Rane ME30B Equalizer Use
After many attempts at equalizing my system with Audyssey MultEQ XT32 using numerous different Audyssey mic positions, plus the Audyssey Pro kit and edit functions, I could not sufficiently reduce the equalized over bright/hard upper mid-range to a level that I was comfortable with. That was despite my frequency response being relatively flat, see below. I even tried various versions of the mid-range “BBC dip” that are provided by the Pro Kit, but while those curves helped a little they were not what was required.
Typical Front Genelec 1038 Plus Sub Frequency Response (Left Hand Side)
The final Audyssey equalization curve may be switched between flat or one of several HF roll-off curves (as shown above). The rolloff selected during the EQ process cannot be adjusted later as it is created when the results are saved to your receiver/processor.
(I did discover that the strange uptick in the HF roll-off above 15K was due to the selection of a 48K sample rate instead of 44.1K. This is an issue with either my USB interface or laptop. So I reverted back to 44.1 and all is OK again.)
In desperation I decide that I needed to manually EQ the LCR front speakers in order to correct/improve the upper mid-range hardness. After some research and reading, in particular white papers by Harmon, B&K and “Mastering Audio” by Bob Katz, it became apparent that a tilted frequency response with a linear fall of about 10dB (1dB per Octave) from 20 Hz to 20 KHz seemed to be the preferred listening curve. A little depressing after I had struggled so hard to get a flat response.
B&K Preferred Listening Curve (Source see below)
Harmon Preferred Listening Curve Comparison (Source see below)
I therefore opted to add three Rane ME30B 1/3rdoctave analog equalizers into the balanced signal feeds to the left, center and right speakers. A useful feature provided by the ME30 equalizers is its filters design and response curve. It uses Rane’s innovative constant-Q (constant bandwidth) design that significantly reduces the interaction of cut and boost from adjacent filters. This simplifies and speeds up the equalization process. My revised system configuration is shown below.
Rane ME30B Analog EQ Addition
I deliberately restricted the cut and boost range to 6dB even though the ME30 can support 12dB. The resulting MLP equalized LCR responses were measured and created using REW, and still include the AVP Audyssey XT32 EQ.
Preference Frequency Responses. Left (yellow), Centre (dark blue) and Right (light blue)
The most noticeable issue was that the ME30 has a fixed 20Hz high pass filter, see below, that in itself was not an issue to the frequency response of the three speakers as they cross over at 40Hz and the Genelec 1038’s -3dB point is 33Hz. However, its group delay, see below, caused the LF summation between the subs and the LCR speakers to rise starting at 80Hz and peak at 20Hz. I managed to contain this LF rise by using the ME30’s to reduce the LF output of the LCR speakers. This approach is not optimal and I will eventually modify each of the ME30’s to bypass this high pass filter. The overall change in frequency response level required me to increase the LCR AVP trim levels by 2dB each in order to match the remaining satellite speakers and the original LCR SPL levels. I could have achieved this level correction by using the controls on the ME30, but chose not to as they had been set to unity gain with no cut/boost, and these level controls are not easy to accurately reset compared to the AVP trim levels. Moreover you should always put gain as early in the signal chain as possible as it helps maintain the signal to noise ratio (headroom issues aside).
ME30B Frequency Response (no cut/boost)
ME30B Excess Group Delay (no cut/boost)
Initially, this “preference” EQ had the desired effect. The overall balance seemed more pleasing and less fatiguing; the mid-range hardness seemed to have reduced significantly, and there was never a requirement to use Audyssey dynamic equalization (DEQ) again.
However, over the next few months the “hardness” returned, not as severe as before, but nonetheless it was there. After re-measuring the system nothing had changed so I can only surmise that my ears had become acustomed to the masking effect of the new response curve and I was again able to detect what Audyssey was doing to the signal.
I will update this post when I re-equalize the LCR speakers without Audyssey, and hopefully with the 20Hz high pass filters bypassed.
Sources
“Relevant loudspeaker tests in studios in Hi-Fi dealers’ demo rooms in the home etc.
– using 1/3 octave, pink – weighted, random noise”.
Paper presented at the 47thAudio Engineering Society Convention
1974-02-26/29
Copenhagen Denmark
http://www.bksv.com/doc/17-197.pdf
“The Subjective and Objective Evaluation of Room Correction Products”
Sean E. Olive, John Jackson, Allan Devantier, David Hunt, and Sean M. Hess
R&D Group, Harman International
http://seanolive.blogspot.com/2009/11/subjective-and-objective-evaluation-of.html
