Revel F35 Speaker Review

[Geert]

Klippel has an application note on measuring vented enclosures https://www.klippel.de/fileadmin/kl...ement_with_Multiple_Drivers_ and_Ports(2).pdf. I might be wrong but the example they give on how they obtain the far field transfer function seems to underestimate the contribution of the port:

The paper also describes merging the near field port measurement with the far field transpons as an alternative.

 

[Geert]

Term directivity doesn't apply in that scenario. Eveything flowing through the tube can be considered "directional" but the term directivity applies for free space wave spreading.

I meant the sound is initially directional when it enters free space, and then spreads after a couple of inches. So when you measure in very close distance you observe directivity. But I never saw any reference to studies.

 

[QMuse]

I meant the sound is initially directional when it enters free space, and then spreads after a couple of inches. So when you measure in very close distance you observe directivity. But I never saw any reference to studies.

I know what you meant but I don't see the point of it. What's there to study?

 

[Geert]

I know what you meant but I don't see the point of it. What's there to study?

I think there is a point if you are questioning to what extend a port is directional, and if it can throw off near field or anechoic measurements.

 

[QMuse]

I think there is a point if you are questioning to what extend a port is directional, and if it can throw off near field or anechoic measurements.

Every port can be measured near field, in fact it is how you do it when you do pseudoanechoic in-room measurement where you sum port response with nearfield woofer response. Hoever, the question remains how Klippel handles rear port measurement.

 

[edechamps]

However, a bug was found in the Klippel software that was causing a few spikes before 1 kHz which I attributed to resonances: They were numerical error, not actual response of the speaker. The cause was asking for too much accuracy from the system! It has guards against this but the limits were not enough to find this problem.

It's great that, from the sheer number of measurements and intense scrutiny, you are able to find issues that even Klippel was not aware of. Really feels like advancing the state of the art of speaker measurement

Do you think you'll be able to regenerate spin data once the bug is fixed so that we can compare before/after?

 

[BDWoody]

I would say the chance of Revel/Harman not being aware that a Klippel is being used on ASR to measure Revel and competitors speakers is exactly zero.

Could their sudden interest in 'technical distancing' be related to just trying to protect trade secrets? If they actually do real deal blind tests on every speaker, and have any and all measurements they can figure out how to make for all those speakers, that's a mountain of correlation I don't know that I'd be so thrilled about giving away either.

If that wouldn't be a competitive advantage, I don't know what would be. Helping Amir find a shortcut to whatever they've figured out might not be a sound business decision.

 

[Absolute]

Could their sudden interest in 'technical distancing' be related to just trying to protect trade secrets? If they actually do real deal blind tests on every speaker, and have any and all measurements they can figure out how to make for all those speakers, that's a mountain of correlation I don't know that I'd be so thrilled about giving away either.

If that wouldn't be a competitive advantage, I don't know what would be. Helping Amir find a shortcut to whatever they've figured out might not be a sound business decision.

Could be, but I think we're looking too far into it. There's enough research and knowledge out there to define what makes a great speaker, but there's also enough information and data that suggest that we'll never be able to produce that one speaker that betters everything everywhere for everyone.

Speaker design is and always will be a balancing act between different contradicting priorities and will therefore never be perfect. I don't believe for a second that Harman/Revel knows something that no one else knows, I just think that they're sticking to what they know and prioritize and try to perfect their particular design compromises. If I'm right, it would be weird of them to come out and confirm that "hey, we just do what we've done for 20 years with focus on minute details".

I don't think - and will never accept - that making spectacular speakers is somehow a mysterious black art. Even Revel produces straight up normal speakers like everyone else has been doing for decades.

 

[BDWoody]

I don't think - and will never accept - that making spectacular speakers is somehow a mysterious black art.

I agree with what you say completely.

I may disagree on whether the accumulated information and intellectual property, and all of that supporting data doesn't show correlations that a smaller data set wouldn't, and that's a good reason to keep it protected from a business point of view.

Whatever compromises they decide to make on any of their hundreds of speakers, presumably based on that data in terms of preference correlation, for them would be just putting an ever sharper tip on the pencil in a constantly iterative process, not reinventing any wheels or claiming magic woo.

 

[Absolute]

I think the compromises is largely a result of design target vs pricepoint. The size and quality of the cabinet must be balanced against the driver choice, design, spl requirements, cost of manufacturing, price point aim, shipping etc.

They don't seem to compromise much to the importance of frequency/polar response, which is the biggest giveaway of what they deem most important.

I don't believe there's much effort being made in disovering if there's inherent advantages to very small things like 0,05 % less distortion at 1 khz vs a 0,2 dB worse response at 1 khz or if a small reduction in diffraction is more important than a little bit reduced dispersion.
Business companies will always seek to do things as cheaply and efficiently as possible where the term "good enough" is far more important than "best that can be done".

Since I'm a cynical bastard I'm willing to bet a corona that any speaker designer is judged after their capability to produce good results with regards to the design goal as cheaply and efficiently as possible in the eyes of their employers, Harman included.

I accept that I might be wrong, but I don't believe the claim that Revel is meticulously blind testing all their products against all the top rivals before they release a new product. Business wise that doesn't make any sense, especially if there's any truth to the notion that they know something that others don't.

I'm rambling because I'm bored, sorry for that. I'll stop. Promise.

 

[BDWoody]

I accept that I might be wrong, but I don't believe the claim that Revel is meticulously blind testing all their products

I find it very hard to believe as well, but I suppose I am letting myself go with those who would be in a better position to 'know,' rather than my general feeling of 'yeah...right.'

IF they actually do them, AND have, on EVERY speaker... whew...I'm exhausting myself...

I'm rambling because I'm bored, sorry for that. I'll stop. Promise.

Right there with ya...

 

[TimVG]

According to Rex Anderson in revel thread on AVS Mark Glazer told that "there is a lot more to loudspeaker performance than just the spin. A less expensive pair of loudspeakers could have a similar spin, but could not produce the dynamics and clarity of the F328 or F226" and there's upcoming White Paper on the F328Be

I like to think a great spinorama is more likely than not the result of proper design in the first place (acoustically, mechanically) and not brute forcing filtering a mediocre design to conform to a certain standard. Also big shout out to Mark Glazer who was the main design engineer on all the Revel models to date, yet remains mostly in the background.

 

[QMuse]

I like to think a great spinorama is more likely than not the result of proper design in the first place (acoustically, mechanically) and not brute forcing filtering a mediocre design to conform to a certain standard. Also big shout out to Mark Glazer who was the main design engineer on all the Revel models to date, yet remains mostly in the background.

That indeed sounds reasonable, but what about cases like Klipsch RP-600M which has excellnt DI's and only average LW, ER and SP? If you assume that inter-specimen variation is small you can easilly EQ it and get stellar spinorama.

 

[TimVG]

That indeed sounds reasonable, but what about cases like Klipsch RP-600M which has excellnt DI's and only average LW, ER and SP? If you assume that inter-specimen variation is small you can easilly EQ it and get stellar spinorama.

That's true of course. What I meant is I have more respect for speakers that have a good spinorama by default, I like to believe care went into designing it at that point. I think that is what Mark Glazer implies as well, that a good spinorama is one thing, but that it's the result of making sure everything that came before it was considered to get to that point, and that not every one of those considered details, ends up being very obvious in the spinorama itself.

 

[Balle Clorin]

Can someone please explain why the F35 has the very uneven vertical directivity? Is in a failure in design, or just inherent in a 2.5 way with 3 woofers?

 

[tuga]

Can someone please explain why the F35 has the very uneven vertical directivity? Is in a failure in design, or just inherent in a 2.5 way with 3 woofers?

A coaxial driver (tweeter concentric to the mid-woofer) will produce identical vertical and horizontal dispersion within the two drivers' combined operating range.
When you disseminate several different "ways" over a long baffle the vertical dispersion varies depending on the listening axis position in regards to the front baffle because each driver will produce increasing beaming with frequency.

 

[lonewolf]

FYI I have been going through detailed analysis with Klippel. So far, no issues are found with low frequency response but work continues.

However, a bug was found in the Klippel software that was causing a few spikes before 1 kHz which I attributed to resonances:

They were numerical error, not actual response of the speaker. The cause was asking for too much accuracy from the system! It has guards against this but the limits were not enough to find this problem.

Will there be updated measurements of this speaker? I imagine it's a lot of work so I can understand if there isn't, but it would be nice to see accurate measurements that don't show those odd spikes.

 

[lonewolf]

Good suggestion. Just looked and the two ports are open. But looking inside, the padding on the right side is not glued and had come loose and moved an inch toward the bottom port. The top port is more open however, there is padding horizontally separating the top of the box and its port, and the bottom. The woofer is in front of the top port so I am assuming just the padding is there rather than a hard barrier. Any ideas if either one of these impacts bass response?

I had a look at both of my F35s and all the padding is glued firmly to the sides. Looking up through the bottom port I can see wood bracing between the top and bottom ports with padding covering the top of the brace. Wood is only on the wall of the speaker and largely open through the middle. If the padding came loose on yours and moved forward by an inch, I would assume it's not blocking any airflow through the port?

 

[BYRTT]

Will there be updated measurements of this speaker? I imagine it's a lot of work so I can understand if there isn't, but it would be nice to see accurate measurements that don't show those odd spikes.

In below analyze F35 verse M16 / R3 / 8341 think its hard to see what is the problem before 1kHz area and that a software bug should help out for F35, compared to beauty curves of M16 one could hope for there is some padding or wire fault that could help F35 improve in directivity index say 500Hz-5kHz and in its not a loaner speaker maybe it will be looked into down the road : )

LOL...

For those four ASR sweeps had a closer look into their directivity index curves 200Hz-10kHz, there is many coherences here and there and could point out 8 times where all four models is on same frequency seen in below animation, maybe @amirm is interested see that data :

 

[lonewolf]

In below analyze F35 verse M16 / R3 / 8341 think its hard to see what is the problem before 1kHz area and that a software bug should help out for F35, compared to beauty curves of M16 one could hope for there is some padding or wire fault that could help F35 improve in directivity index say 500Hz-5kHz and in its not a loaner speaker maybe it will be looked into down the road : )

Interesting graphs, thanks! I'm really curious if those graphs are the true performance of the speaker or if something else was throwing off the results for that speaker. Seems odd that the M16 measured better bass performance, and I would expect (well, hope) the directivity index to be better as well.