Statement (b) is the giveaway — it is exactly backwards. Negative feedback reduces distortion; that is one of the principal reasons for using it. With a feedback factor \(D=1+A\beta\), any distortion generated inside the loop is divided by D:

\(D_f=\dfrac{D_{ol}}{1+A\beta}\)

Because the output is compared with the input, any departure from a faithful copy appears in the error signal and is corrected. Statement (b) being false eliminates options 3 and 4 at once.

Statement (d) — correct. Negative feedback widens the bandwidth by the same factor that it reduces the gain, leaving their product fixed:

\(A_f=\dfrac{A}{1+A\beta}, \qquad f_{H,f}=f_H\left(1+A\beta\right)\)

\(A_f\times f_{H,f}=A\times f_H=\text{constant}\)

This is the gain-bandwidth product: a 741 with an open-loop gain of 200,000 and a 5 Hz corner delivers 1 MHz of bandwidth at unity gain, or 100 kHz at a closed-loop gain of ten. Gain is traded for bandwidth on a one-for-one basis.

Statement (a) — correct in the standard sense. Noise generated within the amplifier is attenuated by the same factor \(1+A\beta\) as the distortion, so relative to the signal it falls and the output signal-to-noise ratio improves. (Noise arriving at the input alongside the signal is amplified equally with it and cannot be improved by any amount of feedback — the improvement applies only to noise originating inside the loop.)

Statement (c) — wrong. Negative feedback does not reduce the input offset voltage. Offset is indistinguishable from a genuine input signal, so it is amplified by the full closed-loop gain and appears at the output as

\(V_{out,offset}=V_{os}\left(1+\dfrac{R_f}{R_1}\right)\)

Reducing offset needs trimming, chopper stabilisation or a better-matched input pair — not feedback.

Assemble. (a) and (d) survive, which is option 2.

The complete ledger of what feedback does, every entry scaling by the same factor \(1+A\beta\): gain down, bandwidth up, distortion down, noise (internal) down, gain stability improved, and input and output impedances raised or lowered depending on the topology.

Hence, the correct statements are (a) and (d).