Kii Three - my impressions and pro reviews

[mitchco]

Wow, I am curious at what sound pressure level this is being listened to at the listening position? Do either of you have a SPL meter or can use a SPL meter app on your phone as a rough indicator?

 

I only managed to get the limiter to kick in on some very bass heavy tracks approaching maximum output. That's perceived as 4 times louder than reference level of 83 db SPL...

 

So what are folks measuring SPL wise?

[mitchco]

Here is the frequency response of both the Kii THREE and the D&D 8c overlaid as measured in my room. Both were tuned to produce a downward tilting response from 20 Hz to -10 dB at 20 kHz as per Harman's research.

 

 

This was achieved using the onboard controls only. The 8c is a bit smoother in the bottom end due to the additional PEQ's that the THREE does not have, but it in a double blind comparison, I doubt I could pick the two apart from a tonal perspective. As others have mentioned, the THREE sounds "dryer" than the 8c is about the only way I can describe any difference.

 

The 8c's treble required next to no adjustment to match Harman's research. However, the THREE out of the box required a -4 dB shelf at 3 kHz to match. So, out of the box, without any adjustments, the Kii's may sound overly bright.

 

Of course, both speakers fr can be made ruler flat using room eq like Audiolense or Acourate. Personally, I would just use partial correction from 500 Hz and below, as both have a wonderfully smooth response above 500 Hz. The cardioid design of both speakers really make a difference in the low end response where you don't get the crazy 25 dB of total up and down swing due to room modes, like in my room. So to get that smooth low end response is a real testament on how well both speakers deal with room modes with their cardioid design and on-board boundary controls.

 

Given it's cardioid design and  pretty smooth low end response, it will be interesting to see if any kind of FIR based room eq makes it's way into these products. The limiting factor is the on-board computers with the DSP chips that limit the number of FIR filter taps it takes to smooth out the low frequencies over time. Most hardware based DSP is limited to around 6000 taps per channel, whereas on a PC with Audiolense or Acourate one can use 65,536 or even 131,072 taps per channel for total low end control over time.

 

 

[mitchco]

Nothing in the REW files as I am talking about the direct sound. I can only attribute the difference to the overall electronic design and perhaps, "unique to the implementation used in the Kii THREE is a combined voltage/current control loop that goes beyond merely a better amp – it actively improves the distortion performance of the drive units which contributes significantly to the extreme resolution of the speaker."

 

Don't know exactly what that means 🙂, but one would need an anechoic chamber and associated mic, preamp, ADC that has low enough distortion to be able to measure properly to confirm.

 

[mitchco]

Dryer sound - less distortion is my best guess, given the electronics quote. That's what is fundamentally different compared to the 8c. "According to Putzeys, distortion of loudspeaker drivers can be optimized by varying the amplifier’s output resistance for different frequency ranges. This requires an amplifier that is specifically tuned to the parameters of the loudspeaker driver. An amplifier like this can never be a commercial good-for-all piece of equipment. They have to be tailor made, and this is what was done for the Kii Three." From Audioxpress's excellent review

 

[mitchco]

Great question! I don't have an answer as I liked both speakers a great deal and could easily live with both. One part of me believes the magic Bruno has achieved as the most neutral speaker out there. i.e. no colour at all - truly neutral. The other part of me is trying to reconcile if I like that sound or not. I think I do. It reminds me of Rythmik subs and their Direct Servo technology. Very understated and dry sounding as their distortion measurements confirm on https://www.data-bass.com versus other subs. but that is what I have in my system today. What can I say 🙂

[mitchco]

1 hour ago, Bernstein said:

Thanks a lot! 
And also thanks for your cool tracks you use for testing. If you have a list of them, it would be nice to share them. We have a joint KII Playlist curated by @firedog on Tidal, so we can add some songs to the collection

 

A partial list here: https://audiophilestyle.com/ca/reviews/nad-viso-hp50-with-roomfeel-headphone-review-r720/#subjective

 

[mitchco]

9 hours ago, TheStupidOne said:

Regarding subs;

 

I'm currently using Minidsp 4x10hd for connections, volume control, timing, eq and crossover. In my case I don't use the subs as subs, I use them as bass speakers crossed over at 200 hz. This "necessitates" running them as stereo and properly time aligned with the corresponding speaker.

 

All that said, I think I would change the 4x10hd for something else in order to better the S/N ratio and to get a bit more PEQ bands to fiddle with, maybe the new SHD studio? Or perhaps OpenDrc so I can load Audiolense filters to it?

@mitchco - any thoughts/suggestions here? I would only need digital outputs.

 

 

 

Hello, @TheStupidOne perhaps a little diagram or words to describe your end to end signal path with type of connections. If using computer with Audiolense, Audiolense can also be used for digital XO duties as well (highly recommended). What convolution engine are you using? The issue with OpenDRC and the like, is the limited number of FIR filter taps. Typically around 6000 taps per channel, whereas on the PC, 65,536 taps (or even double) are available and required if you want deep bass room correction. Have you asked Bernt if Audiolense will take the 90ms of delay into consideration if using Audiolense XO with subs? Also, curious about the 200 Hz XO point. The Kii THREE's cardioid capabilities go right down to 80 Hz, which would seem more of natural XO point for the subs, so you can still take advantage of the cardioid response of the THREE's...

[mitchco]

19 hours ago, TheStupidOne said:

@mitchco; Thank you for replying!

I guess I'm asking how you would set up a signal chain if you were me and used the speakers as tv-speakers 90 % of the time?

 

If that is the case, and you are happy with the SQ, I would simply leave well enough alone. If you wanted to experiment, I would lower lower the XO to the subs to 80 Hz or below. The suggestion from Kii is a good one. The subs should augment the Kii's and not replace their cardioid response I would find the -3 dB point at the LP with just the Kii's and then use that point as the XO point to the subs. That way you are maximizing the response of both Kii's and subs. Fyi, that' what I did with my rig (not Kii's but finding the -3 dB point on my mains) and XO to my subs are at 45 Hz. 

[mitchco]

How are you measuring distortion? At the listening position? If so, then unfortunately, the room dominates below Schroeder and likely not getting real distortion numbers... Distortion of speakers needs to be measured at 1 meter distance and ideally outdoors using a ground plane measurement. Also, does the mic come with a distortion spec? There are only a handful of mics that spec distortion. Earthworks, DPA, B&K and http://www.isemcon.com/datasheets/EMX7150-US-r04.pdf  for example. Most measurement mics are good at measuring frequency response, but not so good measuring distortion (including the mic I am using right now).

 

The Kii's have fantastic directivity down to 80 Hz and designed to avoid SBIR and it is worth the time spending with the measurement mic to take multiple sweeps while adjusting boundary eq to get the smoothest response from the THREE's first in your room (maybe you have already done that). Then bring in the subs. Yah, it could be that the -3 dB point in your room is 30 Hz, but subs are supposed to, you know... sub 🙂 If you still want to offload, then I would start at 80 Hz and work your way down from there...

[mitchco]

@TheStupidOne Excellent! I bet it does sound great! 😉

 

1 hour ago, TheStupidOne said:

It should be noted that my subs are not exactly subs and they are placed right next to the Kiis

 

Ah, I missed this, I get what you are doing now.

 

Are you using TTD in Audiolense?

 

If you are still up for an experiment, pick either the 80 Hz or 200 Hz XO and like in the graphs above, at the same SPL, move the mic from the LP a foot closer (or 2 or 3 ft whatever) towards the side you are measuring and take another measurement. You may want to take 2 to 4 measurements towards the speakers before you are like 3 feet in front of the Kii and sub.

 

If the shape of the distortion curve stays the same, but simply goes up in level, one can conclude you are measuring the speakers (and microphones) distortion. If the distortion changes with each mic position, then it is the room (sometimes combo standing waves and noise floor). Have you plotted the rooms noise floor? Also, have a quick look at what JohnM wrote and check out the link to the help file: https://www.avnirvana.com/threads/noise-floor-seems-too-high.1732/#post-15970

 

Congrats again. good job!

[mitchco]

Hi @aps no, I would always recommend a high pass or better yet, a linear phase digital XO between sub(s) and mains.  A linear phase XO sums properly both in the frequency and time domain. So yes, with a multi-channel soundcard or DAC, AL or Acourate digital XO would be used to manage low/high pass to subs and Kii's. Similar to what is done in this article: https://audiophilestyle.com/ca/ca-academy/ integrating-subwoofers-with-stereo-mains-using-audiolense-r712/

 

The trick is to find the best XO point and this is where the room comes into play. In my case linked above, my mains have good output to 40 Hz or a bit below (i.e. the -3 dB point in the room). Using tools like https://amcoustics.com/tools/amroc I have a large room mode, right at 40 Hz. So to offset that, I chose 45 Hz as being inbetween room modes for my XO point.

 

In the case of the Kii's and @TheStupidOne setup, is using additional woofers to supplement the Kii's. Turned out good! There are no hard and fast rules. However, if using a traditional "sub(s)" with the Kii's, I would take Kii's advice (and mine) of finding the -3 dB point in-room of the mains (whatever they are) and cross there.

[mitchco]

11 hours ago, Bernstein said:

@mitchco

Sorry to bother you, but your a room expert  

What is your proposal and experience for the KII Setup (Height, rake, toe in etc). 

 

No bother at all. I don't know your setup or constraints, but I am a simple guy and follow the same setup as most studios and mastering facilities use to mix and master the music we are listening to. Setup is an equilateral triangle with each speaker toed on in axis to that equilateral triangle. i.e. the left tweeter should be pointing at your left ear, and right tweeter, aiming at right ear, coming to a point just behind ones head. This is how most music is mixed and mastered.

 

Height wise, the tweeter should be at least ear height and ideally mid point between the midrange and the tweeter. No rake as in "exact" mode means the tweeter and midrange (and bass drivers) are time aligned/phase aligned as per the measurement in my review, with 0 degree rake. Now if that is not possible, it is better to have the speakers elevated and pointing down as a compromise.

 

Using a tape measure or laser distance measure, I would ensure exact distances from the rear of the speaker to the front wall and each of the side walls, down to a 1/4 inch tolerance or better if you can. One wants as symmetrical of a setup as possible to maintain the proper stereo image. A 20 kHz wavelength is just over 1/2". It is a pain to do this level of fussing, but it pays off in the image department.

 

Next, adjust boundary and controls to suit neutral or to taste. Neutral is roughly 20 Hz to -10 dB at 20 kHz if 7 to 9 ft away from the speakers. I found the THREE's too bright for my ears out of the box and used a -4 dB shelf at 3 kHz to match the neutral target response. Personally, I would use a measurement microphone and REW to set the boundary and contour controls. Once zeroed in, I would "bracket" on either side of the settings to ensure that is as good as it gets.

 

Finally, and optionally, either use REW's suggested PEQ's or Audiolense or Acourate to tame any further room modes below 500 Hz.

 

The Kii THREE's are one of the most neutral and accurate speakers I have heard. It is worth the effort to got through all of the steps. I get everyone has constraints to their setup, so my steps above are ideally speaking.

 

Hope that helps and you are enjoying the music out of these wonderfully accurate sound reproducers!

[mitchco]

12 hours ago, Bernstein said:

@mitchco Does the soundstage get smaller in such a triangular setup with tweeters aligned to the ears? 

 

Not sure how to answer that as I am following industry guidelines for monitoring music production. A couple of examples are: EBU Tech 3276-1998 Listening conditions for sound programme material

and 

Recommendation ITU-R BS.1116-3 (02/2015) Methods for the subjective assessment of small impairments in audio systems.

 

There are others, but other than the frequency response target and that the monitors are not elevated to accommodate a control room window to the studio, this is how most music is recorded, mixed and mastered, using these guidelines for room setup. The guidelines should be updated based on the work of Sean Olive and Floyd Toole have done showing peoples preference for a more neutral response as per one of my posts above.

 

So if the goal is to reproduce as closely as possible to what was heard in the studio (or venue or concert hall or whatever) by the artists, producers and engineers, I would follow the guidelines. If not, then anything goes 🙂 There is no right or wrong.

 

Every system I have set up or worked on, whether pro or home, I use the equilateral triangle and follow the guidelines. The soundstage always sounds right to me, with a solid phantom center. But that's my preference. I can't listen to speakers that point straight ahead or the triangle is too wide or narrow or the LP is too close or too far from the speakers. In all of those cases, the soundstage does not sound right to me. But again, that's my personal preference to align with industry guidelines. For others, likely different and in some case very different, but I would not argue with anyone's preference if it is different than mine.

 

I don't know the details of your setup to say one way or another, other than to say there is an industry "best practice" that I follow,. It may not be your preference.