This thread should have been so simple. A case of someone happy with their audio system who then removes one component, replaces it with some items that are almost universally reckoned to be significantly superior, and then reports the glorious sound quality improvements that result. What could be simpler? The trouble is, this is not how things have worked out. In simple terms, I have gone from a system that I have been happy with, to one that is now irritating me. As has been reported in this thread, this has been a bit of a baffling journey, at first it seemed to be a case of the SOtM kit having a bit of HF harshness.
Interestingly, a lot of SOtM users have reported stuff along the lines of “it sounded a bit harsh when new, then things improved after 200 hours burn in, and at 400 hours it was really singing”, you know the kind of stuff. Others had good ideas, suggesting that the SOtM’s inherent transparency was exposing shortcoming in other areas of my system, so I need to eliminate SMPS’s, get better cables, optimise the PC that feeds the sMS-200Ultra, maybe add some LPSU’s for the SOtM kit to replace the sPS-500, and so on. Some of this might prove to be good advice, and it is standard stuff, improve transparency in one area and it exposes weaknesses elsewhere, I could sit here typing for hours giving many examples of this phenomena, but I am guessing that most people reading this will be more than familiar with this concept.
The trouble is, the more I experimented, tried different things, listened very carefully to what the system was doing, it became clear to me that none of the above was explaining what I was experiencing. To be clear, I am not stating here that nothing written above may offer an improvement, indeed I plan to try some of the ideas written above, but the key point is that none of this explains what I have experienced when adding the SOtM kit, it just does not fit the facts.
So why does none of this fit the facts? As an example, take a specific section of a track that now annoys me with the SOtM kit. Listen very carefully, A/B switch between the SOtM kit and the mR, the bit section that is annoying me is actually more clearly resolved with the SOtM kit, it sounds more realistic with the SOtM kit. Switching between the mR and SOtM kit does require a lot of cable swapping, then software swapping (Sonicorbiter to Eunhasu), so this is not particularly rapid, but what I could do is swap between Roon AIR and the SOtM kit very rapidly, and Roon AIR is a very reasonable proxy for the mR / MC3+USB combination. With rapid A/B switching, it was clear that aspects that were annoying me in the music were quite simply being reproduced more realistically with the SOtM kit. So the SOtM kit was subjectively higher performance, subjectively better, but subjectively not enjoyable to listen to. So following this logic, if improvements are making the system less enjoyable to listen to, creating some kind of psychoacoustic effect of harshness when none is there, then any other improvements, to power supplies or anything else, will simple exacerbate the issue. Indeed, the cure seemed to be to reinstall the mR, which is lower performance and lacks the SOtM’s kit better clock performance and superior noise rejection. Utterly baffling!
Or maybe not. A short session with a friend on Sunday afternoon and things became clearer, and indeed many thoughts came into my head, all of a sudden memories of past experiences started to connect and I started to make full sense of what is going on. The session last Sunday was a simple thing, I was starting to think very clearly about what was happening, and was about to try something myself, I was playing with Roon’s parametric EQ, and had simply tweaked the problematic 4.75kHz region down by 3dB. Anyway, with this tweak in place, my friend listed to a couple of tracks. They had no idea about the EQ, but did know I was messing about with new kit. One track in, my friend made some positive noises about everything sounding good and then said something along the lines of “the cymbals were sensational in that that track, I don’t know how many time I have heard that track but never heard that before”. So an overall positive response, with the performance in the HF region getting special mention. I then mumbled something about different configurations of the new equipment, and played the same track again, this time with Roon’s EQ turned off. My friends response? No, not as good as the first one, the first one was much better. OK, this was not the most extensive and scientific blind test, it was over in 15 minutes, then we both had other things to do, but it did get me thinking.
Back to the concept of improving an aspect of the system, and this makes thing sound worse. It dawned on me that I have experienced this before, and recently too, when replacing the D800 with the 1000Pro. Is the 1000Pro better than the D800? Yes, definitely. Indeed, I have had the luxury of a rapid A/B switch test between the D800 and 100Pro, thanks to the good people at Oxford Audio. One thing I remember about this audition was how the 1000Pro had far better resolution, was far more realistic sounding, most notably in the upper mid-range area. This test was imprinted in my memory, the 1000Pro clearly better in the higher registers. I can also recall my last weekend with the D800, there was one track in particular that I was listening too and thinking “this sounds sensational, I can’t imagine that with the 1000Pro it will sound better than this.” As things turned out, it didn’t. I remember listening to this same track once I had the 1000Pro, and rather than experiencing immediate audio joy, I started to notice some issues in the HF area. Sound familiar? With the 1000Pro though, this made sense, I knew from the Oxford Audio audition that the Pro is better in this regard, so the Pro’s improved transparency was simply exposing some hash in the recording, or so I thought. I was also thinking that maybe the 1000Pro deserved a better front end than the mR/Mutec, for the record the Oxford Audio audition was with the Aurender N10, a source I like a lot with the Devialet. I have had the 1000Pro for a while now, and I have no doubt that it is superior to the D800. That is only part of this story though, there have been times with the 100Pro when I am listening to something and miss the D800’s slightly more relaxed presentation, despite being the ‘inferior’ performer, it was somehow nicer to listen to, or perhaps more accurately characterised as nicer to listen too with some source material. So over the last year or so, maybe I play something and there is a touch of harshness in the HF, I have put it down to the recording, or maybe start thinking that the 1000Pro deserves a better front end. Looking back, I think this was wrong. Certainly the 1000Pro deserves a better front end, or at least it can benefit from one, but the swapping of the D800 for the 1000Pro has created exactly the same phenomena as swapping the mR for the SOtM kit, so this was the root cause of the niggling thoughts about HF harshness.
Things are now making sense to me, I almost feel like I have discovered some new audio phenomena. I say almost, because I am sure that many must have experienced this before me, and what I am experiencing may be well known to audio specialists or those who study psychoacoustics, I simply do no know. Anyway, this is what I now believe is happening. I think there is some kind of psychoacoustic effect, that when you improve the quality of reproduction, improve the resolution, it makes the frequencies observed in the presence band more apparent, it does not increase them, it just makes them more apparent when listening. This may or may not just be in the presence or HF region, I do no know, but that is where I am experience this phenomena now. To clarify, if I use REW and look at the in-room frequency curve, I have the normal peaks and nulls, a fairly typical bit of bass reinforcement, followed by a thankfully shallow null, normal stuff. One key point here is that ignoring the peaks and dips, the frequency curve is pretty flat from say 200Hz to over 10kHz. I think this is significant, because many room correction ‘house curves’ tend to have a gentle slope from bass down to the HF region. I have never seen a house curve that increases the HF region compared to mid range. A recall reading some material about room correction, this stated that you tend to hear frequency as hearing the peaks in the curve. So if I do have a peak in the presence band, this will tend to be prominent in my listening. Add the 1000Pro, this improves resolution, the presence band becomes more prominent, or perhaps more accurately the perception of the presence band becomes more prominent, then add the SOtM kit, the same thing happens and the effect becomes cumulative. This then tips me over the edge, from something I can live with, something I enjoy, to a situation where the peaks in the presence band annoy me so much I cannot enjoy all the other good stuff that is going on. Using this forum, I can read through old posts that I made after getting the 1000Pro, and indeed posts by others after upgrading to the Pro. I can see I had indeed experienced this phenomena moving to the 1000Pro. I suspect others have too when moving from Expert to Pro, there are quite a few posts about HF harshness or the Pro being less forgiving with respect to treble. Of course, the experience of others will be system dependent. What may be a negative in one system because of the phenomena I describe, could easily be a positive in another system where the in room presence band response is less prominent, where increased perception of the presence band would simply be perceived as greater clarity. So that is the theory, I could write a couple more pages as to why I think the phenomena I am describing is very real, for now, please trust me, I am as convinced as I can be. Plus, this is a theory which lends itself to a little experimentation.
So, armed with REW and Roon’s parametric EQ, I decided to tweak the EQ to remove the peaks in the presence band and thorough to the treble region. Of course, I could try to correct bass nodes and try to mimic a house curve whilst I am at it, but for now, I want to change as little as possible, and just take out the peaks from about 2kHz upwards. I am not an expert in room correction, in fact it is something that I have quite deliberately avoided, it somehow does not fit with my more purist approach to things. So for today, I simply tried doing a minor room correction tweak via a method I understand, then test with REW, tweak the Roon EQ, retest and check if I have the desired curve with presence band peaks removed, then repeat until optimised. A bit crude, but it is a method I understand and can do without spending hours messing about leaning new tricks. Well, that was the plan anyway. When I performed the first baseline measurement, I found that the 4.75kHz peak mentioned previously was not there, what I had now was a 5.75kHz peak. I think I know why this is, last Sunday evening I noticed that the left hand Blade had less toe in than normal, something I had been messing about with earlier in the day. This annoyed me, it just looked a little odd. So I tweaked it back to where it was, or so I thought. OK, I still have a peak, but it has moved, it just goes to show how sensitive the measurements are with respect to minor speaker changes, maybe my bogus measurements in post #131 were correct after all. I also found that that the corrections I thought would be easy to make in Roon, were in fact not easy to make at all. I am new to room correction, and I can now see I have much to learn, but from where I am now the changes I can make with the Roon parametric EQ are nowhere near accurate enough to do anything with my specific peaks, all I can do is get a general reduction across a couple of peaks, Roon’s EQ seams to average things out too much for what I am trying to achieve. In fact, with a bit of playing around this morning, i can get just as nice a curve, with a bit of downward slope, by simply tuning off Roon’s EQ and tweaking the treble down on the Devialet.
I can see some potential with room correction, but I need to do a lot more reading up and learning before I can get it to do what I would like it to do. Also, I would probably be better off investigating some improvements with room treatment first. With the room improved there would be less for room correction to have to correct in the first place. All this will take time, and I do not have much free time at the moment. So, in the short term, I can best achieve what I am trying to do with a simple tweak on the treble control, crude, but no worse than I can achieve with Roon EQ at this point in time. By chance, it would also seem that my semi-accidental speaker positioning tweak has helped. I recall listening to the Blades at Oxford audio one time and I thought the system sounded a little dark, I visited again a week later, this time they did not sound dark at all, I think minor differences in positioning and toe-in make a big difference, maybe more with the Blades than some other speakers.
So as an experiment today, I tweaked the treble to -4dB, which should be enough to kill the presence band issue stone dead, and I sat down to listen to some familiar tracks. Not far through the first track it became clear that the system actually sounded a little dull. These are the margins I am talking about here. In fact, a tweak down -1dB or -2dB seems to be all that is needed.
Realising I am going full circle to a degree here, I thought I would do a simple brain re-set and listen to some familiar demo tracks with Roon AIR, this is an effective mR+MC3+USB proxy and takes me back to where I was before the REF10 and SOtM additions. This was actually a bit of a revelation, Roon AIR sounded dreadful! OK, we are talking relative terms and margins here, so it was not dreadful in absolute terms, but the step back was stark. I think I had been so focused on one aspect of the systems presentation that I had not really noticed just how big a step forward the REF10 and SOtM kit was making.
So after much messing about and much pondering, my conclusion is that the next step is to simply set the treble down by -2dB or similar, this appears to be all that is needed to kill the presence issue and prevent it bothering me with some tracks. Then I can simply see how things go, it should end my brain's preoccupation with the prescence band issue, and I might start to fully appreciate all that goodness that the REF10 and SOtM trio are delivering.
As a final comment, if someone can explain why increasing a systems resolution can increase the apparent prominence in the prescence band, I would be most interested to learn why.
Interestingly, a lot of SOtM users have reported stuff along the lines of “it sounded a bit harsh when new, then things improved after 200 hours burn in, and at 400 hours it was really singing”, you know the kind of stuff. Others had good ideas, suggesting that the SOtM’s inherent transparency was exposing shortcoming in other areas of my system, so I need to eliminate SMPS’s, get better cables, optimise the PC that feeds the sMS-200Ultra, maybe add some LPSU’s for the SOtM kit to replace the sPS-500, and so on. Some of this might prove to be good advice, and it is standard stuff, improve transparency in one area and it exposes weaknesses elsewhere, I could sit here typing for hours giving many examples of this phenomena, but I am guessing that most people reading this will be more than familiar with this concept.
The trouble is, the more I experimented, tried different things, listened very carefully to what the system was doing, it became clear to me that none of the above was explaining what I was experiencing. To be clear, I am not stating here that nothing written above may offer an improvement, indeed I plan to try some of the ideas written above, but the key point is that none of this explains what I have experienced when adding the SOtM kit, it just does not fit the facts.
So why does none of this fit the facts? As an example, take a specific section of a track that now annoys me with the SOtM kit. Listen very carefully, A/B switch between the SOtM kit and the mR, the bit section that is annoying me is actually more clearly resolved with the SOtM kit, it sounds more realistic with the SOtM kit. Switching between the mR and SOtM kit does require a lot of cable swapping, then software swapping (Sonicorbiter to Eunhasu), so this is not particularly rapid, but what I could do is swap between Roon AIR and the SOtM kit very rapidly, and Roon AIR is a very reasonable proxy for the mR / MC3+USB combination. With rapid A/B switching, it was clear that aspects that were annoying me in the music were quite simply being reproduced more realistically with the SOtM kit. So the SOtM kit was subjectively higher performance, subjectively better, but subjectively not enjoyable to listen to. So following this logic, if improvements are making the system less enjoyable to listen to, creating some kind of psychoacoustic effect of harshness when none is there, then any other improvements, to power supplies or anything else, will simple exacerbate the issue. Indeed, the cure seemed to be to reinstall the mR, which is lower performance and lacks the SOtM’s kit better clock performance and superior noise rejection. Utterly baffling!
Or maybe not. A short session with a friend on Sunday afternoon and things became clearer, and indeed many thoughts came into my head, all of a sudden memories of past experiences started to connect and I started to make full sense of what is going on. The session last Sunday was a simple thing, I was starting to think very clearly about what was happening, and was about to try something myself, I was playing with Roon’s parametric EQ, and had simply tweaked the problematic 4.75kHz region down by 3dB. Anyway, with this tweak in place, my friend listed to a couple of tracks. They had no idea about the EQ, but did know I was messing about with new kit. One track in, my friend made some positive noises about everything sounding good and then said something along the lines of “the cymbals were sensational in that that track, I don’t know how many time I have heard that track but never heard that before”. So an overall positive response, with the performance in the HF region getting special mention. I then mumbled something about different configurations of the new equipment, and played the same track again, this time with Roon’s EQ turned off. My friends response? No, not as good as the first one, the first one was much better. OK, this was not the most extensive and scientific blind test, it was over in 15 minutes, then we both had other things to do, but it did get me thinking.
Back to the concept of improving an aspect of the system, and this makes thing sound worse. It dawned on me that I have experienced this before, and recently too, when replacing the D800 with the 1000Pro. Is the 1000Pro better than the D800? Yes, definitely. Indeed, I have had the luxury of a rapid A/B switch test between the D800 and 100Pro, thanks to the good people at Oxford Audio. One thing I remember about this audition was how the 1000Pro had far better resolution, was far more realistic sounding, most notably in the upper mid-range area. This test was imprinted in my memory, the 1000Pro clearly better in the higher registers. I can also recall my last weekend with the D800, there was one track in particular that I was listening too and thinking “this sounds sensational, I can’t imagine that with the 1000Pro it will sound better than this.” As things turned out, it didn’t. I remember listening to this same track once I had the 1000Pro, and rather than experiencing immediate audio joy, I started to notice some issues in the HF area. Sound familiar? With the 1000Pro though, this made sense, I knew from the Oxford Audio audition that the Pro is better in this regard, so the Pro’s improved transparency was simply exposing some hash in the recording, or so I thought. I was also thinking that maybe the 1000Pro deserved a better front end than the mR/Mutec, for the record the Oxford Audio audition was with the Aurender N10, a source I like a lot with the Devialet. I have had the 1000Pro for a while now, and I have no doubt that it is superior to the D800. That is only part of this story though, there have been times with the 100Pro when I am listening to something and miss the D800’s slightly more relaxed presentation, despite being the ‘inferior’ performer, it was somehow nicer to listen to, or perhaps more accurately characterised as nicer to listen too with some source material. So over the last year or so, maybe I play something and there is a touch of harshness in the HF, I have put it down to the recording, or maybe start thinking that the 1000Pro deserves a better front end. Looking back, I think this was wrong. Certainly the 1000Pro deserves a better front end, or at least it can benefit from one, but the swapping of the D800 for the 1000Pro has created exactly the same phenomena as swapping the mR for the SOtM kit, so this was the root cause of the niggling thoughts about HF harshness.
Things are now making sense to me, I almost feel like I have discovered some new audio phenomena. I say almost, because I am sure that many must have experienced this before me, and what I am experiencing may be well known to audio specialists or those who study psychoacoustics, I simply do no know. Anyway, this is what I now believe is happening. I think there is some kind of psychoacoustic effect, that when you improve the quality of reproduction, improve the resolution, it makes the frequencies observed in the presence band more apparent, it does not increase them, it just makes them more apparent when listening. This may or may not just be in the presence or HF region, I do no know, but that is where I am experience this phenomena now. To clarify, if I use REW and look at the in-room frequency curve, I have the normal peaks and nulls, a fairly typical bit of bass reinforcement, followed by a thankfully shallow null, normal stuff. One key point here is that ignoring the peaks and dips, the frequency curve is pretty flat from say 200Hz to over 10kHz. I think this is significant, because many room correction ‘house curves’ tend to have a gentle slope from bass down to the HF region. I have never seen a house curve that increases the HF region compared to mid range. A recall reading some material about room correction, this stated that you tend to hear frequency as hearing the peaks in the curve. So if I do have a peak in the presence band, this will tend to be prominent in my listening. Add the 1000Pro, this improves resolution, the presence band becomes more prominent, or perhaps more accurately the perception of the presence band becomes more prominent, then add the SOtM kit, the same thing happens and the effect becomes cumulative. This then tips me over the edge, from something I can live with, something I enjoy, to a situation where the peaks in the presence band annoy me so much I cannot enjoy all the other good stuff that is going on. Using this forum, I can read through old posts that I made after getting the 1000Pro, and indeed posts by others after upgrading to the Pro. I can see I had indeed experienced this phenomena moving to the 1000Pro. I suspect others have too when moving from Expert to Pro, there are quite a few posts about HF harshness or the Pro being less forgiving with respect to treble. Of course, the experience of others will be system dependent. What may be a negative in one system because of the phenomena I describe, could easily be a positive in another system where the in room presence band response is less prominent, where increased perception of the presence band would simply be perceived as greater clarity. So that is the theory, I could write a couple more pages as to why I think the phenomena I am describing is very real, for now, please trust me, I am as convinced as I can be. Plus, this is a theory which lends itself to a little experimentation.
So, armed with REW and Roon’s parametric EQ, I decided to tweak the EQ to remove the peaks in the presence band and thorough to the treble region. Of course, I could try to correct bass nodes and try to mimic a house curve whilst I am at it, but for now, I want to change as little as possible, and just take out the peaks from about 2kHz upwards. I am not an expert in room correction, in fact it is something that I have quite deliberately avoided, it somehow does not fit with my more purist approach to things. So for today, I simply tried doing a minor room correction tweak via a method I understand, then test with REW, tweak the Roon EQ, retest and check if I have the desired curve with presence band peaks removed, then repeat until optimised. A bit crude, but it is a method I understand and can do without spending hours messing about leaning new tricks. Well, that was the plan anyway. When I performed the first baseline measurement, I found that the 4.75kHz peak mentioned previously was not there, what I had now was a 5.75kHz peak. I think I know why this is, last Sunday evening I noticed that the left hand Blade had less toe in than normal, something I had been messing about with earlier in the day. This annoyed me, it just looked a little odd. So I tweaked it back to where it was, or so I thought. OK, I still have a peak, but it has moved, it just goes to show how sensitive the measurements are with respect to minor speaker changes, maybe my bogus measurements in post #131 were correct after all. I also found that that the corrections I thought would be easy to make in Roon, were in fact not easy to make at all. I am new to room correction, and I can now see I have much to learn, but from where I am now the changes I can make with the Roon parametric EQ are nowhere near accurate enough to do anything with my specific peaks, all I can do is get a general reduction across a couple of peaks, Roon’s EQ seams to average things out too much for what I am trying to achieve. In fact, with a bit of playing around this morning, i can get just as nice a curve, with a bit of downward slope, by simply tuning off Roon’s EQ and tweaking the treble down on the Devialet.
I can see some potential with room correction, but I need to do a lot more reading up and learning before I can get it to do what I would like it to do. Also, I would probably be better off investigating some improvements with room treatment first. With the room improved there would be less for room correction to have to correct in the first place. All this will take time, and I do not have much free time at the moment. So, in the short term, I can best achieve what I am trying to do with a simple tweak on the treble control, crude, but no worse than I can achieve with Roon EQ at this point in time. By chance, it would also seem that my semi-accidental speaker positioning tweak has helped. I recall listening to the Blades at Oxford audio one time and I thought the system sounded a little dark, I visited again a week later, this time they did not sound dark at all, I think minor differences in positioning and toe-in make a big difference, maybe more with the Blades than some other speakers.
So as an experiment today, I tweaked the treble to -4dB, which should be enough to kill the presence band issue stone dead, and I sat down to listen to some familiar tracks. Not far through the first track it became clear that the system actually sounded a little dull. These are the margins I am talking about here. In fact, a tweak down -1dB or -2dB seems to be all that is needed.
Realising I am going full circle to a degree here, I thought I would do a simple brain re-set and listen to some familiar demo tracks with Roon AIR, this is an effective mR+MC3+USB proxy and takes me back to where I was before the REF10 and SOtM additions. This was actually a bit of a revelation, Roon AIR sounded dreadful! OK, we are talking relative terms and margins here, so it was not dreadful in absolute terms, but the step back was stark. I think I had been so focused on one aspect of the systems presentation that I had not really noticed just how big a step forward the REF10 and SOtM kit was making.
So after much messing about and much pondering, my conclusion is that the next step is to simply set the treble down by -2dB or similar, this appears to be all that is needed to kill the presence issue and prevent it bothering me with some tracks. Then I can simply see how things go, it should end my brain's preoccupation with the prescence band issue, and I might start to fully appreciate all that goodness that the REF10 and SOtM trio are delivering.
As a final comment, if someone can explain why increasing a systems resolution can increase the apparent prominence in the prescence band, I would be most interested to learn why.
1000 Pro - KEF Blade - iFi Zen Stream - Mutec REF10 - MC3+USB - Pro-Ject Signature 12
