mids and emphasise the presence region. Not the most pleasant thing to listen too, but it does allow you to hear the offending region with great clarity. I plan to try this with the mR vs SOtM kit, but frankly I could not be bothered to swap the two today. So as a simple experiment I tried this SOtM versus Roon AIR, which does not require any kit and cable swapping. It is a difficult test but I could hear some differences, very subtle though. One thing is clear, even with this horrific presence only set-up, I can hear nothing fundamentally wrong with the SOtM kit, but things like the snare springs of a snare drum do sound a little different. One very curious result, I tried listening to a snare drum section with this bonkers EQ with the Mutec engaged then disengaged with the SOtM kit, it’s subtle, and I am convinced it was far better with the REF10 in play. Might be expectation bias I suppose, but I am sure it sounded clearer. So clearer presence resolution with enhanced presence frequencies might be enough to trigger some kind of psychoacoustics effect? This might be worth revisiting when I have a little more time. (and can be bothered.) Also, I think the bonkers EQ method might be informative when trying different power options with the SOtM items, in terms of clearly indicating any effect in frequency region that my brain is clearly very sensitive too. The same applies to future SMPS elimination tests and similar.
I have to conclude that sometimes the tiniest things can have a positive or negative effect, and sometimes the tiniest negatives can grab your attention to the elimination of most everything else. This might explain why some people are made very happy with things like USB cable swaps and why seemingly insignificant stuff like changing the power supply to a component can yield perceptible results.
So many things for continued investigation. At a basic level the issue is solved, but the puzzle as to why the mR sounds (subjectively) darker does remain. I am listening to the SOtM kit now, it does sound brighter, I am telling you.
Time for a break!
Reverting to using REW using the conventional way and running the measurement sweeps through AIR indicated a similar peak at about 4.6kHz, this looked similar to last week’s measurements, albeit the peak is at about 4.75kHz.
I tried left / right checks, these show a less pronounced peak at about 4.6kHz, with the next dip at about 5.7kHz. Interestingly, the curves for left and right are very similar. I am not an expert in room acoustics, but it does look like this might be a case of left and right peaks reinforcing each other.
With other experimenting I found moving the measurement microphone back 1m or so gives me a dip at 4.6kHz, but invokes a new peak at 5.1kHz. This peak is perhaps less pronounced, but not by much. After a few runs it was clear that the peak is dependent of microphone location, much as you would expect. It was also clear that there is perhaps anther peak at about 3kHz. Sometimes I will listen to music when sat at a computer desk at the back of the room, this is 2 or 3 metres back from the listening position and offset left to right, this position shows a 3kHz as the most pronounced peak, with 4.75kHz peak almost eliminated.
OK, so I move the measurement microphone and the REW curves change a bit, this is no surprise, in fact, it is exactly what you would expect. This did leave the question as to why I now have a 4.75kHz peak in the listening position and earlier REW reading do not show this. What has changed in the room? Nothing of significance. One thing I could think that I had done is move the right hand Blade to get better access to the rack, this was when I was installing the new kit and making a couple of other changes in the rack. The speaker was then moved back to the same location as indicated by dents in the carpet caused by the speaker’s feet. Did the right hand speaker go back exactly where it was? Maybe not, they are on IsoAccousics feet, not spikes, so the location could be fractionally different. As an experiment, I tried reducing the toe-in of the right hand speaker by an absolute fraction, within the tolerance of the carpet indentation, if you like. I could not see this would make much difference, but it did, the 4.6kHz peak was eliminated. If I am honest, I am slightly staggered by this result. I had changed the toe-in of the right hand Blade by an amount that I would consider to be utterly insignificant, certainly nothing to worry about, but this had taken 4dB off the 4.6kHz peak.
With my new but not really changed that much at all speaker location, I tried the sweep at the desk top location, and the 3kHz peak I mentioned before was no longer prominent either.
These results were almost beyond belief, at this point I was genuinely gobsmacked. EDIT - That's because these results were wrong.
To double check this, I reran the REW checks using the slightly trickier method via the SOtM kit and compared to last week’s curve produced in the same way. This still shows a peak at about 4.75kHz, but it is reduced by a little under 3dB, and it indicates that I have gained a peak at maybe 8.4kHz. So this does show that how you use REW can have a big influence on results. Either way, the 4.75kHz peak is indicated as reduced, that speaker nudge has done something. Intriguingly, that 2 or 3dB number keeps reoccurring, either in graphs or in terms of how much I like to turn the treble setting down.
At this point, all I had done was produce some pretty graphs, and although slightly gobsmacked by the graphs, I had little confidence that the system would sound that much different.
OK, REW was saying that my system should sound different, but I was having difficulty believing that rotating one speaker by something like 1 degree would do much.
Next to some listening, I lined up the same playlist of tracks that had alerted me to the treble issue in the first place and sat down for a listen. In summary, a vast improvement. There were a couple of tracks that I think first alerted me to the ‘harsh sounding SOtM’ issue, these sounded fine now, in fact, they sounded excellent. I then tried a couple of tracks that have never sounded that good, they were still a little problematic, but as an experiment I tried using Roon’s parametric EQ to dial out my ‘new’ 8kHz peak, one of those not very sounding tracks suddenly sounded as good as I have ever heard it. This was a pleasing result. With three treble sensitive tracks I re-tried my ‘blind’ treble adjustment test, the results were 0.0, 0.0 & +0.5. Curiously, in subjective terms, I would still say that the mR sounds darker than the SOtM kit, and yet I can pretty much nail the blind treble test on 0.0dB.
Genuinely fascinating. I would now say that the fundamental issue is solved, and I can now see scope for further investigation and improvement. I have been hunting blind with this one, but now I can clearly see my foe and so can go on the attack. (and how cool is Roon? You can adjust the parametric curve on an iPad, sat in the listening position, in real time. Very, very, cool)
In general terms, the rig is now sounding as good as it ever has, plus I can now see easy scope for further optimisation. As one example, I think my room is generally a good one for audio, it is rectangular / symmetrical, a good size (14’ by 20’), carpeted, rear wall broken up, furnished and full of junk. One key negative issue could be ‘first reflection points’, on the left side I have nice floor to ceiling curtains to soften things up a bit, on the right side a solid wall (breeze block and dry wall). Definitely not ideal, definitely not symmetrical, from an audio perspective at least. Plus I have learned that my adjusting the speaker toe-in by basically nothing much can make the difference between a sometimes annoying and fatiguing listen, to audio joy. Definite scope for optimisation.
This does leave one mystery though, why did the 4.75kHz ‘presence’ issue bother me with the SOtM kit, and not with AIR or the mR. Indeed, I watched a couple of movies over the Easter break, with the Blades on front speaker duty fed (analogue) from a Yamaha AV surround decoder. It sounded spectacular. So why did this issue only hit me with the SOtM kit in place?
Considering AIR, the mR and SOtM kit show almost identical frequency curves, a logical conclusion would be that the SOtM kit does actually sound harsher in the presence region. Everyone says the SOtM kit is more transparent, so this is counter-intuitive, but it is logical. Time for another experiment. I set up a very strange Roon parametric EQ, to eliminate bass, mids and emphasise the presence region. Not the most pleasant thing to listen too, but it does allow you to hear the offending region with great clarity. I plan to try this with the mR vs SOtM kit, but frankly I could not be bothered to swap the two today. So as a simple experiment I tried this SOtM versus Roon AIR, which does not require any kit and cable swapping. It is a difficult test but I could hear some differences, very subtle though. One thing is clear, even with this horrific presence only set-up, I can hear nothing fundamentally wrong with the SOtM kit, but things like the snare springs of a snare drum do sound a little different. One very curious result, I tried listening to a snare drum section with this bonkers EQ with the Mutec engaged then disengaged with the SOtM kit, it’s subtle, and I am convinced it was far better with the REF10 in play. Might be expectation bias I suppose, but I am sure it sounded clearer. So clearer presence resolution with enhanced presence frequencies might be enough to trigger some kind of psychoacoustics effect? This might be worth revisiting when I have a little more time. (and can be bothered.) Also, I think the bonkers EQ method might be informative when trying different power options with the SOtM items, in terms of clearly indicating any effect in frequency region that my brain is clearly very sensitive too. The same applies to future SMPS elimination tests and similar.
I have to conclude that sometimes the tiniest things can have a positive or negative effect, and sometimes the tiniest negatives can grab your attention to the elimination of most everything else. This might explain why some people are made very happy with things like USB cable swaps and why seemingly insignificant stuff like changing the power supply to a component can yield perceptible results.
So many things for continued investigation. At a basic level the issue is solved, but the puzzle as to why the mR sounds (subjectively) darker does remain. I am listening to the SOtM kit now, it does sound brighter, I am telling you.
Time for a break!
1000 Pro - KEF Blade - iFi Zen Stream - Mutec REF10 - MC3+USB - Pro-Ject Signature 12
