Posted by John Roberts on December 03, 2003 at 22:26:39:
In Reply to: Graphic EQ Shootout in San Francisco posted by Bink on December 02, 2003 at 02:09:34:
: Hi, Folks --
: This is to announce a Graphic EQ shootout I'm organizing in San Francisco in two weeks. It will be at Noe Valley Ministry, 1021 Sanchez at 23rd Street on December 17th from noon to 10pm. This is a peak-roofed, bright wooden performance space typical of town hall meeting rooms and small wooden country churches. It's also the home of Larry Kassin's Noe Valley Music Series, an eclectic program that ranges from local rock to jazz to world music.
: Our event is on a Wednesday so I hope a bunch of local sound guys and gals will be able to bust out of their hump day gigs and attend. Any interested sound person is welcome and there's no door charge other than an optional donation or the church. There's also no bar or restaurant on site but you have a wealth of choices on nearby 24th.
: We will be testing EQs electronically with SmaartLive stations and then putting the best EQs to the real world trial up on stage with a vocal mic and a pair of wedges. Yes, this is a test of potential monitor world EQs so the only ones I'm interested in looking at will need to have at least 27 bands per channel and have a fast fader control. Digital and analog are both welcome. If digital, make sure there's a fast way to tweak faders!
: Please RSVP me so I can get an idea of numbers. And if you have an EQ to bring send me its manufacturer and model number so I can make sure we all don't bring the same two models! Tell your coworkers, too. Except that FNG who has to do Wednesday night's gig... ;^)
: I'll need a half dozen SmaartLive rigs run by attendee/participants. If you bring your sandbox with all its accessories you'll really be appreciated.
: Needed, too, is a decent small monitor mixer and biamped wedges with amps and DSP. The house system at Noe Valley isn't conducive to touring- and club-style testing since the wedges are passive. They have an older RFI-hammered Mackie 24 and a single pair of UPA/USW mains that are appropriate for the room. Me, I have no gear other than two computers -- I'll need to have all of YOUR help rounding up the right amount of audio gack. If you can bring some gear please email offline with your thoughts.
: I will welcome any ideas from LABsters regarding which bench tests will help us sort out the best few EQs from the pack. What's your favorite torture test for line level audio gear?
: One of the more interesting questions on my mind concerns the different filter types and how well they each tackle the role of monitor wedge EQ. I'd love to get a strong A/B comparison of a carefully tweaked Proportional Q box versus Constant Q, Interpolating Q or Perfect Q.
: Other than all that technical stuff, this is merely an excuse for a party! It will be a kick to meet a bunch of LABsters and other local sound folks as well as anyone whose tour is swinging through town that day. Also present will be host Ramon Sender who helped found San Francisco Tape Music Center with Morton Subotnick in 1961. He's very interested to see SmaartLive in action and might have a story or two to tell.
: Any comments?
: - Michael 'Bink' Knowles
: www.binkster.net
=========================
I'm glad to hear that you're doing this and wish someone would have done this years ago. I like Marty have a slight blood pressure problem when pondering the underserved poor reputation of some brands/models and equally undeserved high reputation of others.
You have already received several good general and pretty obvious specific suggestions, namely actual frequency response of boost/cut filters, and undesirable interactions between filter bands.
Perhaps I can offer a few less obvious things to look for. From my experience with the design of these, noise and distortion is not a huge issue if the better circuit topologies are used. The multiple band pass filters are effectively bandpassing their own noise. Only the main internal summing amplifier is operating at noise gain greater than a few times and most competent designs use a low noise higher performance part in that socket. A second design issue is hum. While all of the band pass filters do not pass hum frequencies their layout is usually pretty spread out and if not careful in designs using an internal transformer they can contribute unexpected hum to the noise floor. Depending how this hum gets introduced into the circuits it can be boosted by HF faders.
Perhaps a useful piece of test equipment is a good old fashioned distortion analyzer (essentially a glorified notch filter) where you can listen to the product output. While I don't expect distortion to be an issue, being able to listen to the noise floor without the signal masking might be instructive and make subtle differences more glaring.
Another post in this thread questioned the use of companding noise reduction in a graphic equalizer (I agree). Again, a THD analyzer will reveal the true noise floor of a companded design as the gain elements will open up to pass the signal. It also might be instructive to review the transient response and tracking accuracy. Tracking should be pretty good (much better than unlinked NR) but the internal gain elements still have to jump from full off to full on for sudden transients. Widely spaced tone bursts or clicks may reveal any problems.
While not isolated to companded systems, it's also worth looking at how you measure signal handling. You are likely to get different headroom measurements at spot frequencies depending upon your boost/cut settings. Don't assume that boosted will be the worst case.
If you don't have access to distortion analyzer with product output you might be able to make a passable notch filter with parametric EQ and a console with a polarity switch. I'm not sure that fine Mackie you mentioned has one so you might make up a pin 2-3 swap cable.
The two-tone (19 & 20 kHz) IMD was one of my favorites for designing phono preamps due to the huge difference between RIAA eq at those frequencies (1kHz was some 20dB hotter). I actually modified an old SMPTE IMD to work at those freq as back in the day such test equipment was esoteric and expensive. In the case of graphic EQ while probably the opposite of what most people do, you could test with HF bands cut and 1 kHz boosted, but your results may not be very comparable if esq. band centers aren't spot on.
As a side note to Marty's comment about the value of filter center frequency accuracy, even though the FLS EQs that I was involved with used precision components (and AFAIK still do), the FLS circuit detects the level coming from the actual band pass filters so even if they weren't perfect frequency center wise the indication would still be meaningful..
I will continue to think about this and if I can think of any other less obvious things to look at or listen for I'll let you know.
Have fun,
JR