Sigberg Audio Saranna (fullrange, cardioid active floorstander) development thread

[Duke]

I don't think this coax will ever measure textbook on-axis, but it evens out nicely off-axis, and sounds great. I am not going to throw it out for the sake of the on-axis prettiness.

Oh I hope you don't think that's what I was suggesting! I was replying to @PB63 with a possible explanation for what he was observing in his measurements.

Earl Geddes is the one who pointed out the on-axis dip to me and showed me how it disappeared off-axis, and he considered this to be an example of incomplete measurements and/or incomplete analysis of one's measurements tricking one into trying to solve a non-existent problem. .

 

[sigbergaudio]

Oh I hope you don't think that's what I was suggesting! I was replying to @PB63 with a possible explanation for what he was observing in his measurements.

Earl Geddes is the one who pointed out the on-axis dip to me and showed me how it disappeared off-axis, and he considered this to be an example of incomplete measurements and/or incomplete analysis of one's measurements tricking one into trying to solve a non-existent problem. .

No, I didn't think you were - I was just pointing it out to anyone lurking about wrinkling their nose at the on-axis response.

 

[PB63]

If it's an on-axis dip in a coaxial that disappears off-axis, it MIGHT be the mouth reflection. Let me explain:

There will be a reflection around the perimeter of the mouth. At a given on-axis microphone distance, there will be a frequency at which the total path length of the reflection energy is 1/2 wavelength longer than the path length for the direct sound. And there will be a cancellation dip at this frequency.

In practice the reflection dip is smeared out somewhat so it's a wider-and-shallower dip than the above paragraph implies.

In my opinion, unless the speaker is going to be used near-field, this dip should NOT be corrected by equalization. Doing so will exaggerate that portion of the spectrum pretty much everywhere except in the on-axis first-arrival sound.

I sometimes use round horns which have the same issue, and the work-around I recommend is to listen from about 15 degrees off-axis.

Hello,
Yes, I'm aware of this. In the beginning, I was a bit put off about it. But by averaging the measurements and using white noise at listening spot and listening, I'm very satisfied with them. I sold my Genelec's last year due to the too wide coverage in high treble, my room is very hard. Although I don't blame the Genelec's for that, it's my choice of decor. They're still in weird shallow test box 50cm wide, so not the perfect and last solution. And I'm moving in August, so no time for building. I think the throat is 20mm in the 6'ers, not home for the moment...

 

[sigbergaudio]

Just noticed the graphs in this post: https://www.audiosciencereview.com/...stander-development-thread.49881/post-1973392 had a really wide range on the Y-axis, like 90dB or something, making them look smoother than with the normal 60dB range. That was not intentional believe it or not, fixed it now.

 

[Zambo]

One of my hands down favorite things with the Manta, is the sometimes magic ability to convey richness, clarity and depth of human voices, combined with the unusually large soundstage.

After listening for a while to the new .. open baffle-hybrid versions of the Saranna, I think Manta will get pretty stiff competition. Is all this sound coming from just two speakers?!

Interesting how you've gone from sealed box (SBS.1) to cordioid (Manta) to open baffle (Saranna). Maybe Sigfried Linkwitz was onto something.

 

[Zambo]

I’m sure they all sound wonderful…looking forward to receiving my Mantas+subs

 

[sigbergaudio]

Interesting how you've gone from sealed box (SBS.1) to cordioid (Manta) to open baffle (Saranna). Maybe Sigfried Linkwitz was onto something.

So, my background is originally in IT and in startups, where you experiment your way to product / market fit, and where you experiment your way to what works. I may have one idea when I start, and things evolve during the development process.

The SBS.1 is a design that likely works best in a sealed configuration due to the size and capacity. Manta works very well in the cardioid configuration. With the Saranna there were some resonance issues that I was interested to see if we could solve with opening up the rear, and the result was very good.

Open baffle / dipole configurations certainly has many benefits. The Saranna is essentially a hybrid where you still get the benefit of bass capacity from the port loaded bass section, and the open, natural sound of the open coax driver at the top. It sounds very similar to the Manta.

 

[sigbergaudio]

Have now spent a bit more time on the crossover design, and the crossover between the tweeter and midrange (1600hz) is now phase coherent.

The bass drivers have a 4th order high pass, so that induces a shift, and will probably stay that way. I find the phase linearity to be most important in the higher frequencies.

 

[kemmler3D]

Have now spent a bit more time on the crossover design, and the crossover between the tweeter and midrange (1600hz) is now phase coherent.

The bass drivers have a 4th order high pass, so that induces a shift, and will probably stay that way. I find the phase linearity to be most important in the higher frequencies.

View attachment 375253

The phase looks fantastic, but are we worrying about 4-15Khz right now? It doesn't look crazy, but it doesn't look anything like perfect either, here. Apologies if I've missed some informative posts in the past several pages...

 

[sigbergaudio]

The phase looks fantastic, but are we worrying about 4-15Khz right now? It doesn't look crazy, but it doesn't look anything like perfect either, here. Apologies if I've missed some informative posts in the past several pages...

This is an in-room measurement, so it's not perfect with regards to room noise.

So no, not super worried. Below is an anechoic measurement (15 degrees) from a little while back (the step down at 15-20khz is gone now, and the rest has been revised/adjusted as well). With the exception of the narrow diffraction at 4.6khz, it's +/-2dB from 4-15khz at this angle. That being said, this coax is not as tidy on-axis as the previous I've used - but similar to the other speakers, on-axis will not be the reference/listening axis.

 

[Zambo]

I’m wondering why you emphasize phase coherence over a flat frequency response?
Floyd Toole et al have found that (1) there is so much that disrupts phase along the way that by the time the signal reaches the listeners ears the phase is inevitably corrupted. And that (2) the good news is that they also found that listeners are not sensitive to phase shifts, but we’re much more sensitive to deviations from a flat smooth frequency response.
Just wondering why there is an emphasis on phase, when it doesn’t seem to be something listeners notice, and which is almost impossible to achieve.
Not being critical, just want to know why you as a designer seem to put a special emphasis on phase coherency

 

[sigbergaudio]

I’m wondering why you emphasize phase coherence over a flat frequency response?
Floyd Toole et al have found that (1) there is so much that disrupts phase along the way that by the time the signal reaches the listeners ears the phase is inevitably corrupted. And that (2) the good news is that they also found that listeners are not sensitive to phase shifts, but we’re much more sensitive to deviations from a flat smooth frequency response.
Just wondering why there is an emphasis on phase, when it doesn’t seem to be something listeners notice, and which is almost impossible to achieve.
Not being critical, just want to know why you as a designer seem to put a special emphasis on phase coherency

First of all I would say the jury is out on the audibility of linear phase. It is subtle for sure, but when it can be achieved, I think we should try to do it. For those who do notice it, it's often perceived as improved precision in soundstage and imaging, which are characteristics I find important.

Second of all, I do not emphasize phase coherence over a flat frequency response. There is no penalty in frequency response linearity by introducing phase linear crossover designs.

 

[jaakkopetteri]

A few questions, if I may - I don't really expect an answer to all of them :
The stock Wavecor sub seems to be linear (by Klippel definitions) to ~5mm. Getting it to 11mm seems like a whole another driver, or are you using different metrics for linearity?
Are there any common modifications you would have driver manufacturers do, not just the sub but the coaxes? I assume mostly aesthetic changes, or maybe something like "OK we can pay 30€ more the driver, can you give the driver some more linear stroke"?

Most of your in-room measurements show a pretty flat response throughout the range. Don't you use a target curve with some bass boost if you do any EQ? Or if you did, what would the curve look like? Mostly asking cause you mentioned boosting the 100-500Hz area, but I wonder how you relate that to the usual "Harman" kind of boost under 150Hz or so

 

[sigbergaudio]

A few questions, if I may - I don't really expect an answer to all of them :
The stock Wavecor sub seems to be linear (by Klippel definitions) to ~5mm. Getting it to 11mm seems like a whole another driver, or are you using different metrics for linearity?
Are there any common modifications you would have driver manufacturers do, not just the sub but the coaxes? I assume mostly aesthetic changes, or maybe something like "OK we can pay 30€ more the driver, can you give the driver some more linear stroke"?

Most of your in-room measurements show a pretty flat response throughout the range. Don't you use a target curve with some bass boost if you do any EQ? Or if you did, what would the curve look like? Mostly asking cause you mentioned boosting the 100-500Hz area, but I wonder how you relate that to the usual "Harman" kind of boost under 150Hz or so

I am in Spain with no computer, so I will get back to the more detailed questions in a week or so.

With regards to the curve, all the in-room measurements I've shared have had a bass boost similar to Harman? Perhaps you've been looking at some measurements that didn't show the entire frequency range?

As a customer you won't need to eq to get a "Harman curve", the Saranna will have this naturally.

 

[jaakkopetteri]

all the in-room measurements I've shared have had a bass boost similar to Harman

Not post #266, but I might have exaggerated about most measurements not having it and got confused by some of your REW measurements being quasi-anechoic. #218 looks almost flat too apart from <60Hz or so

 

[sigbergaudio]

@jaakkopetteri with regards to xmax I suspect you have it confused with a different driver.

"Displacement limiting numbers calculated by the Klippel analyzer using the subwoofer criteria for Bl was XBl at 70% (Bl dropping to 70% of its maximum value) equal to 12.1 mm for the prescribed 20% distortion level (the criterion for subwoofers). For the compliance, crossover (XC) at 50% CMS minimum was 12.5 mm, which means that for the SW223BD03, the Bl is the more limiting factor for getting to the 20% distortion level. However, both numbers were greater than the driver’s 11.7 mm physical XMAX."

 

[kelesh]

Passive version?

 

[sigbergaudio]

Passive version?

Nope.

 

[sigbergaudio]

Not post #266, but I might have exaggerated about most measurements not having it and got confused by some of your REW measurements being quasi-anechoic. #218 looks almost flat too apart from <60Hz or so

This measurement (#266)was at 1m, not in the listening position. 218 was a bit flat, but that was an early prototype from last year, it's different now.

 

[jaakkopetteri]

@jaakkopetteri with regards to xmax I suspect you have it confused with a different driver.

Indeed, I looked at the BD02 mentioned earlier in this thread - that explains it