This is a review and detailed measurements of the JBL Conceal C83 "invisible" speaker. It is on kind loan from a member and costs US $1,100.
As the name indicates, this is basically a flat surface that you cut out in the wall and "tape and mud" the edges. Once painted, it can be made to look just like the rest of the wall and hence "disappear." In the inset, I am showing the stock photo from JBL showing a center woofer and dual higher frequency transducers. All work to excite the front panel.
These type of speakers are high desirable in higher-end stereo, ambient and home theater applications where you want a completely hidden audio system. They require careful project management with the builder/trades to make sure accommodations are made for them during the construction phase of the house. It is highly advised to use appropriate limiter/protections in the amplifier driving them as if you damage them, repair requires ripping them out of the wall to fix! Having heard a number of them, the perceptional effect is quite magical in the way the sound radiates from somewhere in the space with no clues as to where it could be.
These speakers come with a back box meaning the speaker is fully sealed so is not dependent on the wall cavity that you install them on. For easy of testing, I tested the unit as stand-alone speaker instead of building a customer baffle for it. I compared testing this way vs baffle measurements when I tested the JBL Conceal C62 and results were almost the same.
As regular readers know, I use Klippel NFS scanner to measure speakers. The enemy of the system is complexity of the soundfield. The more complex, the more sample points needed. The more sample points, the longer measurements take. With three drivers interfering, the soundfield gets very complicated requiring very high order basis functions to represent it. Anticipating this, I upped the measurement points which requires a 4 hour scan. Even with this, accuracy dropped above 6 or 7 kHz. Klippel NFS makes redundant measurements to compute this. Here is the on-axis actual vs computed one:
We see that the computed measurement in blue (what you normally see in my measurements) starts to separate from actual (red) around 7 kHz. It under represents the energy there by quite a bit when you get above that threshold. To give you an idea of how complex the soundfield is, I captured the 3-D radiation computed by NFS at 10 kHz:
Beautiful acoustic art, isn't it?
Anyway, keep this in mind as you read the measurements below.
NOTE: our company, Madrona Digital is a dealer for these products. Indeed, owner sourced this through us (unsolicited). So feel free to read whatever you want in my commentary.
JBL Conceal C83 Speaker Measurements
Let's start with our usual frequency response measurements:
As with C62, response is extremely variable as the three drivers fight with each other, creating many resonances and cancellations. Letting the reflections sum together acts as a smoothing function:
But there is no escaping the fact that we have anything but a smooth response:
Equalization is a must to a) make the variations smaller and b) give it the proper slope.
The highly variable response makes distortion graphs look bad as well:
Seems a lot of the troughs in the response have peaks in distortion which makes the relative THD graph look, really, really bad:
Note however that power handling was far superior to that of C62. With that speaker, I could not even get to 86 dBSPL and here, I went to 90 dB. Sweeps had some audible issues but not bad.
Directivity measurements are highly variable as well together with some beaming due to use of dual high frequency drivers:
Impedance graph is unlike any other speaker out there (sans the C62):
Waterfall shows resonances we know about:
Step response is not usually informative but here it shows two distinct pulses indicating lack of integration between woofer and high frequency drivers:
I didn't see much point in listening to the speaker.
Conclusions
While building a fully transparent speaker is an achievement in itself, I am left wondering what effort was put into designing the C83. The complex interactions between the three drivers would not lend itself to simple modeling in a speaker design program. Full 3-D FEA would be needed to optimize the design (location of drivers, crossover response, material choices, etc.). The market for these products is too small to justify such level of design but hopefully that will happen one day. Until then, the compromises here are so numerous that I can't recommend the JBL C83 if you have any priority for excellent sound. Would be interesting to see the in-room measurements before and after automated EQ to see if it can be salvaged in use.
The C62 had the same issues but also couldn't handle much power. So if you are going this route, I highly suggest using C83.
------------
As always, questions, comments, recommendations, etc. are welcome.
Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
As the name indicates, this is basically a flat surface that you cut out in the wall and "tape and mud" the edges. Once painted, it can be made to look just like the rest of the wall and hence "disappear." In the inset, I am showing the stock photo from JBL showing a center woofer and dual higher frequency transducers. All work to excite the front panel.
These type of speakers are high desirable in higher-end stereo, ambient and home theater applications where you want a completely hidden audio system. They require careful project management with the builder/trades to make sure accommodations are made for them during the construction phase of the house. It is highly advised to use appropriate limiter/protections in the amplifier driving them as if you damage them, repair requires ripping them out of the wall to fix! Having heard a number of them, the perceptional effect is quite magical in the way the sound radiates from somewhere in the space with no clues as to where it could be.
These speakers come with a back box meaning the speaker is fully sealed so is not dependent on the wall cavity that you install them on. For easy of testing, I tested the unit as stand-alone speaker instead of building a customer baffle for it. I compared testing this way vs baffle measurements when I tested the JBL Conceal C62 and results were almost the same.
As regular readers know, I use Klippel NFS scanner to measure speakers. The enemy of the system is complexity of the soundfield. The more complex, the more sample points needed. The more sample points, the longer measurements take. With three drivers interfering, the soundfield gets very complicated requiring very high order basis functions to represent it. Anticipating this, I upped the measurement points which requires a 4 hour scan. Even with this, accuracy dropped above 6 or 7 kHz. Klippel NFS makes redundant measurements to compute this. Here is the on-axis actual vs computed one:
We see that the computed measurement in blue (what you normally see in my measurements) starts to separate from actual (red) around 7 kHz. It under represents the energy there by quite a bit when you get above that threshold. To give you an idea of how complex the soundfield is, I captured the 3-D radiation computed by NFS at 10 kHz:
Beautiful acoustic art, isn't it?
Anyway, keep this in mind as you read the measurements below.NOTE: our company, Madrona Digital is a dealer for these products. Indeed, owner sourced this through us (unsolicited). So feel free to read whatever you want in my commentary.
JBL Conceal C83 Speaker Measurements
Let's start with our usual frequency response measurements:
As with C62, response is extremely variable as the three drivers fight with each other, creating many resonances and cancellations. Letting the reflections sum together acts as a smoothing function:
But there is no escaping the fact that we have anything but a smooth response:
Equalization is a must to a) make the variations smaller and b) give it the proper slope.
The highly variable response makes distortion graphs look bad as well:
Seems a lot of the troughs in the response have peaks in distortion which makes the relative THD graph look, really, really bad:
Note however that power handling was far superior to that of C62. With that speaker, I could not even get to 86 dBSPL and here, I went to 90 dB. Sweeps had some audible issues but not bad.
Directivity measurements are highly variable as well together with some beaming due to use of dual high frequency drivers:
Impedance graph is unlike any other speaker out there (sans the C62):
Waterfall shows resonances we know about:
Step response is not usually informative but here it shows two distinct pulses indicating lack of integration between woofer and high frequency drivers:
I didn't see much point in listening to the speaker.
Conclusions
While building a fully transparent speaker is an achievement in itself, I am left wondering what effort was put into designing the C83. The complex interactions between the three drivers would not lend itself to simple modeling in a speaker design program. Full 3-D FEA would be needed to optimize the design (location of drivers, crossover response, material choices, etc.). The market for these products is too small to justify such level of design but hopefully that will happen one day. Until then, the compromises here are so numerous that I can't recommend the JBL C83 if you have any priority for excellent sound. Would be interesting to see the in-room measurements before and after automated EQ to see if it can be salvaged in use.
The C62 had the same issues but also couldn't handle much power. So if you are going this route, I highly suggest using C83.
------------
As always, questions, comments, recommendations, etc. are welcome.
Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
