Thursday, September 23, 2010

Squeezed Lobes & 14" Spaced Boundary Arrays - High Frequency Localization Comparison

View QuickTime Comparison Movie [8mb .zip]

No progress I know of yet on making AT4022 capsules into flat pressure response type. I came across a polar pattern for free-field small boundaries on the Crown PZM 180 which I compared to the SASS-P layout here. The small, PZM-mounted mic capsules also have less gain and HF response in the center of the field. The negative effects could be more exaggerated with the larger diaphragm mics. If we are able to convert an AT4022 into a flat pressure response mic, we may still want to alter the boundary orientation from that of the SASS-P.

The other approach I'm exploring here is to take stock AT4022/3032 mics and experiment with orienting the effective polar patterns in small boundaries. The chief drawbacks of designs that use AT4022's flush-mounted in bodies with SASS-P dimensions appear to be weak center gain and HF response concentrated in lobes at 10 and 2 o'clock. The positives of the rigs to retain are "airy" spatiality, superb "reach" for distant subjects in open spaces and "rear" imaging that is easily distinguishable from the front imaging. 

New Test Description
This test incorporates two new array designs. One is based on re-orienting boundaries based on the estimated effective polar pattern above 5K Hz of an AT3032's flush-mounted near a baffle in a 3- 5/8" X 4-7/8" boundary.  The other design uses adjustable boundaries with 14- 3/4" spacing, 1.5" setbacks on the outside boundary edges and the close placement of a tapered baffle to the mic capsules. I added a stock SASS-P, Jecklin Disk and PBB2 to the line-up for references. The goal is to improve center gain and  achieve a more symmetrical HF polar pattern while retaining the "airy" spatiality, "reach" and distinguishable "rear" imaging.

Test Methods

I used five speakers from my surround set-up in this configuration and an inexpensive SPL meter to adjust sample playback level from all of the speakers to 60dB.  I selected the squeaky Red Squirrel chatter sample because it has strong concentrations between 4K-8K Hz for the demands of this test. With sound absorption treatment on the floor, ceiling and walls of my studio, the high frequency content should be quite consistent. I stacked the arrays vertically between 4.5' to 6' in height. 

To this point, I have only evaluated the results on studio speakers set apart 60 degrees and angled towards the listener.

The "lobe" analogy proved to be useful for describing the gross polar pattern from the small boundaries with flush-mounted, large diaphragm mics. The lobes were not closely-associated with the 500-4K Hz "lift" range but with response an octave higher as shown as emphasized ranges on this sonogram. As can be seen in the same sonogram, squeezing the lobes together does improve center gain. The PBB2 with kissing, non-tapered foam was among the poorest in HF response at 0 degrees along with the Spaced Boundary Array angled out 12 degrees. The array with Spaced Boundaries Angled-In at 20 degrees looks and sounds quite promising as far as center gain and it has smooth, off-axis HF taper. The Squeezed Lobe Angled-In 12 degrees also localizes well but I prefer that the 22, 40, 77 & 90 degree samples move further away with the Spaced Boundaries Angled-In at 20 degrees. I suspect this will translate well in large open spaces,

Sounds with primary frequency content under 1500 Hz imaged more smoothly for all of experimental arrays but reflections with the studio make it hard to evaluate subtlties.

As Vicki suggested in some of our correspondence, there is strong suggestion that tapering the foam baffle away from the capsule will help HF response in the center. There remains a chance that the baffle should also "kiss" the mic capsule at the bottom creating  greater proximity for the <1500 Hz sound waves passing through the baffle. The addition of the two pieces of denser, rubber-treated carpet between the two HD foam pieces did not seem to have a big impact on baffle transmission but I didn't test for this thoroughly.

With its free air AT3032's, the Jecklin Disk has strikingly better HF response between 5K-9K Hz. A frequency boost in this ranges aids with localization in general and exists in the SASS-P as well. With the HF-rich sample, the Jecklin Disk seems to render the speaker located at 22 degrees more accurately-- at a position closer to 0 degrees. The 22 degree position jumps too far to the right with the other arrays.  The Disk renders space quite differently with a warmer, proximate feeling even to the rear. Curious about this, I made a quick spoken voice clock test comparing the Jecklin Disk to one of the Squeezed Lobe (boundary) arrays from 6 o'clock to 12 o'clock. One comparison [ 4mb .zip]  displays the samples sequentially for both arrays and the other comparison alternates clock positions from the arrays [  4mb .zip]  There are some very attractive traits to the Jecklin disc it will be interesting to hear whether these increased bass resonance will interfere with lower mid-range clarity outside.

I'm very curious to hear the reach and and spatial, "airyness" differences under 1500 Hz. I plan to run a three rig comparison between (1) PBB2 with a considerably tapered foam nose (still kissing the capsules at the bottom) (2) Spaced Boundaries Angled-In 20 degrees (3) Jecklin Disk.

If the Spaced Boundaries Angled-In 20 Array proves to be interesting, I'll probably compare it to the Squeezed Lobe Angle-In 12 array outside.

Also remaining to be tested more carefully is the important of foam "kissing" and how softer and harder boundaries affect HS response.

 = = = 

The follow-up testing  may be found here.

No comments: