Review of the first revised manuscript:

Review of: "Seeing 'where' through the ears: Effects of learning-by-doing and long-term sensory deprivation on localization based on image-to-sound substitution" by Michael J. Proulx, Petra Stoerig, Eva Ludowig, & Inna Knoll.

The authors have made a substantial effort to improve the original study and manuscript. I feel that my comments have been dealt with in a satisfactory manner. I would recommend accepting the ms for publication however with additional minor remarks and comments.

The authors have added precisions on the strategies used by subjects with system-none p.20. This was useful. However, I think that the reasons for having subjects that used no device and their subsequent expected results should be spelled out more clearly and early in the manuscript (in the methods, and/or in the first results). Otherwise, while reading the manuscript, we do not see why there is an emphasis on RT for these subjects and the interest of a comparison with other conditions, given that it does not take long to touch all the LED's rapidly when there is no searching time). Thus, if the aim of having Sn was to have a baseline by subjects with random pointing (i.e., Random accuracy and RT with no searching time) for highlighting the other results, this should be spelled out.

p. 6. The 2nd {section sign} should be reworded in order to understand that each part of the sentence corresponds to the 3 major aims of the study, e.g. focusing, comparing, and examining (or, first we focused, second we compared, and third we examined).

p. 10. "However, the Bc subjects had better accuracy than the Bt subjects, though this difference was not statistically significant."

Given that the difference between Bc and BT was not significant, this sentence does not seem to me correct: Bc subjects do not have better accuracy. It would be more appropriate to reword the sentence differently, such as: There was no significant difference between Bc and Bt subjects (stats). However, it should be mentioned that the numerical trend indicates a better accuracy for Bc than Bt (means).

p. 8 and 10. Interaction between blindfolding and system use (single subject analyses).
It is somewhat strange that most of the {section sign} does not correspond to the single subject analysis but to a discussion of the results obtained with Sc and Sn subjects. Maybe this needs a light justification, a change in the title, or being displaced?

p. 14. 2. "However, the Sc subjects who used the substitution system immersively in their daily life had superior performance to that of the St, in-laboratory, users of the device in all three experiments. Together, the results indicate that learning-by-doing in naturalistic conditions effectively improved performance on the localization tasks we presented."
This second conclusion of the study is not straightforward. The better results of Sc subjects cannot be for sure attributed to "learning-by-doing" given that it could solely be due to a longer practice with the device (4 hours per day). This is an important point that has to be corrected or convincingly justified.

p.16. and p. 22. The authors conclude p. 22 "Most of our results speak to the question of object localization. However, the increased directness of the grasping in Experiment 3 suggests that the subjects also gained knowledge of the objects' size, shape and orientation".
However, in order to make sure that the subjects really gain knowledge about the object's size, shape and orientation, we have to know if their grip becomes appropriate to the object and matches its overall shape. We do not have clue about this given that we do not know how the precision of the grip is computed and if this way of computing allows differentiating 2 different shapes or orientation. (i.e., p. 22 it is only said that "Direct grasping was attested ...and the hand-posture was largely appropriate to the size of the object"). Thus, either some precisions on the computation of the grip have to be added, or the conclusion has to be moderated.

p. 17. "All Sc subjects were asked to use the system daily for at least four hours, and, like the Sn subjects, wrote or dictated daily reports detailing their experiences".
I do not see the interest of having reports of the subjects that did not use the system, was it for a comparison? In addition, there was no qualitative data provided in the manuscript based on these reports. If these qualitative reports are submitted or available elsewhere, it might be worth mentioning it. Otherwise, this precision does not seem necessary.

p. 17. "Laterality is coded by stereo panning and the time provided by the left-to-right scanning transformation of each image every one or two seconds ..."
Although it is mentioned p. 20, it might be useful to add here the precision that time scanning is fixed, and users can choose 1 or 2 seconds.

p. 18. "see supporting online material, Video S4 for an example image converted into sound".
I've checked the demo. It seems that the 3 squares generate the same sound although they have different elevation (each has 3 different pitches but the same for the 3 squares).

p. 19. "The values of prep are directly related to p values, however, such that values of prep greater than 0.9 are also significant p values less than 0.05".
This sentence is hard to understand and could be reworded.

p. 21. "Five different objects that varied both in size (e.g. a pen, a CD, a trainer, a large box) and in contrast (a white shirt rolled into a ball, a grey plush mouse) to the cloth were used for each test."
This sentence is unclear. How many objects were used in total? It seems that there is missing information: It is said 5 different objects but there are more as examples.
In addition, did the subjects new the different types of object that were used?

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.