Devialet Phantom Reactor 900 Measurements (now with Spinorama)

[napilopez]

Nice work!

What gating do you use for 200Hz upwards and what is the distance from mic to the speaker? Do you put speaker in centre of the room on some stand or..?

I gate measurements at 6.5ms and measure from 1m in the center of my room on an elevated stand, which is placed on a turntable. The room is narrow but has 16ft ceilings, so usually i have a little over 7ms to work with, but i keep it at 6.5 to be safe.

1m isn't ideal for driver integration, but as I only intend to measure bookshelf speakers and the reactors are particularly small, this isn't a problem. Just to be sure, I always repeat a few on and off axis measurements at 2m to make sure there isn't a major difference, and so far no speaker I've measured has shown any meaningful variation at 2m, though obviously I have less resolution to work with at low frequencies at that distance.

For this particular speaker, I could not just turn it on its side to perform vertical measurements because that would impede one of the woofers from moving. So I had to do them more manually by repositioning the speakers and my microphone stand for each measurement using an angle finder, ruler and and a laser. This surely introduced some user error, hence the messiness of my vertical measurements a few posts above. However I do not think it matters tremendously once averaged, especially considering the speaker's even dispersion.

 

[Krunok]

Just to be sure, I always repeat a few on and off axis measurements at 2m to make sure there isn't a major difference, and so far no speaker I've measured has shown any meaningful variation at 2m, though obviously I have less resolution to work with at low frequencies at that distance.

This is not quite obvious to me, I'm probably missing something.
Do you use the same Right window of 6.5ms when measuring at 2m? Why would distance affect LF resolution?

 

[napilopez]

This is not quite obvious to me, I'm probably missing something.
Do you use the same Right window of 6.5ms when measuring at 2m? Why would distance affect LF resolution?

Oh, because I have to reduce the gate to about 3-4ms I think when I move to 2m So not quite the same =] Still decent resolution for the mids on up, but it gets trickier to splice the bass

 

[Krunok]

Oh, because I have to reduce the gate to about 3-4ms I think when I move to 2m So not quite the same =] Still decent resolution for the mids on up, but it gets trickier to splice the bass

Huh.. Why would you want to reduce the gating while you were increasing the distance between mic and the speaker? With speed of 330m/s it takes 3ms for 1m so you would need app 6ms for the direct wave to reach mic which is 2m away. In that scenario you would probably need to set right window to 7-8 ms to block early reflection waves from the floor. Or am I thinking something very wrong here..?

 

[napilopez]

Huh.. Why would you want to reduce the gating while you were increasing the distance between mic and the speaker? With speed of 330m/s it takes 3ms for 1m so you would need app 6ms for the direct wave to reach mic which is 2m away. In that scenario you would probably need to set right window to 7-8 ms to block early reflection waves from the floor. Or am I thinking something very wrong here..?

Obviously I don't want to reduce the gate, but I measure indoors. The further I am from the speaker, the smaller the difference in time between when the sound hits the mic and the first reflection hits.

Also to be clear, the gate I am referring to is the relative gate. So 0 ms is when the sound hits the mic, not the start of the signal.

 

[Krunok]

Obviously I don't want to reduce the gate, but I measure indoors. The further I am from the speaker, the smaller the difference in time between when the sound hits the mic and the first reflection hits.

Also to be clear, the gate I am referring to is the relative gate. So 0 ms is when the sound hits the mic, not the start of the signal.

Aha, so you're using relative gate. In that case I suggest you lower the 6.5ms gate when measuring from 1m as 6.5ms after direct sound reaches the mic is enough time for the reflections from the floor and ceiling to arrive.

Btw, which software are you using for measurements?

 

[napilopez]

Aha, so you're using relative gate. In that case I suggest you lower the 6.5ms gate when measuring from 1m as 6.5ms after direct sound reaches the mic is enough time for the reflections from the floor and ceiling to arrive.

Btw, which software are you using for measurements?

I use REW for measuring, Vituixcad for generating the spins =].

The speaker is elevated over 5 feet from the ground and my ceilings are 16 feet high, so it is always at a nice distance from the nearest surface. The reactor and mic were about 167 cm from the ground, which gives us a reflection time of about 7.2 ms:

In any case, there's no need to guess or do the math; you can, of course, see when the first reflection arrives in the impulse response:

6.5 gives me a nice buffer, and I usually have a bit more, but the reactor is a very short speaker.

 

[Krunok]

I use REW for measuring, Vituixcad for generating the spins =].

The speaker is elevated over 5 feet from the ground and my ceilings are 16 feet high, so it is always at a nice distance from the nearest surface. The reactor and mic were about 167 cm from the ground, which gives us a reflection time of about 7.2 ms:

View attachment 45989

In any case, there's no need to guess or do the math; you can, of course, see when the first reflection arrives in the impulse response:

View attachment 45990

6.5 gives me a nice buffer, and I usually have a bit more, but the reactor is a very short speaker.

Sure, reflection arrivals are visible from the IR.

Here is the link for that calculator, if anybody want to use it.

 

[extenso]

I was quite impressed auditioning the Reactor. The only thing that was mildly annoying was the elevation between approx. 20 and 100 Hz. How could that most easily be taken care of with EQ? I am part of the Apple walled garden, by the way.

 

[napilopez]

I was quite impressed auditioning the Reactor. The only thing that was mildly annoying was the elevation between approx. 20 and 100 Hz. How could that most easily be taken care of with EQ? I am part of the Apple walled garden, by the way.

Indeed, it's not really the type the of problem you normally have with small speakers, and devialet could fix it so easily by simply having EQ in the app. But as you can see in my in-room response, if you have the speakers near the walls (in my case, 1-2 feet from an uneven front wall) you will simply have too much bass.

How would you connect to the reactor?

 

[extenso]

Indeed, it's not really the type the of problem you normally have with small speakers, and devialet could fix it so easily by simply having EQ in the app. But as you can see in my in-room response, if you have the speakers near the walls (in my case, 1-2 feet from an uneven front wall) you will simply have too much bass.

How would you connect to the reactor?

They would be connected to a TV with Toslink, and my navigation hub would be the Apple TV to the TV with HDMI. Regrettably, my room layout will force the Reactors too close to the front wall, amplifying the troublesome frequencies.

 

[napilopez]

They would be connected to a TV with Toslink, and my navigation hub would be the Apple TV to the TV with HDMI. Regrettably, my room layout will force the Reactors too close to the front wall, amplifying the troublesome frequencies.

That's totally fine if you plan on EQ-ing anyway. I'm a firm believer that closer to the wall is better as long as you plan to EQ if you don't have a room the size of an auditorium. See this genelec vid.

Does your TV have any EQ in it? Since the problems are really only in the bass, you can probably just turn down the bass with a basic EQ tone contro until you feel it sounds good.

 

[extenso]

There are some rudimentary EQ options, but won’t they just affect the internal TV speakers?

 

[napilopez]

Depends on the TV! If not you can get some device to act as an intermediary and provide EQ. Devialet plans to add in-app EQ at some point but they don't have a timeline for it because apparently it's hard to do with their DSP for some reason.

 

[extenso]

Depends on the TV! If not you can get some device to act as an intermediary and provide EQ. Devialet plans to add in-app EQ at some point but they don't have a timeline for it because apparently it's hard to do with their DSP for some reason.

Any ideas for such a device?

 

[Frank Dernie]

Any ideas for such a device?

PMFJI

I have not tried it (though I do own one) on my Phantom silvers but the DSPeaker Anti-mode dual core 2.0 can be routed through a toslink connection. One connects the source to the Anti-mode then the output from the anti-mode to the phantoms.
I haven't tried it but it should work.
My speakers are in a big room away from walls and I haven't even tried to equalise trem.

 

[extenso]

PMFJI

I have not tried it (though I do own one) on my Phantom silvers but the DSPeaker Anti-mode dual core 2.0 can be routed through a toslink connection. One connects the source to the Anti-mode then the output from the anti-mode to the phantoms.
I haven't tried it but it should work.
My speakers are in a big room away from walls and I haven't even tried to equalise trem.

Interesting that you should mention DSPeaker. I have the old Anti-Mode 8033 gathering dust after using it in a prior setup. It did its job well, but I guess it should be used only for subwoofers. Being old, it also has only RCA connections.

 

[napilopez]

Just thought I'd update this thread with some things.

Mentioned this in another post, but it's worth noting the Reactors measure significantly worse(a big dip around 3k) if not positioned properly on whatever stand of surface you put them on:

Also thought I'd share polars since some people like them. You can see the speaker approaches omnidirectionality for much of its frequency range, and have extremely wide dispersion up to about 4K, where the HF unit becomes more directional (it's a large unit, about 2 inches). Note that the area from 100-170 degrees is simulated, but it should be a decent representation of what's going on.

Horizontal:

Vertical (messiness in part due to having to reposition the speaker for this data):

For comparison, this is what a more traditional good dispersion looks like (iLoud Micro Monitor):


Some polar charts:

500Hz:


2KHz:

5KHz:

10KHz:

 

[Sancus]

These speakers are really interesting, but setting them up for HT or surround with any other types of speakers seems like a total nightmare due to the input requirements and reported latency.

 

[napilopez]

These speakers are really interesting, but setting them up for HT or surround with any other types of speakers seems like a total nightmare due to the input requirements and reported latency.

Indeed. The reactors have an optical/3.5mm input though, so it's not quite as horrid as the original Phantoms. It's clumsy, but you could just set the volume the same on each unit and run mono RCA to 3.5mm cables. There's also a custom version with a euroblock connector for integrators. But yeah, not ideal.