• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Purifi Audio SPK5 Measurements | SoundStage/NRC

MZKM

Major Contributor
Forum Donor
Joined
Dec 1, 2018
Messages
4,242
Likes
11,464
Location
Land O’ Lakes, FL
https://www.soundstagenetwork.com/i...i-audio-spk5-loudspeakers&catid=77&Itemid=153

This is a speaker design that Purifi has the free plans for on their site:
https://purifi-audio.com/tech/

I see that @sgoldwin sell these on his website for $2830 USD shipped:
https://celuaris.com/products/celuaris-spk5-loudspeakers-featuring-purifi

Both drivers have been measured by HiFiCompass:
https://hificompass.com/en/speakers/measurements/purifi/purifi-ptt65w04-01a
https://hificompass.com/en/speakers/measurements/mundorf/mundorf-amt21cm21-c

Please note that measurements are only anechoic above ~100Hz.
0/15/30 H:
fr_on1530.png

Some peaking around the 3200Hz crossover, but other than that it’s pretty flat. It does have a very low sensitivity of 82-84dB though (they calculate 82dB from 300Hz-3kHz, but on the listening window).

45/60/75 H:
fr_456075.png

This is something I don’t understand. We see a narrowing of directivity starting at ~500Hz, however the measurements of the woofer show it doesn’t start to lose directivity until ~1500Hz. But looking at the Purifi documents, they too show a narrowing of directivity at 30°. I have no clue why this is happening, and maybe that peaking at the crossover region is intentional in order to better match the woofer and tweeter for the Sound Power, or else at 3200Hz the woofer’s SPL would have been much lower off-axis.

Listening Window (different than Spinorama, +/-15° H & V):
fr_listeningwindow.png

Flatter than 0°, so not sure if the tweeter axis is not the reference axis, or it is and those non-linearities vanish quickly vertically.

THD @ 101dB (measured 96dB @ 2m):
thd_95db.png

This is nothing short of amazing, besides the blip around 400Hz, you almost see no distortion in the graph >100Hz
 
Last edited:

Ilkless

Major Contributor
Forum Donor
Joined
Jan 26, 2019
Messages
1,757
Likes
3,437
Location
Singapore
https://www.soundstagenetwork.com/i...i-audio-spk5-loudspeakers&catid=77&Itemid=153

This is a speaker design that Purifi has the free plans for on their site:
https://purifi-audio.com/tech/

I see that @sgoldwin sell these on his website for $2830 USD shipped:
https://celuaris.com/products/celuaris-spk5-loudspeakers-featuring-purifi

Both drivers have been measured by HiFiCompass:
https://hificompass.com/en/speakers/measurements/purifi/purifi-ptt65w04-01a
https://hificompass.com/en/speakers/measurements/mundorf/mundorf-amt21cm21-c

0/15/30 H:
fr_on1530.png

Some peaking around the 3200Hz crossover, but other than that it’s pretty flat. It does have a very low sensitivity of 82-84dB though (they calculate 82dB from 300Hz-3kHz, but on the listening window).

45/60/75 H:
fr_456075.png

This is something I don’t understand. We see a narrowing of directivity starting at ~500Hz, however the measurements of the woofer show it doesn’t start to lose directivity until ~1500Hz. But looking at the Purifi documents, they too show a narrowing of directivity at 30°. I have no clue why this is happening, and maybe that peaking at the crossover region is intentional in order to better match the woofer and tweeter for the Sound Power, or else at 3200Hz the woofer’s SPL would have been much lower off-axis.

Listening Window (different than Spinorama, +/-15° H & V):

Flatter than 0°, so not sure if the tweeter axis is not the reference axis, or it is and these non-linearities vanish quickly vertically.

THD @ 101dB (measured 96dB @ 2m):
thd_95db.png

This is nothing short of amazing, besides the blip around 400Hz, you almost see no distortion in the graph >100Hz

The HD curve is spectacular, it measures like what a good speaker in its size class would at ~85dB (which is the base level Soundstage measures in the NRC).
 

TomJ

Active Member
Joined
Sep 16, 2019
Messages
129
Likes
177
Location
Palo Alto CA
thanks @MZKM for posting. very clean (!) but at the same price, I'd also consider the active KEF LS50 Wireless II with its DSP cross-over...
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,201
Likes
16,983
Location
Riverview FL
Too bad they cut off the distortion graph, in this case at about -45dBfs, or .56%
 

Ilkless

Major Contributor
Forum Donor
Joined
Jan 26, 2019
Messages
1,757
Likes
3,437
Location
Singapore
The blip. Port resonance? The Purifi, as Dennis' experiments show, is quite demanding on port design.
 

Absolute

Major Contributor
Forum Donor
Joined
Feb 5, 2017
Messages
1,084
Likes
2,125
Ok, I'll say it so that the rest of you won't have to. That's horrible. Aside from the distortion, of course.
You're welcome!
 

Absolute

Major Contributor
Forum Donor
Joined
Feb 5, 2017
Messages
1,084
Likes
2,125
I partially kid, of course. But in all honesty, a peaking due to lack of dispersion control in the area from 2,5-5 khz is difficult to get around. It'll most likely sound quite bright in most applications. Brightness is not my cup of tea!
 
OP
M

MZKM

Major Contributor
Forum Donor
Joined
Dec 1, 2018
Messages
4,242
Likes
11,464
Location
Land O’ Lakes, FL
I partially kid, of course. But in all honesty, a peaking due to lack of dispersion control in the area from 2,5-5 khz is difficult to get around. It'll most likely sound quite bright in most applications. Brightness is not my cup of tea!
As shown in the listening window, it is flat, so maybe the tweeter axis is not the reference axis, or maybe it goes away if going slightly above/below. This is where a Spinorama and vertical off-axis plot would come in handy.
 

Matias

Master Contributor
Forum Donor
Joined
Jan 1, 2019
Messages
5,029
Likes
10,799
Location
São Paulo, Brazil
Now that we have UMIK-1 + REW => PEQ to fix amplitude easily and cheap, isn't a smooth dispersion and low distortion more important than having a great linearity? Maybe that is a broad question for another topic.
 

Absolute

Major Contributor
Forum Donor
Joined
Feb 5, 2017
Messages
1,084
Likes
2,125
As shown in the listening window, it is flat, so maybe the tweeter axis is not the reference axis, or maybe it goes away if going slightly above/below. This is where a Spinorama and vertical off-axis plot would come in handy.
Perhaps, but I guess that the +-15 degree listening window as shown here is very inaccurate. It exaggerates the vertical importance and dismiss the much greater importance of horizontal dispersion into the first reflection angles. The 45/60/75 angle measurements tells the true story, imo.
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
As shown in the listening window, it is flat

Must admit, I can't quite see what you're seeing there :) To me, the LW response looks just as far from flat, albeit a little less peaky. All the resonances are in the same places, just smoothed somewhat as a result of averaging:

1603815740287.png


I agree though that the on-axis peak at 3-ish kHz is likely to be somewhat offset in the PR/PIR etc. by a vertical off-axis suckout around the XO. I can see how this could at least be spun as a feature rather than a defect....
 
D

Deleted member 14468

Guest
Hi,

Being the person who got these measured, I'd like to interject. One thing that must be remembered is that Purifi isn't selling this speaker -- it was sent out as an evaluation unit to look at the behavior of their midrange-woofer. As such, its sensitivity in this design is very low, presumably to deliver as much bass as possible, and it's crossover point is at 3.2kHz, which is quite high, but gives a clearer picture of how the driver behaves that high in frequency, which is really the point.

Remember, this same driver can be implemented six ways till Sunday in other designs.

Doug Schneider
SoundStage!
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Hi,

Being the person who got these measured, I'd like to interject. One thing that must be remembered is that Purifi isn't selling this speaker -- it was sent out as an evaluation unit to look at the behavior of their midrange-woofer. As such, its sensitivity in this design is very low, presumably to deliver as much bass as possible, and it's crossover point is at 3.2kHz, which is quite high, but gives a clearer picture of how the driver behaves that high in frequency, which is really the point.

Remember, this same driver can be implemented six ways till Sunday in other designs.

Doug Schneider
SoundStage!

Fair point, Doug. The measurements make a bit more sense knowing that.

I'd also like to note that nobody here was (as far as I could tell) criticising the individual drivers in this design. Hopefully that's clear.

I also can't help wondering: why design an unnecessarily compromised speaker just to demonstrate how the woofer performs high in frequency, given that the woofer itself can be tested independently of any particular implementation? (And indeed has been tested, extensively, and shown to behave very well.) Not expecting anyone to answer this ofc, just wondering out loud...
 
D

Deleted member 14468

Guest
I also can't help wondering: why design an unnecessarily compromised speaker just to demonstrate how the woofer performs high in frequency, given that the woofer itself can be tested independently of any particular implementation? (And indeed has been tested, extensively, and shown to behave very well.) Not expecting anyone to answer this ofc, just wondering out loud...

I wouldn't call that much of a compromise. Also, the idea was to implement in a design that someone could play -- not necessarily take it out of.

But also, from what I understand, in their original demo speaker they had a lower crossover point, but the Purifi driver performed better at the crossover region than the AMT did, so they raised the crossover point to where it is today.

Doug
SoundStage!
 
OP
M

MZKM

Major Contributor
Forum Donor
Joined
Dec 1, 2018
Messages
4,242
Likes
11,464
Location
Land O’ Lakes, FL
Must admit, I can't quite see what you're seeing there :) To me, the LW response looks just as far from flat, albeit a little less peaky. All the resonances are in the same places, just smoothed somewhat as a result of averaging:

View attachment 89852

I agree though that the on-axis peak at 3-ish kHz is likely to be somewhat offset in the PR/PIR etc. by a vertical off-axis suckout around the XO. I can see how this could at least be spun as a feature rather than a defect....
It fits within a 4dB window, which would be +/-1.5dB. So, not as flat as can be, but perceptually pretty flat (though the low-Q dip in the middle of the treble may be audible.
 
Last edited:

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,109
Likes
8,424
Location
NYC
I think, considering the impressive bass extension and low distortion, this is quite good. The LW is almost entirely within +/- 1.5 dB, and if this had been in one of Amir's reviews with their shorter scaling, it would look a better than it does here. Imo we've given passes to worse speakers than this in Amir's reviews.

I doubt the 400Hz blip will be too audible, as it seems to largely disappear by 75-degrees off axis -- it will probably be largely smoothed out in the power response/PIR and at this frequency the room is starting to take over anyway.

That leaves the 3K peak and subsequent scoop as the biggest problems, but not so much a problem as to negate the benefits in low-end extension.

*EXTREME SPECULATION ALERT*
Doing some super rudimentarty messing around with vituixcad --basically guesstimating vertical lobing as we don't have that information -- I think we'd get a PIR and power response roughly like this:

1603824268754.png


This is almost certainly cleaner than it would look in reality, but my guess is that the peaking largely balances out in room, though I do generally believe that at higher frequencies the direct sound takes over so this remains a flaw.

For someone looking for passive speakers with no sub, these look like a great option, assuming they have enough power for their listening levels.

That said, it does seem maybe opting for a different tweeter and settings the crossover just a bit lower would solve that 3K peak but what do I know! I'm not a speaker designer :)
 
Last edited:

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Yeh, look, I realise I'm applying quite high standards in my criticisms of this speaker. Essentially, my rationale is that this speaker is supposedly (or at least is likely to be assumed to be) a reference 2-way design using a woofer that is (not without some basis) touted by its manufacturer to be the best woofer of its class on the market. Moreover, the speaker is being promoted by said manufacturer on its website (I'm not sure if Purifi is also responsible for actually designing the speaker, but that question seems neither here nor there).

So I'm not suggesting this is a bad speaker in absolute terms. I'm just contending that it is a mediocre speaker whose design is somewhat primitive relative to other examples, and that it fails to be the equal of the potential sum of its parts ;)
 
D

Deleted member 14468

Guest
Must admit, I can't quite see what you're seeing there :) To me, the LW response looks just as far from flat, albeit a little less peaky. All the resonances are in the same places, just smoothed somewhat as a result of averaging:

I just want to make sure no one gets the wrong idea. Our measurements are done at 1/24th octave. There is no smoothing that some people might miscontrue, but the Listening Window is "averaged," which doesn't smooth, per se, but gives equal weighting to five measurement axis. That might've been what you meant, but I just wanted to make sure others knew.

Doug
SoundStage!
 
Last edited by a moderator:
Top Bottom