PMC Twenty.21 Bookshelf Speaker Review

[thefsb]

I see Floyd responded to your Yamaha post here. He also posted about them on AVS.

That gave me a trip down memory lane.

The only time I attended a Hi Fi show/exhibition was around 1983 in Edinburgh. The rep in the Yamaha room was very nice and accommodating. After a moment of very polite hesitation he took my copy of Gateway (John Abercrombie, Dave Holland, Jack DeJohnette) and we listened to the whole of Back-Woods Song. I don't recall anything about the equipment except that the speakers were NS1000.

That was a transformational moment and set me up for decades of disappointment in listening to playback of recorded music.

 

[carlosmante]

I bet the Kudos stuff measures real bad, I've hated every pair I've ever heard, same with Neat.

Maybe an acute case of Misophonia.

 

[amirm]

have any of the Salk speakers been measured here?

Not yet but one just arrived.

 

[MZKM]

Not yet but one just arrived.

Really looking forward to that. They usually have an on-axis frequency response, but usually it’s just fine (sometimes great), and they always state they focus on being phase-coherent, so I hope the off-axis results are good.

 

[Pepperjack]

Not yet but one just arrived.

Oh snap. Pretty interested to see what happens with this one. I suppose if it aligns with the data they provide it will be a real good sign for the rest of the family.

 

[MattHooper]

The fact.8 was originally released in 2010 but we reviewed new version, the fact.8 Signature: https://www.stereophile.com/content/pmc-fact8-signature-loudspeaker

As I mentioned in the comments section, your review and the measurements didn't surprise me at all. (Nor did the measurements Amir took of the 21 bookshelf speaker). Having spent time listening to other PMC speakers including the fact.8, they left me scratching my head: "Wait, professional monitor company? Are these supposed to be neutral because they sure ain't!"

 

[mhardy6647]

And also here if you're willing to do lots of scroll https://www.audiosciencereview.com/forum/index.php?reviews/

Scrolling through that list was my gateway drug here. I am just sayin'...

 

[MattHooper]

When I heard the PMCs I thought "well if pro monitor guys use these I guess it's sort of like the old YAMAHA NS 10s." Not exactly the same frequency response, but a heightening of "detail" for hearing in to the mix....or something."

Kal's subjective review mirrored what I heard. Though I may be more harsh: vocals sounded clear, but artificial. The warmth region was down so the "edge" of the sound was emphasized while richness and body was removed, so no vocals sounded natural or human to me.

 

[wwenze]

I'm getting confused. So are we supposed to listen to it on-axis so we can enjoy the treble as intended, or off-axis so we can enjoy the overall sound as intended?

 

[Xyrium]

I'm getting confused. So are we supposed to listen to it on-axis so we can enjoy the treble as intended, or off-axis so we can enjoy the overall sound as intended?

In the case of the PMC 2021, you're not supposed to listen to them at all, reach for your headphones instead.

 

[Ilkless]

PMC Twenty5 just went on sale at the local distributor. What a coincidence

 

[BikeSmith60]

PMC Twenty5 just went on sale at the local distributor. What a coincidence

LOL. Yeah, that coincidence is valid for just about everything when new models are replacing older models. The Twenty5 models are replaced by Twenty5i models since late May 2020. Gee, what a coincidence. SMH...

 

[Frank Dernie]

In the case of the PMC 2021, you're not supposed to listen to them at all, reach for your headphones instead.

Sorry but even these are not bad enough for me to prefer listening on headphones, I have nice ones but emergency use only just because I hate the experience so much.

 

[Koeitje]

How does this spectral plot compare to the M106? There I see more delay at the lower frequencies, but the PMC has a peak at 400hz. Everything higher looks lower on average for the PMC, but less smooth than the Revel. Can somebody explain how I should interpret this?

 

[Listen first]

PMC recommend that you listen to twenty series loudspeakers off axis. This is from the twenty 21 manual:
To further enhance the audio picture or soundstage the speakers can be angled/toed-in’. Start with the speakers angled so they will cross approximately 50cm (2ft) behind the listening position. (See stereo set-up diagram) Varying this angle will also subtly affect the vividness of the audio picture, so again experiment.
Over the years (and this is an old design) they have further developed this concept. This is from the more recent twenty 5 21i manual:
To enhance the audio picture/soundstage and sound balance, the speakers can be angled/toed-in. With the speakers facing straight ahead play a well recorded vocal track. Gradually toe-in both speakers by small increments (max 5° at a time) (see stereo set-up diagram) until the voice is vivid and centrally placed. You can ensure this is best position by going ‘one step too far’ then returning. Then, use a track with extended high frequencies such as cymbals, check that they sound clear but not overpowering. If they sound too loud or harsh subtly toe the speakers out by very small increments until you achieve the ideal balance; then repeat both listening tests to confirm.
So they clearly expect listeners to listen off axis and to use the associated HF roll off to achieve a "balanced" response. Had this of been considered the frequency response measurements would be, flatter and more useful.
As it stands the on axis frequency response measurements are not valid.
Might be an idea to get a better understanding of the intended use of the system before posting findings.
I'm sure the lovely folk at PMC would be only too happy to engage to prevent scenarios like this.
My experience of PMC is that they are a very good company with sound engineering pedigree who make great products.
There are dozens of examples of very good sounding speakers which don't measure ruler flat.
Equally there are dozens of very bad sounding speakers which do measure ruler flat.
If there was a "correct" target frequency response everybody would be using it.
There isn't because there are too many variables involved. Everybody hears differently and the human hearing sensitivity changes hugely across the frequency bands. A flat speaker frequency response to a flat microphone doesn't mean a flat response to the ear brain system, when you consider how hearing sensitivity changes with frequency.
To judge a loudspeakers performance on how close it comes to a flat line is not taking into account the full picture.
And that is coming from someone who measures speakers for a living.
Of course one can use DSP alongside loudspeaker drive units, cabinet enclosures and crossover components to achieve a flat response however many designers don't have DSP available and have to as much as they can in the box.
Even if you do get a flat response out of the box your room will EQ the system as well as your ear brain system.
So you still have to listen and measurements can only be used to guide you in the right direction.

 

[tonapo]

I have a pair of PMC Twenty5 22's and this result may explain some of my experience. A little history first...

My main listening room was our lounge, a family room, that has to accomadate a TV and 5.1 system, I have a Monitor Audio Bronze set-up in there. I got back into the kit side of things as the family got older a few years ago as it was less likely that a finger would push through a tweeter! I use Roon and HQ Player with ethernet across the house. Anyway, because of the room lay-out, and the front left and right speakers, I felt I was not getting the most out of the lounge set-up, so I went and listened to a few speakers (including B&W 704 s2’s, Monitor Audio Gold 200 and Kef R5’s) thinking better speakers would help, I ended up bringing the B&W's home for an extended demo. I liked how they sounded (to a point), but I was getting uneven bass due to rear porting and the fact that one speaker fired into space (to the left of a fireplace) and another into plaster board (it was in a window bay). And my other half was not taken by the looks. So, I started looking at sealed cabinets and front ported speakers and I ended up on the PMC Twenty5 22's, I demo'd at home and they were better in my lounge than the B&W's, so I bought a pair as they were reduced and I traded some kit in too. At the same time I was starting to use REW and then I ended up getting some custom filters from Homeaudiofidelity which I load into HQ Player. I still struggled though with uneven stereo imaging which I put down to positioning and the room layout. So much so, that I moved my whole main listening kit into our snug room, where I can get a better stereo triangle. I measured again, and loaded up some new filters. It does sound alot better in this room, although this review is another interesting angle to it. And, to be clear, I do like how they look, although the 22's could be a little big for the current room. I do listen slightly off axis, with the speakers crossing over behind the listening position.

Anyway, I can dig out the room measurements with the speakers, before and after, if people are interested.

 

[Purité Audio]

I expect they sound pretty decent from the house next door.
Keith

 

[Koeitje]

PMC recommend that you listen to twenty series loudspeakers off axis. This is from the twenty 21 manual:
To further enhance the audio picture or soundstage the speakers can be angled/toed-in’. Start with the speakers angled so they will cross approximately 50cm (2ft) behind the listening position. (See stereo set-up diagram) Varying this angle will also subtly affect the vividness of the audio picture, so again experiment.
Over the years (and this is an old design) they have further developed this concept. This is from the more recent twenty 5 21i manual:
To enhance the audio picture/soundstage and sound balance, the speakers can be angled/toed-in. With the speakers facing straight ahead play a well recorded vocal track. Gradually toe-in both speakers by small increments (max 5° at a time) (see stereo set-up diagram) until the voice is vivid and centrally placed. You can ensure this is best position by going ‘one step too far’ then returning. Then, use a track with extended high frequencies such as cymbals, check that they sound clear but not overpowering. If they sound too loud or harsh subtly toe the speakers out by very small increments until you achieve the ideal balance; then repeat both listening tests to confirm.
So they clearly expect listeners to listen off axis and to use the associated HF roll off to achieve a "balanced" response. Had this of been considered the frequency response measurements would be, flatter and more useful.
As it stands the on axis frequency response measurements are not valid.
Might be an idea to get a better understanding of the intended use of the system before posting findings.
I'm sure the lovely folk at PMC would be only too happy to engage to prevent scenarios like this.
My experience of PMC is that they are a very good company with sound engineering pedigree who make great products.
There are dozens of examples of very good sounding speakers which don't measure ruler flat.
Equally there are dozens of very bad sounding speakers which do measure ruler flat.
If there was a "correct" target frequency response everybody would be using it.
There isn't because there are too many variables involved. Everybody hears differently and the human hearing sensitivity changes hugely across the frequency bands. A flat speaker frequency response to a flat microphone doesn't mean a flat response to the ear brain system, when you consider how hearing sensitivity changes with frequency.
To judge a loudspeakers performance on how close it comes to a flat line is not taking into account the full picture.
And that is coming from someone who measures speakers for a living.
Of course one can use DSP alongside loudspeaker drive units, cabinet enclosures and crossover components to achieve a flat response however many designers don't have DSP available and have to as much as they can in the box.
Even if you do get a flat response out of the box your room will EQ the system as well as your ear brain system.
So you still have to listen and measurements can only be used to guide you in the right direction.

Off-axis performance is also pretty garbage. You fix some of the 5khz+ issues, but introduce new issues below that. Its just badly designed. With regards to your comments about frequency response: the research shows that ruler flat speakers are preferred by everyone. The only variation you will find is a preference for more bass or more treble (most likely due to changes to our hearing due to age), but it still has to be a smooth response. If you take two seconds to think about why a flat representation would be preferred its actually pretty self evident.

Performance also has nothing to do with DSP's or passive crossovers. Take KEF's Reference and R series, or the entire Revel line-up. Those show that you can get really flat performance without any DSP.

As for the influence of the room. You are correct that the room has a massive impact on how a speaker sounds. This also explains why you want a flat response with smooth off-axis performance. It is quite evident what happens if you don't meet these requirements and have to deal with reflections.

For reference, I have owned the Twenty.22's. They were a bit bright, but I mainly had issues with the low-end. It just wasn't as controlled as the DIY monitors I had (bass reflex using ScanSpeak drivers) and the M106's that it was replaced by. I didn't think the PMC was terrible to listen to. I mean I could still be quite happy with them if I hadn't heard those other speakers. People here make it sound like these PMC"s sound like a cheap bluetooth speaker, but that is a bit of an exaggeration.

 

[LTig]

PMC recommend that you listen to twenty series loudspeakers off axis. This is from the twenty 21 manual:
To further enhance the audio picture or soundstage the speakers can be angled/toed-in’. Start with the speakers angled so they will cross approximately 50cm (2ft) behind the listening position. (See stereo set-up diagram) Varying this angle will also subtly affect the vividness of the audio picture, so again experiment.
Over the years (and this is an old design) they have further developed this concept. This is from the more recent twenty 5 21i manual:
To enhance the audio picture/soundstage and sound balance, the speakers can be angled/toed-in. With the speakers facing straight ahead play a well recorded vocal track. Gradually toe-in both speakers by small increments (max 5° at a time) (see stereo set-up diagram) until the voice is vivid and centrally placed. You can ensure this is best position by going ‘one step too far’ then returning. Then, use a track with extended high frequencies such as cymbals, check that they sound clear but not overpowering. If they sound too loud or harsh subtly toe the speakers out by very small increments until you achieve the ideal balance; then repeat both listening tests to confirm.
So they clearly expect listeners to listen off axis and to use the associated HF roll off to achieve a "balanced" response. Had this of been considered the frequency response measurements would be, flatter and more useful.
As it stands the on axis frequency response measurements are not valid.
Might be an idea to get a better understanding of the intended use of the system before posting findings.
I'm sure the lovely folk at PMC would be only too happy to engage to prevent scenarios like this.
My experience of PMC is that they are a very good company with sound engineering pedigree who make great products.
There are dozens of examples of very good sounding speakers which don't measure ruler flat.
Equally there are dozens of very bad sounding speakers which do measure ruler flat.
If there was a "correct" target frequency response everybody would be using it.
There isn't because there are too many variables involved. Everybody hears differently and the human hearing sensitivity changes hugely across the frequency bands. A flat speaker frequency response to a flat microphone doesn't mean a flat response to the ear brain system, when you consider how hearing sensitivity changes with frequency.
To judge a loudspeakers performance on how close it comes to a flat line is not taking into account the full picture.
And that is coming from someone who measures speakers for a living.
Of course one can use DSP alongside loudspeaker drive units, cabinet enclosures and crossover components to achieve a flat response however many designers don't have DSP available and have to as much as they can in the box.
Even if you do get a flat response out of the box your room will EQ the system as well as your ear brain system.
So you still have to listen and measurements can only be used to guide you in the right direction.

Did you read the book by @Floyd Toole ? It covers a lot of scientific research which mostly contradicts what you wrote.

 

[Pharos]

There have been a few ex BBC engineers who have left to form their own businesses, not all of them are highly qualified, and I suspect that they rely to an extent on their BBC heritage to bolster their marketing position, and some may or may not be competent designers.

A flat measuring speaker which sounds bad may do so because of factors other than FR, and my impression is that Tool's work is valid, and will stand test of time.

Re earlier in the thread;
I cannot understand the distinct differentiation between 1/4 wavelength lines and reflex speakers, because it seems to me that the former is working as a reflex, but rather than as in the case of a reflex, lining the cabinet to remove cone rear energy, but allow the fundamental port resonance to be minimally affected, the line attenuates the mid energy gradually along a large length.
I do wonder about the possibility of transitory time delays in the latter.