Tannoy Revolution XT 6 Speaker Review

[amirm]

@amirm Can you load any wav file to test IMD on the Klippel instead of generating it on the spot?

No. There is no way to change the waveform other than setting its start and stop frequencies.

My AP does that with ease of course so I will measure and post it using that.

 

[EchoChamber]

I had a pair of Tannoy 8” studio monitors many years ago. I tweaked them to death but I was never able to like them... Everything was off, tonal balance, lack of detail... the only think that worked more or less was soundstage, but nothing special. I don’t know how they got away selling it to the pro community. I think their driver technology is fundamentally flawed...

 

[amirm]

Since I had the Audio Precision powered up and the Tannoy was still in measurement position, I ran the 32-tone test:

Room noise and variable frequency response makes it harder to interpret the results. It is also late at night and I can't run it too loud. Will test against some other speakers and refine.

 

[Dzhaughn]

I'm British and am sorry about our speakers. One day we'll get one right

Perhaps you could find a way to make them leak oil?

 

[filo97s]

I Have the XT8F, so not exactly the same speaker, but I must admit that from listening tests, I really liked how they sound. For me, better than most speakers in this price range.

 

[tuga]

Our reference Revel M106 clearly has much lower distortion

These comparisons are very useful, thanks!

No wonder then that the XT 6 easily exceeds my criteria of 50 dB while playing at 96 dB SPL:

Shouldn't the threshold be a curve (akin to a loudness contour)?

 

[Tangband]

It seems to be a very bad idea to use one coaxial driver covering the whole spectra. The bass will modulate the tweeter acting as a big variable waveguide. The sound is often even worse in those speakers than what the measurement shows.

 

[Tangband]

Do are there designs that make great coaxial drivers? I can only think of Genelec and devialet phantom..

If you only use the coaxdriver down to about 400 Hz with 18 dB or 24 dB/oct filtering such in the more expensive Genelecs, you can get away with it.

 

[Vuki]

Not surprised these are a mess, I'd love to know how their massive old world ones measure.

Like this

 

[edechamps]

Since I had the Audio Precision powered up and the Tannoy was still in measurement position, I ran the 32-tone test:

View attachment 73082

Room noise and variable frequency response makes it harder to interpret the results. It is also late at night and I can't run it too loud. Will test against some other speakers and refine.

I did a quick re-watch of the relevant Klippel webinar that happened a month ago. Presentation of multi-tone testing starts at 00:54:13. He does a live demo of a measurement at 01:08:42. His measurement doesn't seem to have any SNR issues (as shown at 01:11:10), despite being done in a "normal" acoustical environment (Wolfgang's office). That said, he was measuring a very small speaker that presumably produces way more MTD than the one being tested here.

I noticed a couple differences in measurement methodology:

• His measurement stimulus uses way more tones than yours. About 128 tones if I counted correctly.
• He used in-situ compensation to avoid issues with room reflections. If I understand correctly, "in-situ compensation" means that he used the true frequency response of the speaker (from the NFS), compared it to the frequency response picked up by the microphone in the distortion measurement setup, generated a correction curve from the difference, and applied it to the distortion measurements.

It seems to be a very bad idea to use one coaxial driver covering the whole spectra. The bass will modulate the tweeter acting as a big variable waveguide. The sound is often even worse in those speakers than what the measurement shows.

This is why I suspect multi-tone measurements such as IMD and MTD are especially important with coaxials. The issue you're describing will never be triggered by a single tone measurement.

 

[Tangband]

That's a very good point. Amir stated his linearity test was at 200Hz. But, we know compression changes vs frequency. Here's an example of a test result I ran and I didn't even push it past 100dB @ 1m (mean SPL; the HF response was +6dB above the mean; POS speaker).

From here:
https://www.erinsaudiocorner.com/loudspeakers/jamo_s807/

Note: I call this compression testing. I ain't changing that terminology either.


I would imagine AP has the means to do this. Obviously it can be done manually running multiple voltage-varied sweeps. But that would get old.

Very interesting !

 

[GXAlan]

Since I had the Audio Precision powered up and the Tannoy was still in measurement position, I ran the 32-tone test:

View attachment 73082

Room noise and variable frequency response makes it harder to interpret the results. It is also late at night and I can't run it too loud. Will test against some other speakers and refine.

have you tried any of the IMD music tests?
https://tidal.com/browse/playlist/fefd092e-4912-447e-9914-2a3df3f78613

 

[pozz]

For the audibility side of the distortion curve, it would be fun to start a group effort on modelling curves for simultaneous masking at several SPL points, like the standard 86dB and 96dB Amir uses. Since the curves would be frequency-specific, the overlay on the distortion data would look like a filterbank. IIRC available masking research is limited from 100Hz to 10kHz.

 

[JustIntonation]

I spent a few days developing more tests in Audio Precision. Work is on-going. Setup is indoor where I actually take pictures of the speakers. Acoustics of the space can impact the measurements naturally but I have worked to minimize them. Feedback is welcome from happy members on this. Grumpy ones please find something else to do.

Great to see more distortion measurements!

My happy member suggestion / feedback: Multi-tone IMD measurements. Like the 32-tone multitone you use for DAC and AMP measurements.
Perhaps generated on a pink noise weighting and starting at 30 or 40Hz? Otherwise the regular multi-tone measurement is still sure to be very informative.
This would be a single measurement (though perhaps at different levels / SPL) no need to base it on crossover points for a driver or differences between 3-way and 2-way speakers etc.
I think this could well be the most informative distortion measurement for speakers. Perhaps the resulting distortion products displayed with the equal loudness curve overlay like you did with one of the HD graphs already would be very informative as well.

 

[Billy Budapest]

The master of audio hustle on this speaker

Eh, I wouldn’t call him the master of audio hustle—just a salesman doing his thing. I do think that Upscale Audio does some things very well—for example, if tubes are your thing, their testing for reliability and tube matching is excellent.

 

[hardisj]

Eh, I wouldn’t call him the master of audio hustle—just a salesman doing his thing.

Oh. So no different than the vast majority of "reviewers" on YouTube and the web. Got it.

 

[hardisj]

Very interesting !

I've got some other measurements like this of other speakers. It's the "standard" suite of measurements I include. Easy to read and understand (especially since I do the math and tell you the theoretical mean SPL for each output voltage in the legend). Here's some other examples:

https://www.audiosciencereview.com/forum/index.php?threads/neumi-bs5-bookshelf-speaker-review.14404/

https://www.audiosciencereview.com/forum/index.php?threads/buchardt-audio-s400-my-review.14079/

 

[edechamps]

His measurement doesn't seem to have any SNR issues (as shown at 01:11:10), despite being done in a "normal" acoustical environment (Wolfgang's office).

@amirm Forgot to mention that at some point in the webinars he mentions that distortion measurements typically have to be done in the near field precisely to improve SNR.

 

[hardisj]

Maybe Thomas or Amir should break off this distortion testing discussion in to a separate thread. I feel like all the content here has been aimed at Amir's new data and attempts at best providing it and anyone coming here for discussion about the performance of this speaker is going to be lost trying to find it. Just a thought.


And with that said, some things about the data that interest me...

• HF break-up (the peaks/dips above 8kHz): Surely some issue with the big ol' rubber surround. Additionally, they don't market it as a 'full-range' speaker but I imagine most who buy it will run it that way or cross it in the 80Hz region; below point where excursion is limited significantly and thus the modulation effect of the waveguide (midrange) is prominent on the HF response. For example of what I mean here, if you go to my test of the Kef Q100's tweeter response with the woofer fixed out, at rest, and fixed in (bottom of the page in this link) you'll see how the HF response varies based on the position of the midwoofer (which is the waveguide in this case). Bottom line: this HF response is already nasty and going to be worse than what this single set of data shows when playing music unless crossed over high enough to limit excursion.
• "Highly efficient 89 dB/Watt sensitivity". If you count >10kHz only.
• Lots of little resonances/anomalies seen. Kind of reminds me of the little Neumi BS5 I tested where I commented that, while the response was full of resonances throughout, I didn't feel like they were high enough in level to be terribly annoying but maybe it was enough to keep it from subjectively sounding a notch better. That would be something that I wish was easier to diagnose.
• 6-inch drive unit used. That should put the beaming range at about 2.25kHz (give or take). Manual states crossover of 1.8kHz per Stereophile's review:
  • 

• Matching the above with Amir's HD testing, I can see the case for the increased 500Hz distortion being caused by the port. But I don't really see where the sharp increase in ~1.8kHz-based HD is borne. But that number happens to be the crossover point and I can't help but wonder if we are seeing distortion from the tweeter. Then, you look at the NF plot and see the tweeter's crossover slope is quite steep. So, I'd be inclined to maybe put the blame more on the mid/woofer; maybe surround-edge resonance (which is typical when the surround/cone compliance is different).
• Random gripe but when I google "Tannoy Revolution XT 6" I don't get Tannoy's site until 6 or 7th hit down.

 

[hardisj]

Something else that would be nice to see with designs like this (coaxial/concnetric) is the step response. That way we can tell just how well aligned the drive units are. Especially when companies tout the concentric design in the marketing like Tannoy does with this speaker:
"Dual Concentric driver technology provides class-leading coherence and point-source imaging"

I tested this with Kef's R300/500 concentric and found it to be true. I'm always skeptical of other brands, though. Especially the "coaxial" ones who seem to hope people don't understand the difference between coaxial and concentric.