Cadentia 3 Product Development thread / Audio First Designs

[ThELiZ]

Harry, they look amazing veneered!

I've just had the veneer delivered for my build. Itching to get started! Do you have an updated dispatch ETA for the kit version?

 

[Audiofirstdesigns_Harry]

Harry, they look amazing veneered!

I've just had the veneer delivered for my build. Itching to get started! Do you have an updated dispatch ETA for the kit version?

Thanks!

I aim to ship all the first batch order in the last week of this month, or at least before 2/12 prior to my trip to Hong Kong.

 

[CJH]

FYI, Audio First Designs has collaborated with @pierre and the full set of data is available here:

Measurements for speaker Audio First Design Cadentia 3

www.spinorama.org

Edit: I didn't realize this link was already posted.

Looks like Pierre has the values for impedance and sensitivity switched.
CJH

 

[Curvature]

Looks like Pierre has the values for impedance and sensitivity switched.
CJH

@pierre FYI on the data page.

 

[ThELiZ]

Thanks!

I aim to ship all the first batch order in the last week of this month, or at least before 2/12 prior to my trip to Hong Kong.

That would be ace. A lovely Christmas project for me!

 

[willverduzco]

In for a set. Simultaneously scared and excited to undertake my first DIY build.

Going to be fun putting these head-to-head against my Sointuva AWG, Sierra LX, and a few others to figure out which will end up in which system/room.

 

[willverduzco]

I don't understand enough about multiple regression models to give a confident answer, but it's all in the papers if you're so inclined.

I'd hazard a guess that the Blade's dip at 300Hz may be the culprit:
View attachment 487253

What I can tell you with confidence though is that it is ill-advised (even according to the authors) to read much in to the preference scores.

A 1.0 difference in preference score does not say much and differences smaller than that are basically meaningless.

Keep in mind as well that the preference score only describes frequency response and no other aspect of loudspeaker performance.

It is feasible for a speaker to get a perfect preference score, but only at 50dB SPL and in practice that speaker just breaks apart at higher SPLs, resulting in an awful listening experience.

FYI, calculating the score is incredibly simple, and it's all detailed in the original 2004 paper by Sean Olive.

Preference Score = 12.69 - (2.49*NBD-On Axis) - (2.99*NBD-PIR) - (4.31*log10(low freq extension)) + (2.32*SM-PIR)

All required parameters are visible in the data tab of the measurement pages on Spinorama. Also please note how nothing from a speaker's polar response factors into preference score beyond what can be summarized by PIR. Additionally, the high weighting given to both PIR narrow-band deviation (2.99 vs 2.49 for on-axis) and PIR smoothness indicate just how important PIR is within Sean's model.

For the Cadentia 3 (first party measurement), that's: "NBD: ON 0.238, LW 0.173, SP 0.221, PIR 0.149; SM: SP0.988, PIR0.986; LFQ 0.958, LFX 29Hz"

• NBD (narrow band deviation)
  • ON (on axis) 0.238
  • LW (listening window) 0.173 (not used)
  • SP (sound power) 0.221 (not used)
  • PIR (predicted in-room response) 0.149
• SM (smoothness)
  • SP 0.988 (not used)
  • PIR 0.986
• Low frequency
  • LFQ 0.958 (not used)
  • LFX 29Hz -- take log10 of -6 dB point

For the Blade 2 Meta (eac measurement), that's: "NBD: ON 0.322, LW 0.295, SP 0.275, PIR 0.256; SM: SP0.953, PIR0.895; LFQ 1.098, LFX 26Hz"

• NBD
  • ON (on axis) 0.322
  • LW (listening window) 0.295 (not used)
  • SP (sound power) 0.275 (not used)
  • PIR (predicted in-room response) 0.256
• SM (smoothness)
  • SP 0.953 (not used)
  • PIR 0.895
• Low frequency
  • LFQ 1.098 (not used)
  • LFX 26Hz -- take log10 of -6 dB point

Comparing the two for components relevant to Sean's model:

• Narrow Band Deviation
  • On-Axis: Significant advantage for Cadentia 3
  • PIR: Significant advantage for Cadentia 3
• Low Frequency Extension: Slight advantage for Blade 2 Meta
• Smoothness - PIR: Slight advantage for Cadentia 3

As you may have guessed, the significant FR deviations in the lows and mids are the drivers behind the worse NBD metrics.

Also agree with your note on dynamics. I'd add distortion in there as well as a corollary, even if low amounts were shown to not have a major effect.

 

[willverduzco]

(And all of the above is to further highlight the amazing design work @Audiofirstdesigns_Harry was able to accomplish with this relatively inexpensive set of speakers!)

 

[totti1965]

FYI:
According to Bruno Putzeys, Multitone distortion is one of the most importand measurements (at least in amplifier development). The measurement shows whether a playback device smeares information or reproduces it accurately.

Here a typical o.k. isch Speaker - the 2-way bookshelf Elac Debut 2.0 B5.2, designed by Andrew Jones:

Here the Audio First Designs Cadencia 3:

"Not bad" is the understatement of the year.


Here is Erin´s measurement of the EUR 30.000,--/ pair KEF Blade 2 Meta up to ca. 96 dB:

 

[Marin]

FYI:
According to Bruno Putzeys, Multitone distortion is one of the most importand measurements (at least in amplifier development). The measurement shows whether a playback device smeares information or reproduces it accurately.

Here a typical o.k. isch Speaker - the 2-way bookshelf Elac Debut 2.0 B5.2, designed by Andrew Jones:


View attachment 491061


Here the Audio First Designs Cadencia 3:

View attachment 491058

"Not bad" is the understatement of the year.


Here is Erin´s measurement of the EUR 30.000,--/ pair KEF Blade 2 Meta up to ca. 96 dB:


View attachment 491059

Speaking of apples and pears, let's compare measurements taken by the same reviewer and very important same

equipment.
Also two-way vs. two-way 2,3,1 fight

 

[staticV3]

Speaking of apples and pears, let's compare measurements taken by the same reviewer and very important same View attachment 491649equipment.
Also two-way vs. two-way 2,3,1 fightView attachment 491650

That's what 6x the price will get ya.

 

[willverduzco]

Folks, I think an even larger issue than discrepancies in form factor or 2-way vs 3-way, was comparing a full-range multitone distortion measurement (like what Harry shared in post #15) with band-limited plots, crossed over at 80Hz. Doing this arbitrarily places the Cadentia 3 at a significant disadvantage.

Here is the full-range MD plot of the KEF Blade 2 Meta from Erin's review:

This is exceptionally good for a full-range multitone distortion measurement, but a little less excellent than what was originally shared above.

Now, with that fresh in our minds, let's take another look at the Cadentia 3 -- which Harry mentioned was measured under similar conditions to Erin's testing:

Remarkably, MD performance is even better than the Blade 2 Meta -- and as noted in my previous reply, tonality looks better as well!

Once again, what a truly exceptional design -- and one I cannot wait to build and compare against my other speakers.

 

[totti1965]

Folks, I think an even larger issue than discrepancies in form factor or 2-way vs 3-way, was comparing a full-range multitone distortion measurement (like what Harry shared in post #15) with band-limited plots, crossed over at 80Hz. Doing this arbitrarily places the Cadentia 3 at a significant disadvantage.

Here is the full-range MD plot of the KEF Blade 2 Meta from Erin's review:
View attachment 492843
This is exceptionally good for a full-range multitone distortion measurement, but a little less excellent than what was originally shared above.

Now, with that fresh in our minds, let's take another look at the Cadentia 3 -- which Harry mentioned was measured under similar conditions to Erin's testing:
View attachment 492842
Remarkably, MD performance is even better than the Blade 2 Meta -- and as noted in my previous reply, tonality looks better as well!

Once again, what a truly exceptional design -- and one I cannot wait to build and compare against my other speakers.

Wow!

 

[Audiofirstdesigns_Harry]

From my experience in designing loudspeakers, I personally believe there is a very strong correlation between the IMD performance and some subjective feelings like "Transparency" and "smoothness". It makes sense as IMD is noise and it's inharmonic to the original content.

When comparing the Cadentia 3 and the Fidelia side by side, the first thing that others and I notice is that, although their tonality is very similar, the Cadentia 3 simply sounds cleaner and smoother, and this becomes more obvious at higher SPLs. Back in my time at KEF, I felt the same when comparing the 2-way and the 3-way models too.

By the way, I have made really good progress on producing the first batch (10 sets) of Cadentia 3 orders! They will be shipped this week! And sorry for my delay in email reply recently, I am just completely overwhelmed !

 

[EAXAE]

View attachment 488899

Wow, this photo looks nice. Now I want these speakers for the the looks alone.

Wondering why the baffle is angled, not rounded though. Wouldn't it benefit the aesthetics and diffraction?

 

[Trangbar]

The build tutorial video has now been posted on the Cadentia 3 page:

 

[Audiofirstdesigns_Harry]

The build tutorial video has now been posted on the Cadentia 3 page:

Thank you! I was about to post it here too, but you beat me to it! Appreciate that !

 

[101geo]

Thank you! I was about to post it here too, but you beat me to it! Appreciate that !

Maybe if you spent less time globetrotting...;-)

 

[Audiofirstdesigns_Harry]

Maybe if you spent less time globetrotting...;-)

Hahah indeed, I literally uploaded the video right before my trip to the airport and was writing the newsletter about the video during my flight lol

 

[suttondesign]

I just placed my order for the diy kit.