Basic Op Amp audio Listener

Hello Hello,
So, Please bare with me i'm new to this... The Goal is a small handheld (probe) listening device for trouble shooting audio signal flow. it will be powered off of 2x1.5 volt cells and drive a 32 Ohm, 12mW Nom. 20mW Max. transducer (Speaker). I also wanted to have built in Limiting to prevent damage to the device or the listeners ears. I believe i have accomplished this with the addition of D1 & D2 and then softened there effects by adding the voltage divider made up of R6 & R7. (this should give me -22db Apx. "PAD" when R3=2k increasing to 36db Apx "PAD" when R3 = 10k one thing that concerns me is if this compensate for the added gain. -0.425 at 2k and -2.125 at 10k and how to smooth the transition when D1,2 Reach there 0.7V Apx. forward Bias also it doesn't seem right to just calculate the the Gain when D1,2 are biased as you would if R3 and R6 where in Parallel?)
Op Amp Choice: i have looked at a few options in this regard and think i have settled on a TLV2782 or TLV2783 depending on power switching or IC Shut Down method.. what are your thoughts on this IC? Link to Data sheet below.
http://www.ti.com/product/tlv2782
Other design considerations:
The input must be AC coupled to keep any DC Bias from damaging the circuit.
I want to keep current draw to a minimum to extend battery life.
the input signal should be audible anywhere between -40dbu and +4dbu i've been looking at building 3 units for testing each with a deferent input sensitivity; -10dbu, 0dbu, & +4dbu.
all components must be surface mount and the over all size and number of parts must be keep to a minimum.
Thank you very much for having a look at this and i appreciate all the great input i'm sure you will have..

Did you build up the circuit and try it out? Just curious if it worked as you described.

Hello Cody,

Thank you very much for the reply! i am currently awaiting the Arrival of the 2782 Op Amp and a couple different Diode pairs for testing.. i should have parts and a prototype built by the end of the week, i'll be sure to let you know if it performs as expected.

one thing that troubles me is that shifting center frequency of the band pass filter as either R3 is adjusted to set input sensitivity or as D1, D2 reach there forward Bias. for example:

If R3 = 2k then we have a center frequency of 1,990Hz with an over all Gain of -0.4251, a Q of 0.024, and Damping of 20.. with an input voltage of 1.227Vrms this should give an output voltage of 0.5218Vrms in the pass band, below the required forward bias required by D1 & D2.. But when the input voltage rises to 1.545Vrms the output voltage reaches 0.65Vrms the diodes are biased the predicted response of the circuit changes to a center frequency of 10,975Hz, Gain of -0.0139, Q of 0.0045, Damping of 110.. (also the ESR of C1 & C2 will have an effect on this)

See Attached File: Top = +4db Sen below forward Bias and the bottom is the response when Volt out > required Bias.

as this is quit a substantial change in response and will be audible at the output when an audio signal reaches the device and its amplitude is fluctuating above and below the 1.5V Apx. threshold the continual switching of D1&D2; could make the output unusable.. i have thought of adding a Cap in the with Diodes to slow this but i have yet to come up with a clean way of doing this or a correct value for it.. (this one may just come down to some trial and error)

Please let me know if you have any thoughts regarding this design or a way of Softening / Slowing the limiting circuit.

Thank Again,

Colle