CLIO Clinic

[Ian Mackenzie]

Rob,

Great post.

That is exactly what I did to analyse the 4345 some time back...

When measuring in the near field (less than 2 metres), vertical height will display not only driver vertical polar angle but also crossover and driver interaction. Therefore I tend to take a series of measurements at different heights often with one driver turn on at a time then all on (others padded back) and then an overall measurement further away to interprete what is happening. One measurement tells you very little.

Unfortunately most of us don't have the luxury of a wide open space or a vacant roof top like JBL.

I've been doing Pulse and MLS measurements for a while (about 7 yrs ) at home and obtaining useful data really is an art form when attempted in the home environment.(its not always practical to relocate a 200+lb loudspeaker)

Aside from mic position I found the most intuitive measurments by management of the main room reflections which are often the floor and ceiling.

Placing pillows or bedding covers on the floor half way between the speaker and the mic is useful in contolling floor reflections. Acoustic tiles above the system suspended from a light fitting is also usful on controlling ceiling refections which might otherwise be interpreted as a peak or notch on the amplitude response.

Some accoustic tiles at the sides can also assist in removing a lot of junk in the un smoothed curves.

[Mr. Widget]

Interestingly, the basic results are the same above 1 kHz for each setup. Your space must be nicely damped, as reflections aren't majorly apparent.

Windowing cleans it up below that, mostly, but 7.7ms window renders everything below 150 Hz invalid, no?

Gotta use different methodology to get useful results below that, anyway.

D'Appolito says measure on tweeter axis. It's now clear why.

Big differences in the 800 Hz region between your two test heights. I'm guessing you're crossed over somewhere near there?

Zilch,

Rob is using 1/3 octave smoothing. That removes most of the detail and some of the reflection information.

His time window is 4.5 and 4.7ms. If you read the Stereophile articles that I mentioned in an earlier post and Rob tracked down:

http://stereophile.com/features/105kh/

http://stereophile.com/features/405time/

They explain that it isn't all that simple to determine the accuracy of the lower frequencies from a time windowed measurement.

Unless Rob used a larger room than the 7' plus ceiling height of his listening room, he should be windowing closer to 3ms anyway, but that aside... I would suggest that these are accurate to about 1KHz... below that we are seeing a lot of room modes and early reflections.

Widget

[Mr. Widget]

Placing pillows or bedding covers on the floor half way between the speaker and the mic is useful in contolling floor reflections.

DEQX recommends a two foot thick pile of blankets or whatever on the floor to help.

I use a double bed sized piece of egg crate foam and some sonex panels.. they have the greatest effect at higher frequencies. It is best to have all objects as far away as possible. That includes the mic. I have to move the couch 4-5 feet away from the mic as well. If there is anything near the mic that is closer to it than the speaker's first reflection, it will affect the measurement.

Some accoustic tiles at the sides can also assist in removing a lot of junk in the un smoothed curves.

Yeah... they will sort of help. It is way better to have nothing near by, but if that is simply not possible absorptive panels will help somewhat.


Widget

[Ian Mackenzie]

Yeah,

Best to do it when the wife or who ever is not home.

I find 2 am in the morning a great time for measurements.

Ian

[Robh3606]

Here is the same seated height curve with a 3ms window and 1/12 smoothin. I have attached the JBL 4345 plot as well. Looks close above 1K with the expected comb filtering in the crossover region between the HF and UHF. See that nasty notch at 3K??? Quess what it is??? Hint is it's not stock on a real 4344 and it really surprised me.

Rob

[Robh3606]

Hello Ian

I have to learn how to make some close in measurements. I tried a couple off the 2122 and it was as expected very smooth overall. I definately have to work on getting the reflections down. I am measuring right over an uncovered tile floor. May as well have a mirror down there.

Rob

[Mr. Widget]

Hey Rob,

I still haven't loaded version 7 yet... but have you found out how we can do spatially averaged plots? It may be done by simply using the tool that you use to add separate driver measurements together.. not sure. The Stereophile measurements made by John Atkinson are typically an average of several plots in an arc about the tweeter. I think these will remove some of the comb filtering that we see but don't really seam to hear.

I have printed out the version 7 manual, but have not yet read it.

Widget

[Robh3606]

Hello Ian

It's the H94 Serpentine Lense. Take off the lense and it dissapears. I quess the 2308 and the H94 are not interchangeable. Thanks for reposting your measurements.

Rob

[Lakanta]

Hi Zilch,
I still have Clio 3.2. Pure DOS and ISA-Card. The Hardware/Software are mounted in a old Pentium for 0,00 U$. I used the software/Hardware for years. It is quick and practical. But.... the quality of the measurements! Poor resolution and bad interpolation/Smoothing. Repeating the measurement and you have every time a new Plot. That was one of the reasons I switch to Bruel & Kjaer.
But here a great tip:
Get the book 'Testing loudspeakers by Joseph D'appolito'! He describes how to use CLIO correct and compare it with MLSSA and analog Systems. Perhaps heavy to read but quite a excellent reference!

[Ian Mackenzie]

I am sure that us a useful reference and I think some members may it.
Do you have the ISP code?

Given the sensitivity of these measurements to the room environment I often wonder how serious should the diy/ speaker builder be about measurements taken in the home!

[Robh3606]

Hello Lakanta

"Repeating the measurement and you have every time a new Plot."

Is that within the same session or after you set-up again?? An issue I have is I have to break my set-up down and was worried about repeatabillity. When I had it set-up it seemed to give very conssistant measurements while I didn't change the set-up. As soon as I changed the set-up the measurents changed but were again consistant for that set-up. I have the book and thanks for recommending it. Any tips on the methodology you use to get consistant measurements. I figured if I mapped out specific locations in the room for both the UUT and the Microphone I was at least heading in the right direction. I can see that getting repeatable measurements with this compared to an RTA is night and day. This is much more sensitive and has much better resolution.

Hello Ian

"Given the sensitivity of these measurements to the room environment I often wonder how serious should the diy/ speaker builder be about measurements taken in the home!"

Let you know in 6 months if I think it was worth the investment or not.

Some additional reading:

http://www.stereophile.com/features/99/

http://www.stereophile.com/features/100/

http://www.stereophile.com/features/103/


Rob

[4313B]

I figured if I mapped out specific locations in the room for both the UUT and the Microphone I was at least heading in the right direction.

Let you know in 6 months if I think it was worth the investment or not.

Exactly!

[Mr. Widget]

It is quick and practical. But.... the quality of the measurements! Poor resolution and bad interpolation/Smoothing. Repeating the measurement and you have every time a new Plot.

I guess they have improved it or perhaps yours is defective. I am regularly happily surprised at how consistent my CLIO rig is. Occasionally it seems buggy and will give me erroneous data, but typically multiple measurements result in plots that are exactly the same... to within the resolution of my monitor.

I am not sure how others are using theirs... I have mine set to average 5 series of MLS pulses... this is fairly quick and removes most of the anomalous readings you might get from intermittent background noise.

Let you know in 6 months if I think it was worth the investment or not.

I've had mine for just over a year and I couldn't be more pleased. Is it worth $700 for Lite or $1700 for the whole package??? I guess it depends on what your application is. If you are simply tuning your room that is a tough question... for me, I use it quite often as I design and redesign and redesign and redesign..... various speaker systems.


Widget

[Zilch]

I finished first reading of the D'Appolito book last week. It clearly documents the difficulties with different methodologies, and means to overcome or minimize them. The math can be intimidating, but is, for the most part, gratuitous, as I have discovered is common in much of the literature in this field. It's not a "difficult" book; there should be no problem reading and understanding what's going on there for anyone interested in the subject, though I did doze off a couple of times.

An exception is the chapter on FFT, which is pretty heavy without an engineering/math background. Fortunately, it's non-essential to the major subject matter.

In the frequency domain testing sections, he decries the expense and lack of resolution of analog RTA's (1998) and then simulates an "improved" one using a 1/6 octave DIY warble-tone system for many of the tests. Today's RTA's are inexpensive, FFT digital, and 1/6 octave resolution, so much of that material is directly applicable for those using a contemporary RTA such as the Behringer DEQ2496, and, though I'm not familiar with them, some of the PC software based systems as well, perhaps.

That's about the center third of the book, and anyone using an RTA seriously would do well to read up. There's much more here than MLS testing....

[Working up my automated indexing rotary table RIGHT now for polar measurements of horn/driver combinations.... ]

[Ian Mackenzie]

Yep,

I try and always measure from the same co ordinates..I put masking tape on the floor and leave it....

For an old "diy hacka" like me as long as I can reasonably confirm what was / is meant to be happening with a diy JBL clone system like the 4430 or a 4345 I am a happy camper. I gave up attempting to be a designer "Wannabee" when I first build and heard a JBL network made the right way..the 3134/3135 and 3145 new equivalent.

I am sure Clio is a great tool .

Has anyone ventured to plug the output data of Clio into a crossover Optimiser yet? (Leap, SoundEasy)

Ian