Sony SS-CS5 3-way Speaker Review

[QMuse]

I’m about to be out of the house for a while, so it’ll have to wait (I foolishly deleted the rows to match the octave smoothing rather than hide them, so I manually will have to do the adjustment again).

No problem and no rush, whenever you catch time to do it is fine - thank you!

 

[QMuse]

It seems wrong to me too, yet a difficult conclusion to escape.

You are absolutely right, assuiming equally falt on-axis forumla would reward more negative slope with PIR, even when it leads to absurdly negative slopes. I find that ahrd to accept as intuitively I believe there is such thing as optimal PIR slope instead of "the more negative, the better" logic.

 

[napilopez]

Exactly, it has nothing to do with real world. One can argue that on-axis, as a single point measurement is hardly a better representative of direct sound than LW. I remain sceptical, especially considering such high weighting.

Yeah, my point is I think it made sense for the paper, but it may not make total sense for the preference scores here. MZKM calculates it using LW as well in his spreadsheet, so I personally tend to look at that. Almost all speakers get a boost, but some get it more than others.

Well, it‘s also my understanding the waveguides inherently cause issues on-axis that are not seen even a few degrees off-axis. Since Harman uses waveguides, it would benefit them to focus on the listening window.

True, but those also disappear usually within 10-15 degrees, let alone 30. Then there's the fact most people aren't listening on axis and it's easy to see why that can be a problem. So LW is my preference.

On a separate note, as I said I would earlier in the thread, I tried calculating scores for the Sonos move which is quasi omni. Turns out it may not be a great test case for this as the speaker basically has a bit of a built in tilt starting on axis, so the PIR does end up having a tilt anyway. The super ragged on-axis also complicates things. Though tbf it seems many omnis have a bit of a built-in tilt too.



But for the curious, it gets a 4/5.8, which isn't terrible. It might be better in real life because my measurements used broad 15-degree intervals and incomplete rear hemisphere data, so the data doesn't 'smooth out' as much as it would in real life, plus I doubt the HF high-q squiggles on axis are as audible as they look. For example, here's how the full horizontal data averages out:


If I use 1/6 octave smoothing which may be more representative of real life in this specific case it gets a 5.5/7.3, with NBD on-axis unsurprisingly jumping forward the most. Here's what those curves look like:



So omni's don't necessarily suffer too badly if the on axis has some tilt. Granted, the move isn't a true omni, more like a super-wide, but it's a unique case. It sounded very timbrally neutral to me too, which isn't a surprise as at 130cm the room curve looked like this:

this is a compression test, but it still shows the in-room response ends up surprisingly smooth and with the proper tilt.


Looking forward to when Amir gets to test more unusual designs.

 

[MZKM]

Looking forward to when Amir gets to test more unusual designs.

I wonder what the measurements for an Ohm Walsh would look like, as the tweeters are aimed upwards and diagonally.

 

[richard12511]

I wonder what the measurements for an Ohm Walsh would look like, as the tweeters are aimed upwards and diagonally.

I really want to see this. I'm also really curious to see how omni designs do in mono listening tests.

 

[GXAlan]

I have a very oddball speaker to test in the future.

50/50 odds on a Saluting Soldier Panther purely for the oddballness.

 

[edechamps]

@QMuse @andreasmaaan There has been a lot of debate previously on the interpretation of the SM metric. It is a complicated topic that is headache-inducing and quite counter-intuitive. I would urge you to read this post if you are interested in understanding how SM is defined and how it works. The current version of Loudspeaker Explorer now has SM computation as a feature and you can see a detailed computation with charts for every step (under "Olive Preference Score", "SM"). See here for an example computation.

The simplest definition of SM I was able to come up with is: SM describes how much of the curve deviation from a flat, horizontal line can be explained by the overall slope (as opposed to just jagginess). Which is hard enough to wrap your head around as it is. And even then, that description glosses over a few details.

And to further clarify: Assuming that all data points are equally-spaced from the PIR slope (and all else is equal), a speaker with a steeper PIR slope will always receive a better rating for SM_PIR. Correct?

And furthermore, there is no effective limit to this, i.e. a hypothetical speaker with a PIR slope approaching ∞ dB/decade would (all else equal) achieve a superior rating for SM_PIR than a speaker with a shallower PIR slope (once again assuming all data points are equally-spaced from each speaker's respective PIR slope). Also correct?

Yes, this is correct on both points. However, do keep in mind that the total Olive score might not improve with larger slope, because NBD_PIR is part of the score too, and that variable worsens as slope increases. So in the end, there is a delicate (and non-linear, I suspect) balancing act going on between SM_PIR and NBD_PIR. For this reason, it is very difficult to draw conclusions from SM_PIR or NBD_PIR in isolation.

Given that correlation coefficient biases SM in favour of steeper slope, why do we speculate that Olive's model uses it (as opposed to a formula that is independent of it)?

It's not speculation. The Olive paper literally defines SM as the squared correlation coefficient.

It seems also the case that correlation coefficient would be biased in favour of a steeply upward-sloping SM_PIR slope - is this correct?

Yep, and actually that's a very good point which I didn't realize until now. That's a great illustration of why the model might perform poorly with speakers that do not behave similarly to the speakers used in the original study!

I don't think so. -30deg slope and +30deg slope should result in the same corr. coefficient, just the sign would be opposite.

Sadly, SM is defined as r², i.e. the correlation coefficient squared. It will favor any kind of tilt in any direction.

Ok, so do we think that Olive deliberately built in SM_PIR such that it would advantage speakers with the steepest possible downward-sloping PIR?

I doubt that. Looking at the wording of the paper, it's way more likely Olive defining SM as r² was an accident, because it makes very little sense to use r² to represent the "smoothness" of a curve. It's quite possible Olive got confused about what r² meant, which is understandable given that @MZKM, @bobbooo, myself and others were just as confused too (@daverosenthal is probably the only one who got it right from the very beginning). That doesn't make the model invalid though, just weirder and harder to reason about.

 

[QMuse]

Yep, and actually that's a very good point which I didn't realize until now. That's a great illustration of why the model might perform poorly with speakers that do not behave similarly to the speakers used in the original study!

Sadly, SM is defined as r², i.e. the correlation coefficient squared. It will favor any kind of tilt in any direction.

Uh, thank you for this explanation @edechamps , but this.. this is sooo weird.

 

[BYRTT]

.....Those 2 radar charts are practically identical, but LW and PIR curves aren't.

Webservice:

 

[GelbeMusik]

The simplest definition of SM I was able to come up with is: ...

Really? Where is that formula? I'm sure I've seen worse ...

 

[andreasmaaan]

@edechamps many thanks for the great explanation, which basically clarified all the unknowns that remained for me

I doubt that. Looking at the wording of the paper, it's way more likely Olive defining SM as r² was an accident, because it makes very little sense to use r² to represent the "smoothness" of a curve. It's quite possible Olive got confused about what r² meant, which is understandable given that @MZKM, @bobbooo, myself and others were just as confused too (@daverosenthal is probably the only one who got it right from the very beginning). That doesn't make the model invalid though, just weirder and harder to reason about.

The idea that it was an accident seems quite plausible, I agree.

It does seem particularly strange given the "balancing act" (or perhaps "tug of war" would be a better analogy!) you described between SM_PIR and NBD_PIR.

It's not speculation. The Olive paper literally defines SM as the squared correlation coefficient.

Here I think you misread my earlier post. What I actually asked was:

Given that correlation coefficient biases SM in favour of steeper slope, why do we speculate that Olive's model uses it (as opposed to a formula that is independent of it)?

I guess now what I'd love to know now is how Olive arrived at each of the formulae defining each of the parameters.

Did he use a kind of intuitive trial and error? Did he design an algorithm that did it (I presume not)? Was some other process involved?

 

[andreasmaaan]

However, do keep in mind that the total Olive score might not improve with larger slope, because NBD_PIR is part of the score too, and that variable worsens as slope increases. So in the end, there is a delicate (and non-linear, I suspect) balancing act going on between SM_PIR and NBD_PIR. For this reason, it is very difficult to draw conclusions from SM_PIR or NBD_PIR in isolation.

Just some further off-the-cuff speculation here, mostly to add to the weirdness:

Given steepness of slope positively affects SM_PIR while negatively affecting NBD_PIR, surely this would imply a maximum possible preference rating that is in fact below 10.

Which seems particularly counter-intuitive given that the preference rating is "out of" 10.

 

[napilopez]

Just some further off-the-cuff speculation here, mostly to add to the weirdness:

Given steepness of slope positively affects SM_PIR while negatively affecting NBD_PIR, surely this would imply a maximum possible preference rating that is in fact below 10.

Which seems particularly counter-intuitive given that the preference rating is "out of" 10.

Well, it's certainly exceedingly difficult to get a 10, but it's not impossible to get it in the 9s. The D&D 8C is the best scoring speaker of the ones I've measured and plugged into MZKM's sheet, with an 8.4 Ignore LFX score. Out of curiosity, I tried smoothing the measurement to 1/1 per octave before exporting and that got a 9.3, and those aren't perfectly straight lines. So I'm sure one could devise a graph that would score in the high 9s.

All this assuming nothing has changed since MZKM last shared his sheet with me.

 

[napilopez]

Actually @andreasmaaan I might've spoken too soon. I also smoothed the Neumann KH80 and Buchardt S400 as well, which both scored 8.1s with my normal 1/24 smoothed measurements, and curiously, they also maxed out at roughly 9.3 when I smoothed them all the way (ignore LFX). So maybe that's just about the limit.

 

[andreasmaaan]

Actually @andreasmaaan I might've spoken too soon. I also smoothed the Neumann KH80 and Buchardt S400 as well, which both scored 8.1s with my normal 1/24 smoothed measurements, and curiously, they also maxed out at roughly 9.3 when I smoothed them (ignore LFX). So maybe that's just about the limit.

Perhaps you could share the formulae? Presumably the maximum rating can be deduced from the relationship between the formulae (although I'm almost certainly not the person to work out mathematically what it is).

 

[napilopez]

Perhaps you could share the formulae? Presumably the maximum rating can be deduced from the relationship between the formulae (although I'm almost certainly not the person to work out mathematically what it is).

I'm not your guy either, I barely got through calculus . MZKM shared his master sheet here though.

 

[Music1969]

 

[amirm]

View attachment 65495

That's strange. I can't find any review from them much less measurements.

 

[andreasmaaan]

Actually @andreasmaaan I might've spoken too soon. I also smoothed the Neumann KH80 and Buchardt S400 as well, which both scored 8.1s with my normal 1/24 smoothed measurements, and curiously, they also maxed out at roughly 9.3 when I smoothed them all the way (ignore LFX). So maybe that's just about the limit.

Actually @napilopez, I just remembered that Olive gives the target slopes for the top-rated speakers from the sample in the study.

If you plug the slope values for NBD_PIR, SM_PIR and NBD_ON from the "All Tests" column into the equation on the right and ignore LFX, you should end up with a the slopes that achieve the best possible rating, as (IIUC) it was these target slope values that the model was based upon in the first place.

 

[QMuse]

Actually @napilopez, I just remembered that Olive gives the target slopes for the top-rated speakers from the sample in the study.

If you plug the slope values for NBD_PIR, SM_PIR and NBD_ON from the "All Tests" column into the equation on the right and ignore LFX, you should end up with a the slopes that achieve the best possible rating, as (IIUC) it was these target slope values that the model was based upon in the first place.

View attachment 65500

Slope of -2.1 equals to 15dB drop over the range of 20Hz-20kHz.

Slope of -1.75 equals to 12dB drop over the range of 20Hz-20kHz.