This is a review of the Purifi SPK5 reference speaker design for their PTT6.5W04-01A 6.5 Inch Woofer. It was built and sent to me by the company, Celuaris (member @sgoldwin). The unit was shipped from UK and did not travel well. Just about everything that could come loose, did come loose. I had to take the back apart and reconnect everything and solder a few components in the Purifi crossover board. Final product from the company will not have these issues. Still, I think we can get a lot of valuable data from the design as is.
EDIT: An updated sample was tested that fixed the enclosure leakage. As such please refer to new measurements than the ones here: https://www.audiosciencereview.com/...ds/celuaris-spk5-purifi-review-speaker.20125/
Celuaris has built the speaker to Purifi spec and visually it looks to be so:
The backside is butchered up by me opening it back together (screws were very stiff resulting in me stripping a bunch ).
The port is "S" shaped inside. There was barely enough room for it inside and since it was not glued together, it may not be in the optimal position as designed.
The crossover board is external and uses Purifi layout and build:
I did not appreciate the black silkscreen/solder mask one bit. It completely covered the traces making it very hard for me to figure out what went where. Why on earth a development board was taught to be better this way? Confusing the implementer can't be a good thing. I expect more from Purifi. Looks should be secondary to easy of use.
Note that the bent components is my fault. I had to flip it multiple times and they started to loose their straight orientation. Fortunately electrons don't care.
As you see, fancy capacitors and such are used throughout. Combined with high cost of the Purifi woofer ($365?), this would not be a cheap speaker if so dressed.
For my testing I had to put the crossover on top of the speaker separated by a 2 inch piece of foam. The Klippel NFS robotic scanner goes around the speaker and nothing can be separate from it or it will dangle. I don't think that made much of a difference in measurements but thought I mention it.
Note that this test is for experienced readers to analyze the design components here and less so about the finished speaker. Company plans to send me one of those later for proper review.
The measurements were made at about 61 degrees F which was comfortable to me but not if you are from a sunny place.
I used over 1000 measurement points which gave me less than 1% error for most of the spectrum with the exception of 2.5 to 8 kHz where the error crept up a bit more to around 2%.
Purifi SPK5 Measurements: Spinorama Frequency Response
As usual we start with anechoic response of the speaker plotted in multiple ways:
Well, this is not that good. Starting from left, we have a dip in upper bass frequencies. Things get better until we start to approach the crossover frequency and now we face a broad dip from 1.2 kHz to some 7 kHz! There was talk of Purifi adjusting the crossover region, pushing it up to 3.5 kHz. Don't know if that is the cause but what is there is definitely not optimal. We then hit a resonance around 8 kHz and a few more peaks with elevated response from the tweeter.
I researched other measurements and while there is some agreement with this data, they vary fair bit so I thought I dig in to see if what we have is correct. Let's start with near-field response of various radiating sources:
I had to adjust the port level so don't take that as absolute. What is strange about the port is that it does not augment the low frequency response of the woofer as we normally see. It simply reinforces what it can produce as is. I could be OK with it but not that sharp resonance that follows between 300 and 400 Hz. That corresponds with a dip in our frequency response so perhaps this is out of phase and sucks out the energy in that region.
Moving to the next region, that crossover region doesn't look right to me. We have that long and gradual slope for the woofer before dropping like a rock. Yet the tweeter has a sharper slope. Can't imagine those summing correctly. They do agree with the electrical measurements of the crossover as provided by Purifi though:
The red curve when applied to near flat response of the woofer by itself (as provided by Purifi) will result in what I have measured. So we are seeing the response as designed. Just doesn't sum well.
Next is the resonances in the tweeter. I looked up the specs for the Mundorf AMT AMT 21CM 2.1 tweeter and this is what they show:
A very large vertical scale of 80 dB (as opposed to standard 50 dB) helps to hide the same peaks we see in our measurements as marked. Was there not better choices here?
Purifi shows the following measurements which were sadly gated and in-room and hence, lack the resolution to see the port problems:
Notice that the tweeter ringing is there. What is different is that they show same ringing in the crossover region which hides the wide dip that I measured. Disappointing that a company like Purifi would not bother getting a proper anechoic measurement. If they had, these problems may have been fixable.
Anyway, going back to our regular programming, here is the early window reflections:
It seems to be more "correct" which helps tame the on-axis response issues when we look at the combined result:
Horizontal beam width shows decent response until the narrow width tweeter starts to beam toward the end (gets directional):
And the same as contour plot:
Vertical dispersion says you better stay at tweeter level or slightly below:
Purifi SPK5 Distortion Measurements
The main claim to fame of the Purifi woofer is lower distortion. So let's look at our standard measurements:
Ignoring the bass for a moment, I am not happy to see the rising distortion in lower treble. Anyone want to guess as to the source?
Here is our 96 dBSPL @1 meter graph:
Was hoping to see nothing above 50 dBSPL but we have fair bit. I leave it up to you all to dig up past measurements and compare to other speakers with similar sized woofer.
The distortion amount is much worse than what Purifi shows from my quick look. Looking at contributions to it, I see a lot more of it from the port than the woofer. My reassembly of the speaker may have something to do with this as well. Not sure.
Here is the impedance graph:
Our port issue/resonance is clearly apparent. Can't figure out the messiness around 30 to 40 Hz though. Anyone has a guess for that?
Since this is an ultra technical write up, I thought I include the waterfall/CSD plot as well:
I have not tried to adjust thresholds and such. But we can see some of the same issues we have seen in other measurements.
Purifi SPK5 Speaker Listening Tests
I placed the speaker in my usual stand in my listening room with the crossover separated from it (my poor Revel Salon 2 speaker acted as a crossover stand!). I always start with female music tracks to see if the highs are produced well without being too bright or recessed. Well, here they were clearly recessed and not that nice. So the EQ tool came right out resulting in this set after fair bit of experimentation:
The corrections follow the response errors. Lifting up the upper bass resulted in much needed bass energy as without it, there is just not a lot of it there. Broad EQ around 2500 Hz dealt with the crossover deficiency quite well, lifting the female vocals out and giving them the treatment they need: a bit brilliant and forward without it being too much. Actually, it was a too much at first without the higher frequency corrections. A PEQ dialed to lower the resonance at 8 kHz was very helpful as was the quick shelving filter above that.
At first I had the room mode dip active but as noted, there is not a lot of bass coming out of this speaker so I took it out. This helped and by this time, the response was very nice and high quality. Lots of detail without any bass notes becoming overbearing. It was a bit like getting rid of a room mode using EQ and the feeling of relief you get after that.
I played by "speaker killer tracks" and the SPK5 handled them quite well up to a point which was way ahead of what most small woofers can handle. It then followed with a bit of distortion. The woofer was going nuts as far as excursion by this time which would have really messed up the sound of other woofers. Push a bit more though and you are greeted with a static pop that implies the coil getting outside of the magnetic field.
There was a big miss though and that was any kind of physical bass response. Even at much elevated levels, you could barely feel the bass. You could hear it well and on techno tracks, celebrated it. But as a sensation which I crave, almost none was there. I wonder if this is due to the tuning of the poor which doesn't lower the response range.
Overall experience was a very light on its feet speaker that produced very satisfying sound with equalization. With out, it would not be my cup of tea.
Conclusions
Great to see innovations in woofer technology. Alas, the lesson of research in the last four decades was ignored in this development reference design: getting the tonality right. There could not be a more clear example here. I don't care how low the distortion is if you screw up the vocals. Or sound too bright. Fix these -- as I did with EQ -- and then the lower distortion of the bass (appears to ) come forward.
Thinking out loud, I wonder how much bass energy we loose by lowering distortion. Harmonics of low frequencies perceptually would raise the total energy in that region. I wonder if this, together with port issue, was behind lack of sufficient impact in bass response.
I am left thinking if a 6.5 inch driver is the right choice for this technology. Maybe an 8 inch driver would have been a better showcase so that we could have bass and low distortion. Alternatively maybe other designs would do better. I don't know. What I do know is that this was not a good attempt by Purifi to showcase their speaker.
Let's hope some issues here are due to difficulty I had in putting this speaker in working condition. I welcome detailed analysis from you all though on how it could be improved if the problems are real.
Overall I cannot recommend the Purifi SPK5 without equalization.
------------
As always, questions, comments, recommendations, etc. are welcome.
Appreciate any donations using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
EDIT: An updated sample was tested that fixed the enclosure leakage. As such please refer to new measurements than the ones here: https://www.audiosciencereview.com/...ds/celuaris-spk5-purifi-review-speaker.20125/
Celuaris has built the speaker to Purifi spec and visually it looks to be so:
The backside is butchered up by me opening it back together (screws were very stiff resulting in me stripping a bunch ).
The port is "S" shaped inside. There was barely enough room for it inside and since it was not glued together, it may not be in the optimal position as designed.
The crossover board is external and uses Purifi layout and build:
I did not appreciate the black silkscreen/solder mask one bit. It completely covered the traces making it very hard for me to figure out what went where. Why on earth a development board was taught to be better this way? Confusing the implementer can't be a good thing. I expect more from Purifi. Looks should be secondary to easy of use.
Note that the bent components is my fault. I had to flip it multiple times and they started to loose their straight orientation. Fortunately electrons don't care.
As you see, fancy capacitors and such are used throughout. Combined with high cost of the Purifi woofer ($365?), this would not be a cheap speaker if so dressed.
For my testing I had to put the crossover on top of the speaker separated by a 2 inch piece of foam. The Klippel NFS robotic scanner goes around the speaker and nothing can be separate from it or it will dangle. I don't think that made much of a difference in measurements but thought I mention it.
Note that this test is for experienced readers to analyze the design components here and less so about the finished speaker. Company plans to send me one of those later for proper review.
The measurements were made at about 61 degrees F which was comfortable to me but not if you are from a sunny place.
I used over 1000 measurement points which gave me less than 1% error for most of the spectrum with the exception of 2.5 to 8 kHz where the error crept up a bit more to around 2%.
Purifi SPK5 Measurements: Spinorama Frequency Response
As usual we start with anechoic response of the speaker plotted in multiple ways:
Well, this is not that good. Starting from left, we have a dip in upper bass frequencies. Things get better until we start to approach the crossover frequency and now we face a broad dip from 1.2 kHz to some 7 kHz! There was talk of Purifi adjusting the crossover region, pushing it up to 3.5 kHz. Don't know if that is the cause but what is there is definitely not optimal. We then hit a resonance around 8 kHz and a few more peaks with elevated response from the tweeter.
I researched other measurements and while there is some agreement with this data, they vary fair bit so I thought I dig in to see if what we have is correct. Let's start with near-field response of various radiating sources:
I had to adjust the port level so don't take that as absolute. What is strange about the port is that it does not augment the low frequency response of the woofer as we normally see. It simply reinforces what it can produce as is. I could be OK with it but not that sharp resonance that follows between 300 and 400 Hz. That corresponds with a dip in our frequency response so perhaps this is out of phase and sucks out the energy in that region.
Moving to the next region, that crossover region doesn't look right to me. We have that long and gradual slope for the woofer before dropping like a rock. Yet the tweeter has a sharper slope. Can't imagine those summing correctly. They do agree with the electrical measurements of the crossover as provided by Purifi though:
The red curve when applied to near flat response of the woofer by itself (as provided by Purifi) will result in what I have measured. So we are seeing the response as designed. Just doesn't sum well.
Next is the resonances in the tweeter. I looked up the specs for the Mundorf AMT AMT 21CM 2.1 tweeter and this is what they show:
A very large vertical scale of 80 dB (as opposed to standard 50 dB) helps to hide the same peaks we see in our measurements as marked. Was there not better choices here?
Purifi shows the following measurements which were sadly gated and in-room and hence, lack the resolution to see the port problems:
Notice that the tweeter ringing is there. What is different is that they show same ringing in the crossover region which hides the wide dip that I measured. Disappointing that a company like Purifi would not bother getting a proper anechoic measurement. If they had, these problems may have been fixable.
Anyway, going back to our regular programming, here is the early window reflections:
It seems to be more "correct" which helps tame the on-axis response issues when we look at the combined result:
Horizontal beam width shows decent response until the narrow width tweeter starts to beam toward the end (gets directional):
And the same as contour plot:
Vertical dispersion says you better stay at tweeter level or slightly below:
Purifi SPK5 Distortion Measurements
The main claim to fame of the Purifi woofer is lower distortion. So let's look at our standard measurements:
Ignoring the bass for a moment, I am not happy to see the rising distortion in lower treble. Anyone want to guess as to the source?
Here is our 96 dBSPL @1 meter graph:
Was hoping to see nothing above 50 dBSPL but we have fair bit. I leave it up to you all to dig up past measurements and compare to other speakers with similar sized woofer.
The distortion amount is much worse than what Purifi shows from my quick look. Looking at contributions to it, I see a lot more of it from the port than the woofer. My reassembly of the speaker may have something to do with this as well. Not sure.
Here is the impedance graph:
Our port issue/resonance is clearly apparent. Can't figure out the messiness around 30 to 40 Hz though. Anyone has a guess for that?
Since this is an ultra technical write up, I thought I include the waterfall/CSD plot as well:
I have not tried to adjust thresholds and such. But we can see some of the same issues we have seen in other measurements.
Purifi SPK5 Speaker Listening Tests
I placed the speaker in my usual stand in my listening room with the crossover separated from it (my poor Revel Salon 2 speaker acted as a crossover stand!). I always start with female music tracks to see if the highs are produced well without being too bright or recessed. Well, here they were clearly recessed and not that nice. So the EQ tool came right out resulting in this set after fair bit of experimentation:
The corrections follow the response errors. Lifting up the upper bass resulted in much needed bass energy as without it, there is just not a lot of it there. Broad EQ around 2500 Hz dealt with the crossover deficiency quite well, lifting the female vocals out and giving them the treatment they need: a bit brilliant and forward without it being too much. Actually, it was a too much at first without the higher frequency corrections. A PEQ dialed to lower the resonance at 8 kHz was very helpful as was the quick shelving filter above that.
At first I had the room mode dip active but as noted, there is not a lot of bass coming out of this speaker so I took it out. This helped and by this time, the response was very nice and high quality. Lots of detail without any bass notes becoming overbearing. It was a bit like getting rid of a room mode using EQ and the feeling of relief you get after that.
I played by "speaker killer tracks" and the SPK5 handled them quite well up to a point which was way ahead of what most small woofers can handle. It then followed with a bit of distortion. The woofer was going nuts as far as excursion by this time which would have really messed up the sound of other woofers. Push a bit more though and you are greeted with a static pop that implies the coil getting outside of the magnetic field.
There was a big miss though and that was any kind of physical bass response. Even at much elevated levels, you could barely feel the bass. You could hear it well and on techno tracks, celebrated it. But as a sensation which I crave, almost none was there. I wonder if this is due to the tuning of the poor which doesn't lower the response range.
Overall experience was a very light on its feet speaker that produced very satisfying sound with equalization. With out, it would not be my cup of tea.
Conclusions
Great to see innovations in woofer technology. Alas, the lesson of research in the last four decades was ignored in this development reference design: getting the tonality right. There could not be a more clear example here. I don't care how low the distortion is if you screw up the vocals. Or sound too bright. Fix these -- as I did with EQ -- and then the lower distortion of the bass (appears to ) come forward.
Thinking out loud, I wonder how much bass energy we loose by lowering distortion. Harmonics of low frequencies perceptually would raise the total energy in that region. I wonder if this, together with port issue, was behind lack of sufficient impact in bass response.
I am left thinking if a 6.5 inch driver is the right choice for this technology. Maybe an 8 inch driver would have been a better showcase so that we could have bass and low distortion. Alternatively maybe other designs would do better. I don't know. What I do know is that this was not a good attempt by Purifi to showcase their speaker.
Let's hope some issues here are due to difficulty I had in putting this speaker in working condition. I welcome detailed analysis from you all though on how it could be improved if the problems are real.
Overall I cannot recommend the Purifi SPK5 without equalization.
------------
As always, questions, comments, recommendations, etc. are welcome.
Appreciate any donations using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
Attachments
Last edited: