So, as a discovery, I'd like some opinions-

What other problems do you see with a flat-line orientation towards SQ?

What advantages do flat-line responses seem to offer?

Can the incorporation of Fletcher and Munson data into system design improve subjective quality?

Basically the problem is a flat response doens't tell the entire story, reflections, phase problems, delays are all audible but not measureable with a RTA.

i know sometimes i like to adjust the eq as some songs to me sound like they should have more of this or less of that.

but i think a flat freq response gives more balance and makes the different sounds not overwhelm others, but sometimes you want that!

Something else: an RTA reading does not take into account is left/right bias (i.e. the mic will just take the amplitude of the sound wave incident on it, it cannot de-couple left speaker vs. right speaker to determine that left speaker is accounting for 70% of the sum at a particular frequency whereas right speaker is only 30% due to installation location/angle/etc.)

Why would anyone like a ruler flat (as far as an rta is concerned) system when they all typically sound very thin in the low end and extremely bright on the high end?

Build the system right, use an RTA for level setting ONLY, and get an EQ if you need one, then tune by ear, and enjoy your system.

At low amplitudes, the most sensitive region is the midrange.. Less sensitive and sloping to very insensitive are the regions of low bass and high treble.

As the amplitude goes up, however, the fletcher curve begins to flatten out.. That is, the ear starts to hear more like a flat line. Thats why stereo's usually sound better when turned up...

If a stereo is tuned for flat response at low volumes, then when turned up, the subjective performance will become _g_a_r_b_a_g_e_. The bass and treble will be over-emphasized and the midrange will be lacking not only prescence, but balance and tonality. If a system is tuned for a flat response at high volumes, then it may sound actually decent when played at high volumes, as this coincides with the ear in a very generalized sense.

Now, considering the non-linearities of the ear, a flat line doesn't sound good. However, at decently high volumes, a flat line represents the best way to remove the speaker effects from the music.. That is, if the frequency response of the speaker is flat, then everything that you hear must come from the music, balance-wise and tonal wise.

What I mean by a fletcherized response is a stepwise equalizer.. The Rockford EPX will do it.. As the volume changes, the unit switches from one preset to another.. You could equalize low levels and high levels all to different slopes.

However, what I'm thinking about is a more gradual effect than huge changes at one volume level..

I'd imagine the impulse/step response in car would be quite atrocious.

I can see how the fletcher munson curves may have an effect on perceived sound provided you aren't listening at the same reference level that the guy who mastered the music was. This is generally why the loudness button is on your source.

It is either the responsibility of the end user to listen to the music at the appropriate level for best sound reproduction, or the recording company's responsibility to release different versions of the same material mastered for different levels. (like that would ever happen!)

That said, I dont think the FM curves play a very significant role. The relative frequency response between say the 80 and 120 phon level is pretty much the same. The main variations are at the lowest frequencies, where the least detail is located. This would encompass 99% of all music levels played, including the one the recording engineer uses.

I think a flat response is a sign of a good system, but it isn't the end-all, be-all characteristic to strive for. Headphones for one, sound like **** with a flat response.

A level based EQ would be an enormous pain in the ass to accomplish in continuous time. Music by nature is very dynamic and is never in a "steady-state". An acoustic level based EQ would be switching modes continuously. Read, ALWAYS. To do it digitally, you would need say 65,000 different EQ settings for a smooth transition. ( I chose that number only because CD's are 16 bit and they had to have chosen that number for some good reason) I dont think this would work with mic's either. It could work at the line level with a host of sensors, monitoring temperature, humidty, etc. But still, I dont think this is at all the answer.

Home speakers can do a pretty damn good job of sound reproduction** without any EQ's and a fairly flat response. So, I think the problem lies solely in the acoustics of the car.

** (Dunlavy audio labs says they have done a comparison between their SC-V speakers and a live string quartet. Their tests "have revealed that experienced audiophiles cannot consistently identify which they are listening to with scores very close to random guessing" www.dunlavyaudio.com > tech talk )