XSA Labs Vanguard Speaker Review

[sarumbear]

Klippels are not measurements, rather assessments.

You may want to check the dictionary for the meaning of those words.

Anyhow, as you are not interested in rational conversation I’ll stop responding to you. Besides, we went off-topic with discussions of Klippel.

 

[Vladimir Filevski]

Maybe the difference between the measurements made by Amir and those made by the designer are due to the fact that the xover is not optimized for on axis response, a smoother one being obtained at 15-30° off axis

Maybe the problem here is that Amir takes the on axis response as absolute reference for all his measurements, and if this is not the designer's one the resulting measurements are kindof a mess...

I am now a bit confused, ...

Yes, you are very confused. You don't know the basics of loudspeaker measurement.

Dont agree, except 0° on axis, these are all averaged measurements around 0°, as is listenning window for example.

Where do you see the response at 15° 30° 45°, etc...?

Here:

On the left side are horizontal degrees, from 0 to 180. You can choose any angle, including 15, 30, 45, ... On the right side is normalized SPL described in colors. It is easy to see that there is no off-axis frequency response which is better than the on-axis.

Yes but there is something in Amir s loudspeaker measurements that do not convince me at all. And this is not the first time.

This is a very big nonsense. And that is an understatement.

After all a designer is perfectly allowed to choose the design axis freely and it does no have to be necessarily 0°
And further measurements should be made taking this into account, taking as reference the axis giving the flattest and smoothest response as presumably the design axis.

The flattest and smoothest response is on the axis! See above measurement.

Nevertheless cabinet diffraction can produce such bump in the midrange if taken strictly on axis.

That bump is present in all off-axis measurements, so it is not diffraction.

 

[badspeakerdesigner]

Interesting, as a member of DIYaudio it's cool to see one of the popular members designs tested. It's disappointing performance to say the least and makes me a little wary of diving into some his designs. I'd say this is example of a designer running into the fact that Klippel is several orders of magnitude more accurate of analysis system than whatever the designer is using. It's happened before (Paul C overnight sensations Erin's Audio Corner review). I'm sure the diyaudio thread is going to be a place to bring some popcorn for awhile.

I have this DC130 woofer, albeit the round frame variant. I use it in a big ATC lookin three way speaker. I can tell you a few things about it. It's actually a lovely woofer and pretty easy to work with (although the new Dayton klippel optimized signature woofers make this a poor buy anymore). This Vangaurd design is crossing it way to high, the limit IME is 3k and that's it. I settled on 2.5k. I'm also no longer using the tweeter in this picture, but a waveguided ND25FW as it sounds much better (dispersion matching is very important and very audible). The woofer in my little enclosure measures very flat from about 150hz to 3k then rises up with mild breakup. IME you simply cannot use the woofer as high as this reviewed design does, it does not sound good. The rise in the midrange reminds me of the ATC SCM19 reviewed here and my experience tells me that the xover design is lazy. Almost all woofer look like that when placed on a typical sized baffle and flattening it down is one of the first vital steps for woofer integration. The nulling in dispersion on the horizontal plane tells me the drivers have too much overlap there.

The response doesn't really resemble any LS3/5A response I've seen so I don't think they nailed that. Ok so it does kinda look like it has somewhat decent adherence to some LS3/5a speakers. I'd argue one could aim for something similar yet improve on the shortcomings of the target design and offer a better sounding speaker.

For those interested, my in room response looks like this. These speakers here cost me about half of the cost of the Vanguard if you count the three month build time as $0 in labor and forget tooling costs.

 

[AdamG]

OK, we drifted right off the road and I called for a Wrecker to come and pull us out of the Thread Drift Sand Dune!

This is an Official Review Thread and my Boss is a relentless workaholic so please help me get this Train back on topic…

Thank you for your support and cooperation.

 

[Sokel]

Bamboo is no surprise as it trends as an ecological solution and if wood color is your thing I guess is nice and respected if reduces the ecological print.

BBC dip is not what that response looks like thought,I think that the lovers of the particular response will not find comfort there.

Thanks Amir!

 

[SpeaKa]

They state on their website, that it was the first time ever, that someone implemented a "inverted Harsch crossover topology" in a passive commercial speaker. Must be for some reason, other designers avoided to use it.

But otherwise I somehow like the looks and the absence of a port. A simple redesign of the (quite simple) crossover could save this flawed design.

Thank you Amir: more speaker testing, please!!!

A "S. Harsch" Crossover would mean to use a 4th Order Lowpass XO combined with a 2nd order Highpass XO for the tweeter. This is used to block out the cone breakup of woofers. Inverting this principle does not really sense for me and with the shown number of parts, here could be such a type of XO.

Confusing for me: The bar core inductor at that size should have at least 1.5mF or maybe far more than that. For a 2nd order lowpass that would end up in a crossover frequency around 1.5kHz or even lower, which is not confirmed by Amirs measurements...

 

[dukanvadet]

The designer have commented on the review at diyaudio. Seems like he trusts his own measurements over amirs, not sure why. Or does he mean there is something wrong with the DUT? Not sure. I read alot of his posts of what he done as diy projects and thought it seemed good but obviously this is not.

He also carries on the comparison with ls3/5. Odd both because its only similar in the basic shape and size and because its not really that great by todays standard

 

[amirm]

I see that he says the current production parts are all bad. I post a message to contact me to test another one but for some reason, my account is set to zero and my post is awaiting moderator approval. Looks like they updated their forum software???

 

[xrk971]

Hi Amir,
Thanks for reaching out to me on DIYA. We look forward to getting you a set of Vanguards with nominally operating drivers as it seems this particular unit tested may have off-spec drivers. I have sent you a PM.
Thanks,
Viet

 

[kencreten]

You guys are spending too much time at ikea.
I don't mind the use of Bamboo. Lot more sustainable than wood. Could even double up as a cheese board.

Someone mention a deal on cutting boards? ......

 

[RF Air]

Thank you Amir for the Review and all the efforts you make to provide us details that serve the Forum Community.

It is interesting to me that products that are found to be substandard in design and performance are defended by posters with obscure notions and obfuscation of facts. Hence the need for the service you provide for our community (and our Moderators).

Any other product is calculable by standard, like a measure of fuel, or a shoe size. Interesting that we can measure Audio Science, but derision exists to refute the standards.

 

[Maiky76]

This is a review, listening tests, EQ and detailed measurements of the XSA Labs Vanguard 2-way sealed speaker. It is on kind loan from a member and costs US $999 for a pair.
View attachment 300870
The bamboo version I have is machined so well that the corners are almost knife sharp! Magnetic grill is a nice touch but for testing, I removed it as you see. Not much to the back panel as to show that it is a sealed design:
View attachment 300871

Seems like the company name used to be Vera-Fi.

If you are new to my testing, please watch this tutorial on understanding speaker measurements.

XSA Labs Vanguard Speaker Measurements
Let's start with our usual anechoic frequency response measurements created by Klippel Near-field Scanner:
View attachment 300872
Well, this is not what we want to see! We have massive and broad peaking of the response from about 700 to 3 kHz and a couple of notches. There seems to be no attempt at creating a neutral response. Sensitivity drops to a very low 80 dB or even lower at some frequencies. Company post a DIY style measurement in a forum showing something different:
View attachment 300873

While their is also pretty bad it shows that using the wrong protocol for testing can hide actual design issues.

Simple near-field driver measurements show the core problem:

View attachment 300874

Notice the wide overlap between the tweeter and woofer. The two add up to create the "hill" we see in the anechoic frequency response measurement.

The mid-woofer is close enough in size to the tweeter which makes for decent directivity. That makes it easier to EQ but without, it causes off-axis to be just as bad as on-axis:
View attachment 300932
This makes predicted-in-room response decidedly poor:
View attachment 300876

Allowing the woofer to play too high seems to be the reason for higher distortion in that region:
View attachment 300877

View attachment 300878


Digging into directivity again, we see confirmation of what I stated above:
View attachment 300879

If we didn't have that notch between 4 and 5 kHz, this would have been a pretty good response. For that reason though, you better toe in the speaker toward you:
View attachment 300880

Vertically we have to stay at tweeter axis as is typically the case with 2-way traditional designs like this:
View attachment 300881

Company spec of 8 ohm impedance seems right for the lowest it gets:
View attachment 300882

Finally, here is our waterfall and step responses:
View attachment 300883

View attachment 300884

XSA Labs Vanguard Listening Tests and Equalization
I usually start with my female vocals but this time, I had one of my bass tracks queued up so hit play on that. I was surprised how little volume I was getting out of the speaker even after I cranked it way up (I have a 400 watt Mark Levinson amplifier driving it). I stopped and switched to my regular female vocal and volume shot way up. The sound was inoffensive and one could think this is good sound (if you didn't play anything with bass in it). Recalling the poor frequency response measurements, I brought out the EQ and first filter I dialed in was for the broad excess energy in lower treble:

View attachment 300886

That pulled the female vocals back, restoring a neutral response although I could see someone just listening to female vocals liking the stock sound maybe better. To my ears, there was not much warmth so I pulled up the response in upper bass as measurements showed. That was nice but then I had some boominess on a track. I have a room mode around 105 Hz and that is where this speaker peaks as well so I pulled that down some. Now the sound was much more balanced but still not to my liking. Many of my reference tracks didn't sound good.

I remembered that in the past I have had good luck in getting more bass out of sealed cabinets by pulling up the bass response a bit. If the drivers handle it without getting distortion, the effect is like having a port but without its drawback of transmitting cabinet resonances. Indeed that was the case here with that 70 Hz filter. For the first time, I could actually get some physical sensation of bass out of the floor.

With all of these filters in place, the sound was good enough for about half of my tracks. Distortion was not audible unless I really pushed the speaker where it started to creating some ticking sounds. It was plenty loud just playing one speaker so I think you will be fine with a pair.

Conclusions
Company website is high on review quotes and low on information. But reading around it seems that the company thinks they have designed something akin to BBC LS3/5A (???). Whatever the aim was, it is a complete miss in practice. They need to up their game in measurements and spend more time performing proper AB testing using other speakers blind, or as i do, with electronic EQ. Just listening to one speaker, especially one's own design, is likely to lead you into a ditch and that is precisely what we have. So while I appreciate the woodworking here, the speaker itself is a miss on grounds of simple principles of designing a neutral speaker that appeals to many audiophiles.

Fortunately the rather similar sizes of the drivers makes directivity decent enough that simple EQ filtering makes a dramatic difference.

I cannot recommend the XSA Labs Vanguard speaker.

----------
As always, questions, comments, recommendations, etc. are welcome.

Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

Here is my take on the EQ.

Please report your findings, positive or negative!

The following EQs are “anechoic” EQs to get the speaker right before room integration. If you able to implement these EQs you must add EQ at LF for room integration, that is usually not optional… see hints there: https://www.audiosciencereview.com/...helf-speaker-review.11144/page-26#post-800725

The raw data with corrected ER and PIR:

Score no EQ: 3.2
With Sub: 6.0

Spinorama with no EQ:

Poor
Decent DI

Directivity:

Better stay at tweeter height or just above
Horizontally, better toe-in the speakers by 10/20deg and have the axis crossing in front of the listening location, might help dosing the upper range.

EQ design:
I have generated two EQs. The APO config files are attached.

• The first one, labelled, LW is targeted at making the LW flat
• The second, labelled Score, starts with the first one and adds the score as an optimization variable.
• The EQs are designed in the context of regular stereo use i.e. domestic environment, no warranty is provided for a near field use in a studio environment although the LW might be better suited for this purpose.

Score EQ LW: 5.1
with sub: 7.5

Score EQ Score: 6.1
with sub: 8.4

XSA Labs Vanguard APO EQ LW 96000Hz
July242023-160405

Preamp: -1.8 dB

Filter 1: ON HPQ Fc 52.58,    0.00,    1.15
Filter 2: ON PK Fc 133.99,    -3.79,    1.54
Filter 3: ON PK Fc 1160.68,    -5.46,    1.39
Filter 4: ON PK Fc 2181.91,    -3.61,    2.21
Filter 5: ON PK Fc 3228.22,    -1.39,    2.02
Filter 6: ON PK Fc 7937.79,    1.89,    0.51

XSA Labs Vanguard APO EQ Score 96000Hz
July242023-160405

Preamp: -1.3 dB

Filter 1: ON HPQ Fc 52.09,    0.00,    1.15
Filter 2: ON PK Fc 132.87,    -3.59,    1.20
Filter 3: ON PK Fc 1175.02,    -5.39,    1.27
Filter 4: ON PK Fc 2391.92,    -4.61,    1.77
Filter 5: ON PK Fc 3661.89,    -4.10,    4.99
Filter 6: ON PK Fc 4354.75,    3.75,    2.39
Filter 7: ON PK Fc 4982.74,    -2.75,    4.99

Spinorama EQ LW

Spinorama EQ Score

Zoom PIR-LW-ON

Regression - Tonal

Radar no EQ vs EQ score
Large improvements

The rest of the plots is attached.

@pierre

I used your EQ to compared with my Score:
1. The sensitivities should be pretty much the same, it's just your normalization frequency and mine that are not the same. I had to off your EQ by 4.2dB, so total my EQ is -5.5dB yours is -5dB.
2. My scores are always lower than yours for some reasons... I think my scores are 0.1 lower for textbook score and identical for LF to @MZKM 's.
Most probably linked with LF calculation. Could be instructive to see why we have a difference.
3. Since you have the same number of BQ and the same sensitivity I though it could be interesting to see what's different

No EQ: 3.2/6.0 (3.4/6.1 with your calculation)
Pierre EQ: 5.6/7.8 (5.85/? with your calculation)
Score EQ Maiky: 6.1/8.4

EQ Design

Zoom:

Pierre EQ spinorama

Radar:

 

[sarumbear]

We look forward to getting you a set of Vanguards with nominally operating drivers as it seems this particular unit tested may have off-spec drivers.

So do I. I am very much intrigued how off-spec drivers caused the 2-octave overlap in the crossover.

PS. On this review it took 111 posts until the "it was a broken unit" comment to arrive

 

[AdamG]

So do I. I am very much intrigued how off-spec drivers caused the 2-octave overlap in the crossover.

PS. On this review it took 111 posts until the "it was a broken unit" comment to arrive

Well they didn’t toss out the old gem, “It must have been damaged in shipping”. So I see that as progress.

 

[GDWL34]

This is a review, listening tests, EQ and detailed measurements of the XSA Labs Vanguard 2-way sealed speaker. It is on kind loan from a member and costs US $999 for a pair.
View attachment 300870
The bamboo version I have is machined so well that the corners are almost knife sharp! Magnetic grill is a nice touch but for testing, I removed it as you see. Not much to the back panel as to show that it is a sealed design:
View attachment 300871

Seems like the company name used to be Vera-Fi.

If you are new to my testing, please watch this tutorial on understanding speaker measurements.

XSA Labs Vanguard Speaker Measurements
Let's start with our usual anechoic frequency response measurements created by Klippel Near-field Scanner:
View attachment 300872
Well, this is not what we want to see! We have massive and broad peaking of the response from about 700 to 3 kHz and a couple of notches. There seems to be no attempt at creating a neutral response. Sensitivity drops to a very low 80 dB or even lower at some frequencies. Company post a DIY style measurement in a forum showing something different:
View attachment 300873

While their is also pretty bad it shows that using the wrong protocol for testing can hide actual design issues.

Simple near-field driver measurements show the core problem:

View attachment 300874

Notice the wide overlap between the tweeter and woofer. The two add up to create the "hill" we see in the anechoic frequency response measurement.

The mid-woofer is close enough in size to the tweeter which makes for decent directivity. That makes it easier to EQ but without, it causes off-axis to be just as bad as on-axis:
View attachment 300932
This makes predicted-in-room response decidedly poor:
View attachment 300876

Allowing the woofer to play too high seems to be the reason for higher distortion in that region:
View attachment 300877

View attachment 300878


Digging into directivity again, we see confirmation of what I stated above:
View attachment 300879

If we didn't have that notch between 4 and 5 kHz, this would have been a pretty good response. For that reason though, you better toe in the speaker toward you:
View attachment 300880

Vertically we have to stay at tweeter axis as is typically the case with 2-way traditional designs like this:
View attachment 300881

Company spec of 8 ohm impedance seems right for the lowest it gets:
View attachment 300882

Finally, here is our waterfall and step responses:
View attachment 300883

View attachment 300884

XSA Labs Vanguard Listening Tests and Equalization
I usually start with my female vocals but this time, I had one of my bass tracks queued up so hit play on that. I was surprised how little volume I was getting out of the speaker even after I cranked it way up (I have a 400 watt Mark Levinson amplifier driving it). I stopped and switched to my regular female vocal and volume shot way up. The sound was inoffensive and one could think this is good sound (if you didn't play anything with bass in it). Recalling the poor frequency response measurements, I brought out the EQ and first filter I dialed in was for the broad excess energy in lower treble:

View attachment 300886

That pulled the female vocals back, restoring a neutral response although I could see someone just listening to female vocals liking the stock sound maybe better. To my ears, there was not much warmth so I pulled up the response in upper bass as measurements showed. That was nice but then I had some boominess on a track. I have a room mode around 105 Hz and that is where this speaker peaks as well so I pulled that down some. Now the sound was much more balanced but still not to my liking. Many of my reference tracks didn't sound good.

I remembered that in the past I have had good luck in getting more bass out of sealed cabinets by pulling up the bass response a bit. If the drivers handle it without getting distortion, the effect is like having a port but without its drawback of transmitting cabinet resonances. Indeed that was the case here with that 70 Hz filter. For the first time, I could actually get some physical sensation of bass out of the floor.

With all of these filters in place, the sound was good enough for about half of my tracks. Distortion was not audible unless I really pushed the speaker where it started to creating some ticking sounds. It was plenty loud just playing one speaker so I think you will be fine with a pair.

Conclusions
Company website is high on review quotes and low on information. But reading around it seems that the company thinks they have designed something akin to BBC LS3/5A (???). Whatever the aim was, it is a complete miss in practice. They need to up their game in measurements and spend more time performing proper AB testing using other speakers blind, or as i do, with electronic EQ. Just listening to one speaker, especially one's own design, is likely to lead you into a ditch and that is precisely what we have. So while I appreciate the woodworking here, the speaker itself is a miss on grounds of simple principles of designing a neutral speaker that appeals to many audiophiles.

Fortunately the rather similar sizes of the drivers makes directivity decent enough that simple EQ filtering makes a dramatic difference.

I cannot recommend the XSA Labs Vanguard speaker.

----------
As always, questions, comments, recommendations, etc. are welcome.

Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

I have only heard the LS3/A's once on a friends set-up.,Vintage Leak amplifiers and a LINN Turntable. You can't get more vintage then that. I really enjoyed the nearfield experience and would love a pair. Let's not forget they were orinally made to fit in vans along with other audio equipement! This is the UK in the 1970's and 80's. I understand the BBC spent millions of taxpayers money developing this loudspeaker. The license fee paid to the BBC has probably covered costs since 1972! Surfing on the popularity of a loudspeaker design, making it cheaper and avoiding the license fee means only one thing. It can't sound the same, never will whatever you do to it and res

 

[badspeakerdesigner]

So do I. I am very much intrigued how off-spec drivers caused the 2-octave overlap in the crossover.

PS. On this review it took 111 posts until the "it was a broken unit" comment to arrive

The thing I noticed that caused me to contact PE to get a new dc130, was that one was extending far further than the other with the same 2500hz low pass. The bad one was extending all the way to 4-5khz.

 

[sarumbear]

The thing I noticed that caused me to contact PE to get a new dc130, was that one was extending far further than the other with the same 2500hz low pass. The bad one was extending all the way to 4-5khz.

Maybe the driver impedance is so different at that range that the LP filter was not working as designed.

 

[sarumbear]

I understand the BBC spent millions of taxpayers money developing this loudspeaker.

Here is the original BBC paper. It will be nice for members to at least have a look at the frequency response of the design and then ask themselves where is that infamous BBC dip...

http://downloads.bbc.co.uk/rd/pubs/reports/1976-29.pdf

Using modern drivers and measuring methods this manufacturer failed to copy the FIFTY year design.

 

[CDMC]

Turns out the design of the speaker was actually inspired by the Rogers LS3/5a https://v2.stereotimes.com/post/xsa-labs-vanguard-bookshelf-loudspeakers-by-greg-voth Voodoo style engineering.

Well they do have a lot in common, both are rectangular boxes and have two drivers in each box.

 

[tmtomh]

I understand the skepticism about the "off spec drivers"/"broken sample" claim here, and based on the review and the insights shared in this discussion, I share that skepticism.

Fortunately, the guiding principle of this forum allows us to put aside our skepticism and simply wait for the maker to ship Amir another sample, and for him to test it. The measurements will tell the tale.