Dear Editor,

We like to thank referee A for following our argumentation and for the
encouragement to outline the wider impact in the last round.

We also thank Referee B for his appreciation of our manuscript and of the
results and the fast reaction, as the work by Liu et al. [PRL 113, 133601
(2014)] from the Du group was just published two days ago.
However, we are surprised that he/she only reads our manuscript as a comparable
work to the work just published. Our work was developed independently, as the
referee acknowledges. In fact, we could not find any reference to it on the
arxive, otherwise we would have mentioned it at an earlier stage.

More surprising is the referee assertion that "Of course Liu et al's work covers
more and the envelop conversion, which is the full content of this manuscript,
is only part of Liu's PRL work". The referee does not indicate anything to
corroborate this statement, and we think it is the other way
round. In both works there is a study of the effect of the temporal waveform
of an heralded single photon in the probability of loading a cavity, in
accordance with what was already predicted and demonstrated for coherent
states by Bader et al. in [ref 17].
In our manuscript we also demonstrate the time reversal of the heralded
photon, something that can be inferred from the experimental results of Liu et
al., but is not presented as such in their work. Moreover, we present a
technique to generate exponentially rising photons without the use of
lossy techniques like temporal filtering with modulators, allowing an
efficient manipulation of states that is not achievable in the setup presented
by Liu et al.

Here we address the specific points raised by Referee B:

1) The Eq (4) is valid only for the situation when the cavity decay time
matches the coherence time of the photons \tau. This is clearly stated in the
text above the equation. All the measurements presented in our paper are
performed under this condition. We feel that inclusion of different time
constants, as suggested by the referee and shown in Liu's work, is a technical
detail that is somewhat tangential to the physics we want to present.

2) In the work of Bader et al., [ref 17], the estimated energy accumulated in
the resonator is normalized to the total incident energy, as clearly stated
just after eq. 8 of [ref 17]. Energy conservation makes it thus impossible to
obtain a value larger than 1, independently of the classical or quantum nature
of the incident light.  The same constraints apply to our experiments.
We also feel that the validity of statements made for classical light for
non-classical light in terms of propagation through linear optical elements
(and the cavity is such an element, very much like lenses, mirrors,
interference filters and the like) both in time and space is of reasonably
common agreement among physicists and we don't think it warrants specific
mentioning.

3) We actually strive to keep the manuscript simple to make the underlying
physics accessible to a wider audience, without bloating it with
unnecessary equations and explanations. The only thing that is specifically
pointed out as missing by the Referee is the design of the cavity. We believe
the design of a linear optical cavity is reasonably well explained in
undergraduate courses in optics, but we do give a reference to a textbook [ref
24] where some of the commonly established details are nicely explained. We
believe that our description of the cavity is actually very detailed, as we
specify the radii of curvatures and the reflectivities of the mirrors. We also
explained how we measured it, which should permit a reader to understand what
losses the measurement captures. Our cavity in this respect is nothing 
particularly different than any other standard optical cavity - apart from the
asymmetry in the reflectivities, and for that we believe to have provided a
sufficiently detailed characterization. What we did not explicitly mention in
the manuscript is the usage of a fused silica spacer to keep the mirrors at
the quoted distance of the 55mm. We believe this is also a standard technique,
but we have included it now in the updated version of the manuscript for the
benefit of a less experienced reader.

4) In the manuscript, second page, first column, we already specify that:
"All four modes are collinear and propagate in the same direction", explaining
that the pump, signal, and idler modes overlap completely. We also specify the
atomic transitions involved in the generation. Those are the only two
differences with respect the previous description of the source in our
references [27] and [16].

5) The cavity does not require active stabilization. The passive mechanical
stability provides a long enough time to perform the measurements after
alignment. The detuning values we quote were verified with a probe laser
locked to a Rb transition, frequency-shifted by an acousto-optical modulator.

6) In the revised manuscript we now cite the recent Liu. et al. work in our
references.


List of changes:

1. In the description of the cavity,
" The mirrors are separated by 5.5 cm corresponding to a free spectral range
\Delta\nu_f = 2.7 GHz.Therefore, an incident photon of Fourier bandwidth
1/(2\pi\tau) = 27 MHz interacts effectively with only one longitudinal mode of
the cavity, ensuring that Eq. (2) is an adequate model"
is now changed to
"The mirrors are mounted on a 55 mm long fused silica spacer, corresponding to
a free spectral range \Delta\nu_f = 2.7 GHz. Therefore, an incident photon of
Fourier bandwidth 1/(2\pi\tau) = 27 MHz interacts effectively with only one
longitudinal mode of the cavity, ensuring that Eq. (2) is an adequate
model. Temperature control of the spacer allows precise tuning and
stabilization of the resonance frequency of the cavity." 

2. We added the following sentence at the end of the manuscript:
 "Note added in proof: We became aware that during the review process, a
  similar work independently performed by Liu et al. has been published
  [28]."

3. We added the reference of the work by Liu et al. to the bibliography.

With this, we hope to have addressed the issues raised by referee B. We
believe many concerns were probably rooted in the extremely short time the
referee must have had between the publication of the Liu work, and the
comparison with our manuscript, and thank her or him again for the quick
response.

With Best Regards on behalf of all authors,

Christian Kurtsiefer