TACT RCS 2.0 Users Group

One additional suggestion which is a substitute for indpendent measurement. Get a good Mercury Living Presence CD that was recorded in Mono. A stereo recording played in mono won't work because you do't know the actual left right time response. I recommend the Dorati recording of Tchaikovsky's Swan Lake with the Minneapolis Symphony. It doesn't matter if you don't like classical this is a test record. Listen to the last 15 minutes with your eyes closed. If your curves succesfully correct for the room and speaker anomolies, the image should be rock solid between the speakers. If the instruments wander depending on what is playing or how many other instruments are playing, you should check your speaker placement and remeasure. This should work with just about any reasonable target frequency response curve.

I would also recommend that people looking to understand the response of their room try ETF 5.x. Look at http://www.etfacoustic.com/. It is some times simpler (and cheaper) to correct the response this way.

Question 1: I found the TactG1 best suited my taste in terms of the provided target curves. I have a PSB Stratus Gold-i speakers (full range). Other curves would work better for speakers with a more restricted LF response.

Question 2: TactG1 was a good starting place, but what I'm running now is a hybrid of TactG1 and TactF1 (I spliced the HF portion of TactF1 in place of the highend of TactG1), and I tweaked the midrange (500 -1500 Hz) up a bit, with a max rise of about 1 dB at 1 kHz for a bit more presence, and boosted the upper bass / low midrange a a skoche for a tad more warmth and bass oomph. I find the Tact curves just a little too dry (I add ed 0 dB boost at 250 Hz rising to about 2 dB boost at 100 Hz, then smoothly sloping to match target curve at around 50 Hz). I'd be happy to e-mail my target curve to anyone who's interested.

Question 3: Get the house as quiet as possible. I wait until everyone's gone to bed (although St. Bernards snoring in the mudroom carry a long way), turn off the central heat, and wait 'til the refrigerator stops running. Turn up the preamp volume to a LOUD listening level, and use a lot of averaging (I use 20 averages, but 100 would be better -- an effective SNR increase of 7 dB in the measurement).

Question 4: tweak and listen. An experiment I'd like to do is to program a Fletcher-Munson equal loudness curve as a target curve. See, for example,
http://www.sfu.ca/sca/Manuals/ZAAPf/e/equal_loud_cont.html The curve describes the relative power level needed for a sound at any frequency to sound as loud as a
reference tone at 1 kHz. In effective, it's an equalization curve for the "typical" human ear. It shows that at moderate levels (say 80 dB SPL at 1 kHz) significant
boost and buck (a few dB to over 10 dB) of various frequencies is needed for those frequencies to sound as loud as the reference 1 kHz tone. So even though the target
curve is decidely unflat, using this target should result in a systems that SOUNDS flat. By the way, see the Stereophile archives for J. Gordon Holt's article on why "Flat is Bad" (or similar such title).

Question 4, cont'd -- I'm also suspicious that the Tact inverts polarity (am running in DD mode)... I found that with the Tact in the signal chain, my systems sounds a bit more open if I switch on polarity inversion in my preamp (this effect passed the "honey, do you hear this?" A/B blind test). I suppose what I ought to do next time I calibrate is turn polarity inversion on, calibrate, and then turn polarity inversion off and run the system that way as a matter of course.

Question 5: I remeasure every time the room configuration changes. I've found that once I've calibrated the room, even relatively small changes are audible degradations
in sound quality -- a large pile of books on the coffee table for example changes things noticeably. I've changes some components ranging from cables to preamps and not
recalibrated. If these components do not have flat frequency response, then any sonic impact they provide will be (or should be) washed out the next time I recalibrate.

I am in complete agreement with you on the Fletcher-Munson thing, 1439bhr. The high end has eschewed equalization for good reasons, but if Fletcher-Munson is true, we are giving up a lot for the sake of purity of signal. I happen to listen mostly at lower levels. Am I trying for Fletcher-Munson compensation in my choice of components? Could be.

As I understand it, F-M compensation varies by volume. So an F-M target curve for 80 db would be different from one for 70db. Can of worms. If the software could adjust variably for volume, now that would be something!

Which raises another question. Is the volume of the system at the time of TacT measurement something to be concerned with? The TacT literature suggests not, but I wonder. With no preamp in the system (using the TacT as the only preamp), the test signal is pretty freaking loud. Should it be calibrated to normal listening levels (can it be?) or does it not matter?

Drubin, quite right about the F-M curves being indexed by reference SPL level... you'd have to program a set of target curves (no problem-o with TacT's 9 EQ memories). I think what happens in real life is that room gain gives the bass boost which is a key feature of the F-M curve... so even a speaker that measures flat anechoically will approximate a F-M curve. I think a lot of what goes on with system matching (including cable-ology) is an indirect attempt to equalize the system frequency response (to an actual F-M curve? actually, to the actual response of thatindividual's ear)

As far as volume level goes, it affects the SNR of the calibration measurement, so higher volume is better from that standpoint. On the other hand, I could see that loud pulses could start driving an amp and speaker into dynamic compression, which would alter the measured frequency response (the amplitude-compressed pulse would have differenct spectral content than the assumed spectral content of the digitally generated reference pulse). I haven't experimented with this issue; I simply set my preamp gain to what corresponds to a fairly loud listening level.