SR-2x Surround Speaker
Performance: Origin of those graphs
The measurements for the graphs shown here were not intended to be published, they are sourced from measurements performed during development. Measurements and analysis for development purposes are done to look into specific issues, and do not need to look nice or be done according to a protocol that can easily be replicated. However, this now provides sufficient documentation for the performance of the SR-2x for its intended use, as-is.
Shape
Shape determines many of the acoustic properties of the speaker. Location and angle of drivers determine radiation pattern, depth and shape of the cabinet affects how back wall boundary changes frequency response.
A surround speaker for wall mount should also look nice when mounted on a wall, not too deep. Fortunately, a shallow cabinet is also best for acoustic properties. A square box on the wall looks bad, and it also has worse performance acoustically.
The designed cabinet is shallow, with angled front panels, and looks very nice. The measurements show performance is excellent, allowing for flexible mounting options, because it works well both on-wall and free-standing.
Concept
SR-2x is based on the S1.2 surround speaker and results from practical testing of this S1.2 in different radiation configurations. The goal is to create sound objects and atmosphere that fills the room with ambience, with sound objects placed around precise enough to resemble real sound sources, from anywhere in the room, and sound should never seem to originate directly from a speaker.
It is also a goal to minimize number of surround speakers. To be able to seamlessly place objects anywhere around the listener, even from above, with no ceiling speakers, without need for more than 4 surrounds. It should work well even with only 2 surrounds, in a 5.1 system.
S1.2 had different configuration options, each with its own characteristic presentation of the sound. Wide is the normal, with a wide pattern that resembles a sound source far away. Bipole is a very wide pattern. Dipole radiates no sound directly on-axis, giving a very diffuse sound. They were all tested, and the conclusion from those tests were that the wide was better than the others, creating both this sense of sound from around the room combined with precision.
Point source, like ordinary small speakers, was also tested, using only one part of the two units on the S1.2. The results from this test was interesting and revealed very important characteristics for surround sound. As point-source, the immersive and realistic surround sound collapsed into small speakers on the walls.
S1.2 Wide:
S1.2 Bipole - Dipole - Point source:
The SR-2x has a pattern similar to the Wide, and this configuration is the correct one for all use in todays multichannel systems. It presents both ambience and space, combined with precision, with no need for excessive number of speakers.
How many surround speakers?
How many SR-2x surround speakers are required to get good surround sound?
2.
In many small rooms, only 2 SR-2x will provide a very good surround sound with nice, room-filling ambience and sound object both towards side and back, even above. This will be a 5.1 system. If the room has some distance from the listening position to the back wall, 2 additional SR-2x can give further improvement. Then you have a 7.1 system.
If the room is large, it is possible to add more SR-2x units, and they can be mounted in the ceiling as Atmos/DTS-X height channels.
Development process
SR-2x was developed purely using simulation and measurements. No "tuning-by-ear", no listening to try to find out if it sounds "right". I ended up listening to them after they were completed, set up as main speakers in Room2, just because it was easy and convenient to do so, and it is always fun to listen to speakers when they are finished, and in this case it was also interesting to hear how they performed, with the rather unique radiation pattern that they have.
Because they are based on the S1.2, the concept was already verified in listening tests, I knew how the radiation pattern should look like, and same drivers, so crossover could be based on the S1.2.
The simulation model for SR-2x is rather simple, just sufficient to show effects of internal cabinet dimensions, damping material and acoustic ports.
Simulation model includes all crossover parts, it is complete in the sense that it is possible to observe every aspect of the design before it is even built.
SR-2x has acoustic ports, kept from the S1.2, and this is a decision that results from simulations. The ports reduce cone excursion at the lowest frequencies, giving increased output capacity and reduced distortion. Simulation is then used to investigate where to place those ports for best response and minimize midrange resonance and leakage.
Here is the port output compared for two versions of the cabinet, the green is a tall cabinet with driver at top and ports close to bottom:
We see the tall cabinet with ports near bottom has a different response, and in this case will not work optimal for intended purpose.
This is how the lf driver cone, port and sum looks like in the simulation:
Blue = port, green = lf driver cone, red = sum.
Nearfield measurements from driver cone and port verifies that the built cabinet performs as intended, and matches the simulation quite well:
Blue = measured driver cone, orange = measured port output.
Measurements are not exactly the same as the signals from the simulation, because output from other sources are present and adds to the measured signals.
As can be seen in those graphs, the SR-2x is no text-book example speaker, something strange happens at the crossover, and the port contribution is unusual. Crossover was changed from the original S1.2 into a simpler version of the high-slope design technology developed for the F205, it gives the speaker a little better off-axis performance. Port tuning is not a standard bass reflex tuning, it is not meant to be, the ports are designed primarily to reduce cone movement and does not increase low frequency output, and the system performs more like a sealed box in the time domain.
Group delay can be examined from simulation, we see GD is low and does not increase excessively at lower frequencies like a standard ported box tuning does:
(ignore deviations around 40Hz caused by numeric errors in transfer of numeric data from the simulator, note that the SR-2x has a usable frequency range from 100Hz and up only.)
The design is modeled in 3D CAD so it is possible to see how it looks before it is built, and details from the cabinet construction can also be implemented in the simulation model, to verify how different design choices affect performance.
After it is built, it can be measured, to verify that it works as intended. If necessary, details in crossover component values can be adjusted, internal damping in the cabinet can be changed. Simulation is used in parallel with measurements in this phase.
