Sigberg Audio Manta (12" wideband cardioid active speakers) development thread

[sigbergaudio]

As mentioned in the SBS.1 thread, we've visited Seas and borrowed their Klippel.

Which means we've verified the cardioid effect in a proper anechoic chamber.

Here's 0 & 180 degrees. As can be seen, there's a reel effect all the way down, allthough somewhat limited under 200hz. From 600hz and up the attenuation is 20dB or more.

To get a clearer picture of how much of an effect the acoustic cardioid ports have, here's a similar measurement of our SBS.1 speakers that have traditional sealed enclosures:

 

[sigbergaudio]

And 0-90.

A little inconsistent directivity in the treble area, but remember this is a pretty rough prototype enclosure. There are a number of ports and holes in the baffle and whatnot, and all of these are with almost no rounding, so this will be improved. Otherwise reasonably even directivity.

And 15deg / listening axis separately:

 

[thewas]

And 0-90.

A little inconsistent directivity in the treble area, but remember this is a pretty rough prototype enclosure.

The problem is unfortunately more driver related though with almost all OEM coaxial drivers, there is a reason why the KEF and Genelec coaxial drivers are they way the are with specially designed transitions from tweeter to woofer and KEF being optimising them for 30+ years and 13 generations now. Hope though it will improve like you say from the enclosure and looking forward to more measurements like this.

 

[sigbergaudio]

The problem is unfortunately more driver related though with almost all OEM coaxial drivers, there is a reason why the KEF and Genelec coaxial drivers are they way the are with specially designed transitions from tweeter to woofer and KEF being optimising them for 30+ years and 13 generations now. Hope though it will improve like you say from the enclosure and looking forward to more measurements like this.

I'm not sure about that, this coax is incredible well behaved. I revised the crossover for our SBS.1 speakers today (they use the same coax driver) after seeing a somewhat similar pattern on that. I don't think the results are horrible.

Here are an earlier measurement of the treble and midrange invidiually (still on the SBS speakers), Note: This isn't on-axis but measured at 60 degrees off-axis. Again I'd say it's pretty well behaved.

 

[thewas]

I'm not sure about that, this coax is incredible well behaved. I revised the crossover for our SBS.1 speakers today after seeing a somewhat similar pattern on that. I don't think the results are horrible.

Horrible they are not of course but on that price range horrible would be not really an option , still there is a difference though to reference coaxial drivers like the above mentioned.

 

[thewas]

I'd say for instance Kef R5 (admittedly in their lower-end segment) looks way worse:

I'm not sure if anyone has measured the reference series.

Don't agree, the above plot has just some wiggles but is more continuous without larger directivity discontinuities, also to be seen on the more compararable to your bookshelf R3:

 

[thewas]

Kef Reference 3 (0-15-30)...

Nope, that is
Top curve: 45 degrees off-axis response
Middle curve: 60 degrees off-axis response
Bottom curve: 75 degrees off-axis response

SoundStageNetwork.com | SoundStage.com - NRC Measurements: KEF Reference 3 Loudspeakers

www.soundstagenetwork.com

and again much better having then yours, sorry.

 

[sigbergaudio]

@thewas Don't agree about the R5, but the R3 is significantly more tidy.

Unfortunately I can't produce similar graphs right now, but with the revised crossover there's no obvious discontinuity here either:

I guess I have to go back to Seas at some point (will do anyway when the Manta is finalizing). And to point I'm not trying to say there's anything wrong with Kef, they're obviously extremely competent (as is their drivers).

Both Kef and Revel are companies I hold in high regard and who have similar philosophies.

 

[thewas]

Unfortunately I can't produce similar graphs right now, but with the revised crossover there's no obvious discontinuity here either:

There are still but with your new tuning it is quite better, anyway when your tuning is completed we can talk about the full spinorama. Also as said the Seas etc coaxial drivers while not bad cannot compete with KEF or Genelec coaxial drives which is a shame also for us DIY hobbyist.

 

[sigbergaudio]

There are still but with your new tuning it is quite better, anyway when your tuning is completed we can talk about the full spinorama. Also as said the Seas etc coaxial drivers while not bad cannot compete with KEF or Genelec coaxial drives which is a shame also for us DIY hobbyist.

I guess it depends on your definition of problems. There are lots of expensive speakers out there with traditional drivers (so no coax at all) with way worse directivity problems in their crossover regions. And it sounds great (in both speakers).

EDIT: For clarity I'm talking about the coax driver as seen in SBS.1, not the Manta.1 prototype. I haven't even started optimizing the crossover on this particular speaker. And as mentioned there are a number of ports, where I haven't investigated or addressed potential diffraction issues.

 

[thewas]

Sorry, I looked at the wrong graph. I'm glad their Reference series looks better than my prototype.

Also their entry Q or R bookshelves do from directivity point of view (normalised plots)

I was starting to wonder what they were doing over at Kef.

Almost 40 years of optimisation of the coaxial drivers, including since many years full FEM and fluid models enablingh them to fully customise the directivity, just look at their white papers.

I do still stand by the point that I don't think there's a lot of inherent problems with this driver though.

As said it is one of the best OEM but on the other hand also not on the highest level compared to the above mentioned ones.

 

[thewas]

I guess it depends on your definition of problems. There are lots of expensive speakers out there with traditional drivers (so no coax at all) with way worse directivity problems in their crossover regions. And it sounds great (in both speakers).

Everything is relative of course, here we are discussing just them measurements you kindly provided, not many small companies have the courage to do so, you have my highest respect for that.

Now about what sounds great, there is of course also the matter of personal taste, although the Toole/Olive research shows a clear correlation to measurements and also depends a lot on room acoustics, for example acoustically more damped rooms or nearfield listening is more forgiving in directivity flaws.

 

[sigbergaudio]

@thewas Just out of curiosity to understand if we read graphs similarily.
You say the Kef R5 has "no larger directivity discontinuities".

40 & 50 degrees doesn't look to be tracing very well above 3khz.
40, 50, 60 & 70 all have dips in different areas.
80 & 90 have dips from 3-4khz that are 10dB deep.

In general from 5-10khz the different angles don't seem to track each other very well.

If you compare this to SBS.1 (even before revision), there are no dips deeper than 5dB, and they're not even that deep in reality because it's part of an intentional BBC-dip. All angles trace each other almost perfectly following the same dips and peaks with the exception of 15 degrees in the upper end. What am I missing / what parts are you intepreting differently?

This is not argue the point of which is better, I'm curious what you are looking for / what you are considering a problem (or not).

 

[Hugo9000]

Sorry, I looked at the wrong graph. I'm glad their Reference series looks better than my prototype. I was starting to wonder what they were doing over at Kef. I do still stand by the point that I don't think there's a lot of inherent problems with this driver though.

Since you are a manufacturer, you should edit your post so that you don't perpetuate FUD about a product from another brand. It's not enough to acknowledge in a later post that you labeled that graph incorrectly.

Thanks.

 

[sigbergaudio]

Since you are a manufacturer, you should edit your post so that you don't perpetuate FUD about a product from another brand. It's not enough to acknowledge in a later post that you labeled that graph incorrectly.

Thanks.

Of course, done.

And again I have nothing but respect for Kef.

 

[thewas]

This is not argue the point of which is better, I'm curious what you are looking for / what you are considering a problem (or not).

Narrow dips, especially when they are at different frequencies are not such an audible problem as not monotonously continuous directivies like for example on your above plot where the directivity rises, then widens, rises again etc. If you have good quality 0-90° measurements in both axes you can compute both normalised directivity plots, the various directivity indexes (early and full) and the different "bounces" (floor, ceiling, side walls) which show that better.

 

[sigbergaudio]

I deleted the posts where I referred to Kef with exception to the last one (where it was included for the discussion on how to intepret the measurements).

It was my never intention to say anything bad about Kef, but it turned out wrong. My apologies.

 

[Absolute]

Comparing drivers between different speakers are pointless. Any disturbance in the force could be from anything. We need to look at measurements of the drivers alone to compare driver quality isolated imo.

Do you have independent third-party measurements of the drivers @sigbergaudio ?
I think I remember seeing Kef's coax measured somewhere, I'll try to look it up.

Edit; here it was ; http://www.audioexcite.com/?page_id=3614

 

[thewas]

Comparing drivers between different speakers are pointless. Any disturbance in the force could be from anything. We need to look at measurements of the drivers alone to compare driver quality isolated imo.

Sure, but we are discussing full loudspeakers though so crossover tuning and enclosure should be also considered.

Do you have independent third-party measurements of the drivers @sigbergaudio ?
I think I remember seeing Kef's coax measured somewhere, I'll try to look it up.

Edit; here it was ; http://www.audioexcite.com/?page_id=3614

Mind you this is from their previous, now long discontinued R series.

 

[sigbergaudio]

Comparing drivers between different speakers are pointless. Any disturbance in the force could be from anything. We need to look at measurements of the drivers alone to compare driver quality isolated imo.

Do you have independent third-party measurements of the drivers @sigbergaudio ?
I think I remember seeing Kef's coax measured somewhere, I'll try to look it up.

Edit; here it was ; http://www.audioexcite.com/?page_id=3614

Agreed, so that's where the discussion sort of went the wrong direction. My point wasn't to imply problems with Kefs drivers, but more towards your point. That even though issues could be found with the measurements of our (prototype) speakers, that doesn't necessarily mean there's a problem with the driver itself.

I do agree with @thewas that there are some inherent problems with many coax drivers, but there are also some inherent advantages with most coax drivers (such as the obvious, that it is a point source) - so as with everything else in speaker design there are always pros and cons.

With this particular driver we have found there to be many positives and few negatives.

I'm not aware of any third party measurements of the driver itself.