Unfortunately its hard to judge from the S&R isobars how well the angle measurements match the manufacturer ones.While this is not wrong, the 944K isn't their flagship, and since the SAR measurements match really well those given for the 944K:
I guess we can trust those given for the 901K:
which are definitely cleaner. They don't seem to have the same design as well, where one has a plate in front of a low crossing woofer and the other has the midrange and tweeter stuck to their small siblings' grill.
Thank you for mentioning this. With regards to distortion, transducers produce about an order of magnitude more distortion than the amplifier. If you want to lower distortion, spend money on the transducers first. At all of our price points, we carefully optimize transducers for linearity and symmetry using Finite Element Analysis. The associated reduction in 2nd order harmonic distortion causes 3rd to dominate. We are also using a 2 layer coil on the woofer with an oversized magnet, so that helps a little, and flux stabilization rings both above and below the top plate on the midrange. At a higher price point, we would be adding flux stabilization (and a few other features) to the HF transducer.Why does nobody comment on the prevalence of the third harmonic?
The cheap class D amps... I can understand that a second dominant harmonic is preferred, within reasonable values, or clean profile, but that of the third dominant harmonic...
http://www.sengpielaudio.com/calculator-thd.htm
Well, this is the hairiest situation, where you don't know who made an error - I don't think either is lying. Since even the on-axis is different (e.g. at 1.5k and 7k) while SAR matches perfectly for the 944K, I'm inclined to think that this third-party tester might be wrong in some way.
One possible reason for the differences could be the measuring distance, Hifi-Selbstbau measured them at one meter, while the Geithain measurements don't mention a distance and a higher distance (which is sensible for larger speakers) often smoothens the curves a bit. Nevertheless the vertical directivity problem of pseudocoaxial speakers like the Geithain remains, but our ears are luckily not so sensitive in vertical deviations.
Coincidence? No, I think that's because the speaker just doesn't measure that well in other listening windows. But I'm a cynical bastard!
Class D amplifiers are going to be <0.1% distortion until they reach their limit and start to clip. Transducers are around 10x more. Also, the characteristic of the distortion curve is different. The amplifier is very clean and abruptly transitions to distortion at clipping, whereas the transducer starts with a modicum of distortion even at low level and gradually increases. Also, the distortion coming from the amplifier is relatively uniform across the frequency range, Really, the transducers dominate harmonic distortion until the amplifier clips.Then is not the class D amp the guilty of H3 dominant. The key is how much total distortion in anechoic.
https://www.kaliaudio.com/lone-pine-studio-monitors
Kali LP-6
Kali IN-8 without ghaphs
https://www.kaliaudio.com/independence
Thank You. Now you have an accurate review and metrics that is representative of the product that is offered for sale.So I finally braved the cold garage and measured my purchased sample of Kali IN-8. Here are the results in a more extensive manner per our later tests.
Let's get the most important message out:
View attachment 48334
Yup. The problem with anemic bass is gone. This makes sense as Charles reported that the tested sample had a bad woofer. Likely that lowered its output capability. Sans the cancellation around 8 to 10 kHz, this is a very smooth response. It shows up just as well in predicted in-room response:
View attachment 48335
I ran a few tests at different output level to see if there is any bass compression and there was none:
View attachment 48336
Tested level was third from the top. At the highest level, the IN-8 was quite unhappy (audibly) in attempting to reproduce the very low frequencies but even then, the in-band response was not impacted.
Unfortunately I did not run the levels in the ascending order so the distortion graphs are for lowest output level. Here they are in percentages:
View attachment 48338
Advanced Measurements
View attachment 48340
View attachment 48339
View attachment 48341
View attachment 48342
View attachment 48343
Eye-candy Speaker Measurements
I drilled down into the soundfield (all up wave front from the speaker) at crossover frequencies of the woofer to midrange and midrange to tweeter:
View attachment 48344
The microphone is at the tweeter level so naturally the loudest sound is coming from the woofer so below the "nref" axis. We have one unified speaker instead of two distinct ones playing their own tunes which is what we want to see.
Same story repeats in the transition between the midrange and tweeter:
View attachment 48345
So boring in a good way.
Measurement Accuracy
Here is the polar plot for the highest measured frequency of 20 kHz:
View attachment 48346
This produces the most elongated measurement allowing us to see if it peaks at 0 degree. And it essentially does. I am not taking a protractor and poking a hole in the tweeter to align it with the microphone.
Conclusion
Seems like the mystery is resolved. The Kali IN-8 is indeed a well-designed speaker when it is not broken.
Now someone needs to buy this speaker from me so that I can use the money to eat lunch tomorrow. No, it is not at a discount. It is the only Kali IN-8 with $100,000 measurement data!
I will link the review post to this one.
A conversation based on data... How refreshing!
