• 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!

Thomann/Swissonic A306 - Review & Measurements by Erin

i would like to add in passive world, similar situation arises +
2.5-way if done right can be cheap and with good sensitivity, many times with impressive performance even,
3-ways do be done right have to cost more than usual budget 2..2.5-ways,
altought i do think room EQ and speakers are by far the biggest game changers in room for sound
 
altought i do think room EQ and speakers are by far the biggest game changers in room for sound
Definitely think so.

1. Compared to many years ago, we now have a lot more information, on speakers, to aid with a purchase decision.

2. DSP has contributed to lowering the cost of speakers, to achieve what would otherwise have needed expensive physical/electronic engineering as you would find in something like zero DSP speakers such as made by PSI.

3. DSP enabled speakers, have made it easier to add on user controls to EQ, within the speaker.

4. DSP outside of the speaker, aks DRC (Digital Room Correction - which should correct some of the speaker anomalies as well as room anomalies) seems to have evolved quite a lot in recent years, as a tool to take things one level further to address any remaining non linearities, in the speaker, as well as counteract the effects of the room. The challenge with this DSP typically done outside the speaker is, I have not seen any proper comparison of the various approaches, at the budget end, e.g IK Multimedia ARC vs Sonarworks vs Acourate vs Audiolense vs the various options that can be achieved via payware tools like REW + RePhase. And on that latter end, there is so much being published by individuals, on blogs or Youtube, but a bit difficult to get one's head around what each person is doing, and what the differences between their approaches are. REW and Rephase and similar tools have tons of options/parameters, so it does get pretty confusing........So much of what occurs in these tools is proprietary, some of it a bit voodoo, and without any real proper scientific confirmation.

Without real comparisons, especially when for most of these tools - the measurement microphone + software (and sometimes an external processor based standalone DSP unit, such as from MiniDSP), comes out at about £400 to £500 as the cost of entry. And you have to spend that much as a guinea pig, not knowing really how effective the results will be, compared with any other method.

From anecdotes Trinnov seems to be highly effective - for those with very hefty budgets.

Hopefully the next few months we'll see much more comparative information, for now most of the info is from the manufacturers of each solution or implementors who are singing praises - and earn a fee for implementation. No real world product comparisons. Might be that these speaker DSP/DRC manufacturers are too afraid to submit their products for a truly independent comparison, lest the truth be revealed, that its all hype without substance!!.
 
I have published my reimplementation to GitHub and it should be ready for use. The FIR filters have not yet been implemented (and I perceived no difference) but that will soon be done.

The project includes detailed labels and comments, which should enable fairly straightforward customization. The project is paginated for clarity. If you are unfamiliar with Sigma Studio, I would advise caution as you are playing with a potentially unprotected tweeter.

1704454690794.png
1704454404287.png
1704454438486.png


With the exception of one small diffraction artifact which should not be corrected with EQ, we a have response of ±1dB from 49Hz to 20KHz, with an F3 of 45Hz.

onaxis.jpg
 
Last edited:
With the exception of one small diffraction artifact which should not be corrected with EQ, we a have response of ±1dB from 49Hz to 20KHz, with an F3 of 45Hz.

294473839-1f1651e7-0fc8-42e6-88dd-72be4b054df4.png
Awesome! PS, bottom image does not seem to be loading.
 
Awesome! PS, bottom image does not seem to be loading.
That's odd. It should be fixed, now.

The chart assumes that the -2dB pad is engaged as seen on Erin's page. I will likely revise the project to incorporate that filter by default, as there's no need for flat to be anything but the default behavior.
 
Last edited:
The chart assumes that the -2dB pad is engaged as seen on Erin's page. I will likely revise the project to incorporate that filter by default, as there's no need for flat to be anything but the default behavior.
Amazing work, thanks!!

Since you will revise the project, would it be possible to extend the low end before rolling it off from the current 50hz to 45hz, it isn't a big difference but why not if the speakers can manage it.
 
Last edited:
Amazing work, thanks!!

Since you will revise the project, would it be possible to extend the low end before rolling it off from the current 50hz to 45hz, it isn't a big difference but why not if the speakers can manage it.
I can certainly do that but the headroom penalty would be significant, as that is far below the port tuning frequency. There is already a mild boost in that region and as you may be aware, a 3dB boost means twice the power draw. We have only 50W to play with.

I will be sharing the corresponding VituixCAD project on the GitHub page shortly, which will enable you to make any desired adjustment yourself while visualising the result.

I may implement a switchable DRC algorithm to extend the F3 to 38-40Hz at low listening levels but we shall see.
 
I can certainly do that but the headroom penalty would be significant, as that is far below the port tuning frequency. There is already a mild boost in that region and as you may be aware, a 3dB boost means twice the power draw. We have only 50W to play with.

I will be sharing the corresponding VituixCAD project on the GitHub page shortly, which will enable you to make any desired adjustment yourself while visualising the result.

I may implement a switchable DRC algorithm to extend the F3 to 38-40Hz at low listening levels but we shall see.
Yes that's something that would make sense mostly at lower volumes, indeed. Probably not a valuable idea on my side.

Most use them with a sub anyway and 50hz is below the x-over frequency one would use. Also, the A306 already goes very low for their size and type of speaker.

Thanks again!
 
Below the tuning frequency of the port, cone and port are out of phase. It's like putting two subwoofers next to each other and inverting the phase on one: the main tone gets cancelled and all you hear is a lovely cacophony of port noise and other distortion - which is kind of neat to test different port designs, but people already figured out how to design ports decades ago.
1704645339567.png


You could delete the port entirely and extend the lower frequency response with a lowshelf filter, but honestly, just get a woofer.
 
Below the tuning frequency of the port, cone and port are out of phase. It's like putting two subwoofers next to each other and inverting the phase on one: the main tone gets cancelled and all you hear is a lovely cacophony of port noise and other distortion - which is kind of neat to test different port designs, but people already figured out how to design ports decades ago.
View attachment 340327

You could delete the port entirely and extend the lower frequency response with a lowshelf filter, but honestly, just get a woofer.
I turned off my dirac calibration and sub management and did a lot of listening. While I don't disagree with you theoretically, in practice I just don't get any "lovely cacophony of port noise and other distortion". Erin's objective test and subjective impression don't support that either.

I think that you are exaggerating by writing "lovely cacophony of port noise and other distortion".

Do you have the A306?
 
any plans with the a305? The supply must be running out and I'm still on the fence D:
 
any plans with the a305? The supply must be running out and I'm still on the fence D:
I have no plans to purchase the A305 at the moment but the A306 should soon become available once more.
 
I have no plans to purchase the A305 at the moment but the A306 should soon become available once more.
They suddenly had 12 or so b-stock 306's on offer yesterday. By now they're gone. Are you saying the claims of this lineup's phaseout may be premature?
 
I have published my reimplementation to GitHub and it should be ready for use. The FIR filters have not yet been implemented (and I perceived no difference) but that will soon be done.

The project includes detailed labels and comments, which should enable fairly straightforward customization. The project is paginated for clarity. If you are unfamiliar with Sigma Studio, I would advise caution as you are playing with a potentially unprotected tweeter.

View attachment 339858View attachment 339856 View attachment 339857

With the exception of one small diffraction artifact which should not be corrected with EQ, we a have response of ±1dB from 49Hz to 20KHz, with an F3 of 45Hz.

View attachment 339878

hi, very good job!

Just a question: are the original measurements from a single speaker or were they averaged between a large number of speakers?
Unfortunately I don't know Swissonic's production tolerances.
If we are doing a great job, but then in reality the data is not consistent with production we are screwed :-(
 
hi, very good job!

Just a question: are the original measurements from a single speaker or were they averaged between a large number of speakers?
Unfortunately I don't know Swissonic's production tolerances.
If we are doing a great job, but then in reality the data is not consistent with production we are screwed :-(
Both the original Thomann corrections and my own are based upon measurements of a single randomly selected unit. Quasi-anechoic measurements of my own pair (both random samples) indicate reasonably narrow sample variation.

Most of my filters are quite broad and do not target high Q variations that are likely to vary between samples.
 
Both the original Thomann corrections and my own are based upon measurements of a single randomly selected unit. [...]
That's not entirely true. I had the luxury of having not one but two engineering samples. :D
But hey, I got two A306 and a Beyerdynamic MM1 at home, and it's a 6" woofer so a NF measurement should be OK up to the end of the jagged part. What mic do you use for measuring?
 
That's not entirely true. I had the luxury of having not one but two engineering samples. :D
But hey, I got two A306 and a Beyerdynamic MM1 at home, and it's a 6" woofer so a NF measurement should be OK up to the end of the jagged part. What mic do you use for measuring?
I use a UMIK-1.
 
Hey everyone. I suspect I have snatched up the last of the "current" stock of the A306, however that does mean that the ones that were available were some B-stock examples. According to Thomann, the equipment should be 100% working and might show traces of use, however I have noticed some buzzing from the woofer of the one unit (especially at lower volumes). I suspect this is voice coil rub, as it all but disappears when pressing on the surround at about 6 o'clock. From this I have 2 questions:

1. Are there any "soft" meathods of addressing the VC rub, such as test tones, burning-in, manipulating the cone etc. which do not involve opening the speaker (or dust cap)?
2. What would be the best way to test whether there are no other severely detrimental effects on the frequency responce of the units? I have a USB mic (but no SPL-meter), could I perform a measurement which would confirm that the speakers are within acceptable limits other that the abovementioned?
3. Given the issue, should I rather exchange these units for A305's or A308's (or otherwise)?

Thanks for any input!

Edit: I have tested by swapping the inputs and the issue remains on the same woofer, and with different sources/amps. I have also rotated the unit 180 degrees to no avail.
 
Back
Top Bottom