Tannoy Revolution XT 6 Speaker Review

[amirm]

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).

We are trying to figure out different things. I like to find out the limit of the speaker and express it as a single value so we can compare speakers. Most compression is going to happen in bass so we should just pick one frequency and stick with it.

Another reason to do this is the fact that these tests are extremely loud so I don't want to sit through full frequency sweeps at 115 dB or whatever!

I just ran a couple more frequencies:

So yes, it varies. There is 2 dB difference with 100 Hz and 1 dB with 300 Hz. Any of the numbers are better than no numbers.

 

[Tom C]

The new distortion measurements are super cool.

 

[phoenixdogfan]

I think you're really on to something with distortion measurements at various SPLs. Obviously good linear FR is paramount, but if that's only reproducible by a speaker at lower volumes then it's missing a key performance feature needed to satisfy most listeners' need for verisimilitude in sound reproduction.

In other words, most people want speakers that play loud without sounding strainded.

 

[LDKTA]

More data! Yes. That distortion is serious.

 

[MediumRare]

We are trying to figure out different things. I like to find out the limit of the speaker and express it as a single value so we can compare speakers. Most compression is going to happen in bass so we should just pick one frequency and stick with it.

Another reason to do this is the fact that these tests are extremely loud so I don't want to sit through full frequency sweeps at 115 dB or whatever!

I just ran a couple more frequencies:
View attachment 73072

So yes, it varies. There is 2 dB difference with 100 Hz and 1 dB with 300 Hz. Any of the numbers are better than no numbers.

The compression figures, on first glance, appear to reveal bad behavior in other frequencies that appear, i.e., changes in the tonality of the speaker and potentially accentuation of certain frequencies at higher listening levels. To me that reveals far more than one linearity number at any particular frequency.

 

[Canuck57]

She must be stronger than she looks if she is helping lift them.

She would help Arim get something up for sure!

 

[Blumlein 88]

I guess the only complaint is adding all the distortion analysis to a speaker that was going to score so very low even if it has zero distortion. But you have to start somewhere.

There is always the Neumann KH80 DSP measurements #4 with distortion results. HAHAHA!

 

[amirm]

The compression figures, on first glance, appear to reveal bad behavior in other frequencies that appear, i.e., changes in the tonality of the speaker and potentially accentuation of certain frequencies at higher listening levels. To me that reveals far more than one linearity number at any particular frequency.

Those variations are not very audible though:

They are just 0.25 dB above 200 Hz in that test and go down from there. There is an exponential drop off starting at 70 Hz.

 

[MediumRare]

Those variations are not very audible though:

They are just 0.25 dB above 200 Hz in that test and go down from there. There is an exponential drop off starting at 70 Hz.

Agree, for the speaker shown, there's not much happening above 150 Hz. However, below 150 Hz there are some frequencies up and some down by potentially audible amounts. So measuring at 200 Hz - or 100 Hz - or any one Hz - would miss that. I suppose I, at least, have no idea what is out there to find, and at what frequency I would find it, kind of like your temperature-dependency discovery. So why not try this with a few speakers and see what comes out?

 

[SimpleTheater]

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

I’d love to see a pair of Fyne Audio F501’s tested.

 

[oldsysop]

Distortion, what interests me most about a measurement.
Thank you, thank you, 3 times thanks Amir !!!

 

[Xyrium]

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

Last time I checked, the Celestion speakers in my guitar cabs are doing quite well though!

However, I'll admit, just knowing about the dismal distortion performance of this model would really T"annoy" me if I owned them.

Sorry, had to....

 

[Sonny1]

After the last Tannoy, I guess I shouldn’t be surprised but I had high hopes for this one. Great work on the review Amirm, as usual.

 

[617]

The impedance graph of the Tannoy revolution is walking down the street when he sees a sign at a hotel saying 'ugly competition, winner takes home 1000 dollars. Knowing how ugly he is, he goes into the hotel and inquires about the competition. However, he gets turned away; the rules clearly state 'no professionals'.

 

[jurop]

One 'like' button is not enough for this, at least two or three were needed. Finally, serious compression and distortion measurements, which is what IMHO really matters when talking about 'dynamic range' or 'liveness' of a speaker! As soon as I can - very busy period thanks to a strong post-covid rebounce with business - I'll sit and write down a couple of personal considerations. Every review, more and more thanks and kudos to @amirm.

 

[MattHooper]

Amirm, you seem to be gaining the same type of experience as John Atkinson, in terms of measuring speakers and comparing the results to subjective impressions. Many times in JA's remarks, issues in the measurements directly correspond with colorations he heard in the speaker.
But also, fairly often, he reports that though X or Y behavior in the measurements look bad, in practice, subjectively, it didn't nearly sound as bad as expected, or was rarely audible.

This is one reason why I have taken the comments of some in the Stereophile comment sections with a grain of salt, where some people absolutely tear up a speaker design for any number of faults or bad practice. Sometimes I've listened to the speaker in question and quite liked them, or they weren't the sonic disaster one would infer from the enraged comments.

 

[Dennis Murphy]

Perhaps I shouldn't continue hoping that Amir will someday test my Wharfedale D320's.

View attachment 73055



I've joked before about Amir hoisting a 304 lb Westminster Royal up onto his Klippel platform.

View attachment 73059

Been there. Done that. The speaker, I mean. I have a rich friend with a big room. It sounded better than it measured.

 

[YSC]

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

 

[6speed]

@amirm Can you load any wav file to test IMD on the Klippel instead of generating it on the spot? If so, then just create a 32 tone test like what you use on your DACs. I would upload my own results, but I have a Behringer UMC and the ADC generates its own IMD at -60dB, which is higher than the distortion from my speakers.

 

[Ilkless]

Don’t sweat it, this is just the misshapen brother of the magnificent specimen that is the KEF R3.

Getting coaxials right seems to be really hard. I’m kind of curious how the Gradient speakers with the SEAS coaxial would measure.

SEAS coaxials drivers measure better than Tannoy and worse than KEF, Technics and Genelec. About the same level as TAD based on directivity and top-octave roughness. However, KEF has some pretty bad voicing choices that detract from the potential of their drivers, so a good SEAS implementation with smoother directivity and FR below the high treble can be pretty compelling too. The LS50 passive is the worst offender, but the R and Reference stuff are definitely the best short of Genelec.

The Gradient Revolution had flawless polars <10kHz even by today's standards over 20 years ago (old Stereophile off-axis). Since then, Gradient's latest revision uses a custom SEAS coax that pushes any anomalies to 16kHz (basic in-room measurement of the current passive Revolution by a Polish mag, the active Revolution might be an even further improvement). The smaller Gradient 1.4 uses an even newer SEAS coax than the current Revolution (which hasn't seen a further revision yet), and that has completely eliminated any nulls, entering the very top-tier of direct-radiating coaxs. Either way, I have not seen a Gradient speaker with anything less than exemplary polar measurements overall (see also Helsinki 1.5).

Cheaper SEAS coax active monitors exist too, using their off-the-shelf/entry-level OEM drivers with more anomalies, which are nonetheless subtle should FR and directivity below the top-octave be well-managed. The Emes Black has flawless directivity matching at crossover, but the 10kHz anomalies from the entry-level SEAS coaxial (still much better than Tannoy). KS Digital C5-Reference has a bass bump deliberately by choice, but they have a DSP crossover and EQ that pushes the cheap SEAS coax to it's limit: +/-1.5dB throughout much of the audioband.