Peer Review History
Original SubmissionDecember 8, 2019 |
---|
PONE-D-19-33956 Machine Source Localization of Tursiops truncatus Whistle-like Sounds in a Reverberant Aquatic Environment PLOS ONE Dear Dr. Woodward, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. We would appreciate receiving your revised manuscript by Feb 23 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file. If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols Please include the following items when submitting your revised manuscript:
Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out. We look forward to receiving your revised manuscript. Kind regards, Haru Matsumoto Academic Editor PLOS ONE Journal Requirements: When submitting your revision, we need you to address these additional requirements. 1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf 2. Thank you for stating the following in the Acknowledgments Section of your manuscript: 'We thank the National Aquarium for participating in this study, as well the National 402 Science Foundation (Awards 1530544, 1607280), the Eric and Wendy Schmidt Fund for 403 Strategic Innovation, and the Rockefeller University for funding.' We note that you have provided funding information that is not currently declared in your Funding Statement. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form. Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows: 'MO Magnasco, DR Reiss Awards 530544, 1607280 National Science Foundation The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.' Additional Editor Comments (if provided): Dear Dr. Woodward, The manuscript has improved but it is still lacking details and scientific discussions. Although the application of machine learning to animal localization is unique and the results are interesting, I have to agree with both reviewers that the manuscript needs a major revision. As guidelines, 1) as reviewer 2 pointed out, PLOS One draws a broader audience and not just for experts in machine learning or bio-acoustics. For that, it needs more explanation without losing your audience. 2) The experiment must be repeatable with details of experimental set-up (e.g., hydrophone locations in XYZ as pointed out by both reviewers). I am sure that marine bio-acoustics researchers are eager to apply your ML method to the other animals in the ocean if you can describe the experiment clearly with a few jargon. Unlike the ocean, the aquarium tank is a controlled environment. 3) Re-submission without the details of your aquarium experiment set up and methodology, especially the dimensional information, will jeopardize further review. 4) Before resubmitting the revision, please ask your coauthors to proofread your manuscript carefully in order to save the reviewer's time. Regards, Haru Matsumoto [Note: HTML markup is below. Please do not edit.] Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Partly Reviewer #2: Partly ********** 2. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: No Reviewer #2: Yes ********** 3. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified. Reviewer #1: Yes Reviewer #2: Yes ********** 4. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: No Reviewer #2: No ********** 5. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: This paper aims to tackle an important challenge of localization: associate the recorded whistle to the sound-producing animal. Instead of TDOA-based localization algorithms, the authors used classify the approximate locations and estimate the locations (regression) through random forest, support vector machine and other machine learning algorithms. It was a promising work but the data could be summarized more clearly, experiment setup laid out more precisely and results reported more systematically. It seems to be a work written by a junior academic member without proof-reading from other co-authors. There're simply too many errors and inconsistency all over the manuscript. I recognized that there might be a diamond in the rough. However, the paper have to be organized in a better shape so that scientific findings can be communicated effectively. Detailed comments are as follows: In abstract, please refrain from using abbreviations without the full names. What is MAD? What is IQR? Why is SRP the abbreviation of Steered-Response? What does P stand for? It's not clear what it means by "...comparable accuracy - even when interpolating at several meters - in the lateral directions when deprived of training data at testing sites..." and "...in all domains..." Line 120-123: It'll be easier to understand through an equation or an example on the figure. Line 127: It's almost impossible to imagine what y(t) might look like. The denominator arcsin(Phi(t)) are very likely to be incorrect since the input to arcsin needs to be [-1, 1] whereas Phi(t) can be outside this range. Line 129: Why does m=0.8 correspond to two harmonics? Line 133: What is "pre-speaker"? In the legend of Figure 1, it says duration = 1 s whereas Figure 1 only shows roughly 340 ms. In addition, please use SI units everywhere in the manuscript, i.e., "s" instead of "second" or "sec". Details and examples of SI units can be seen: https://www.bipm.org/en/measurement-units/ Line 151: what are the depths of the hydrophone? What's the separation of the four hydrophones in a location? If there's no significant separation of depth, how can you expect that they are able to locate the depth of locations? Line 163: the authors mentioned "Audacity AUP sound format". AUP is not sound format. It's Audacity's format for organizing projects. Line 175: "...with sinusoids and quasi-sinusoids with the same parameters grouped together given their close similarity." There might be too many "with"s. Line 205 - 207: Why does 27,126 Fourier transform elements correspond to 0 to 27,126 Hz? What is your FFT/DFT size? In order to have 1 s frequency resolution, DFT size needs to be 192,000! Line 210: Why is the number of features 897,871? I calculated it to be 897,736 (=6,601 x 136) Line 251-255: the training and testing data are unclear from the authors' description. Neither were they found in the previous section. The second way seems to imply that the authors used "leave-one-out" to do the validation in that training the model by all the data points except one that is tested. However, if this is true, what are the testing data for the first way? Did the authors use all training data for testing, which is definitely incorrect way? Line 285-289: 6,788 or 6,778 features? Two different numbers appeared. The authors must do a better job to proof-reading before submitting the paper. Line 295: "...achieved 100.0% cross-validation and 100.0% test accuracy..." I doubt that this statement is true. What is your training/validation data split? After 10-fold cross-validation, how do you use the data to train the final model? Line 296-297: How were you able to have test accuracy 97.75% and 99.44%, respectively? Line 171 mentioned that there are 1,605 recorded tones and 10% were used for final testing. Thus, either 160 or 161 were the number of testing tones. Either number would not result in 97.75% or 99.44% unless repeated experiments were conducted and average results were reported. In addition, confidence intervals were reported. There must be repeated experiments and is was not mentioned at all in the method. Line 305: What are "EP front" or "EP Wall"? A figure might be needed for readers to understand. All figures need to have better quality. Legions are difficult to read, almost unintelligible. Reviewer #2: This work represents a non-deterministic approach to source separation/localization using machine learning methods. The work likely has merit as proof of concept, but the lack of details make it non-repeatable and I have some concerns about the methodology. Subsequently, I cannot recommend publication as presented. My review largely echoes those of previous reviewers and while I note that there has been some positive movement towards explaining the methods and results in full, much remains to be done before this can be accepted into the bioacoustics and source separation cannon. A major concern is how the sounds were generated. It appears that the sounds were produced sequentially at each of the speaker locations and the authors used machine learning (ML) to discriminate between these locations. Thus the authors ask, which of these locations could this sound have come from. This is a valid question from a theoretical standpoint but it’s value is only useful when animals are greater than half a body length apart and whistles are not produced in isolation. To make it applicable to the field the authors should ask, ‘Which animal did each source come from’. In doing so the authors would play sources from multiple locations simotaneously. This is especially important given that in ML, it’s difficult to determine which features are the most useful in the discrimination task. If the features are relative amplitude of the first arrival at the four hydrophones then the value of the proposed method is limited. Technical Concerns No definition about how tonal extraction was done. Were the 2 second clips taken from 1 after the onset of the tonal recording? From the peak time? From which hydrophone was the timing obtained? For example, they sound clips could be slightly offset on each hydrophone. Did an analysist manually select the clips? How the clips are extracted will determine the features useful for source separation and subsequently what their system uses to discriminate. Organization The authors have not organized the Methods section of the manuscript in a manner consistent with the field of bioacoustics and machine learning which is causing considerable confusion. To be consistent with the field, better document their work, and convey their results to the broadest audience possible, the authors should arrange the methods section as follows. Array setup This including speaker and receiver positions. Failing to include the hydrophone locations within the body of the manuscript is inexcusable. Where hydrophones are placed relative to the sources and eachother is fundamental to the study and the ability of the methods to generalize. As noted by the previous reviewers, the authors need to include this in the methods as well as in the figure. Signal generation This section does not need to be as extensive and could possibly go in supplemental information. One or two paragraphs with the included table will suffice. The authors choice to exclude harmonics may problematic without considerable explanation. In doing so they have produced idealized tonals which are not representative of biological signals. This choice needs to be justified. Signal acquisition This should include how the authors parsed the signals of interest from the recordings. It is not clear how long each recording used to train the regression trees was. Did they collect two seconds around the peak of the received signal? Was it the onset or was some other method used? This is a critical and not well documented aspect of their work. Further down the authors refer to ‘snipits’ of data. The word choice is fine but what each snipit is and how it was generated needs to go in this section. Feature Extraction. This section is key in understanding and replicating their work. As noted by previous reviewers, it’s still not clear. I advise that they exclude or move to the supplemental information any bits of the feature extraction analysis that wasn’t used in the final model and clarify what was included. Introduction The value of this work concerns sound source discrimination. The introduction needs to be structured to highlight this. The authors have gone through some efforts to provide background information on the biological motivations for the system but have not done due diligence to the extensive body of work available on source separation in marine mammals. However, there are some issues in what they are presenting in overstating the value of captive studies in signature whistles. While captive studies were integral in the initial study of whistles and have some value in ontogeny initially, the limitations are considerable. I suggest that the authors re-structure the introduction to focus only on caller discrimination and limit the potential applications to small portions of the discussion. Clarify the issue with source separation in a reverberant environment. Please see work by EM Nosal on sperm whale source separation, a variety of papers by D. Mellenger and other contemporaneous researchers. At minimum, these papers would demonstrate the proper terms in the bioacoustics field. Methods 86 - The paragraph starting on line 86 is a single sentence. This does not help the reader understand the work. 92 – This paragraph could be considerably simplified Consider rewording: We generated 128 unique tonal sounds with pitch and duration within published ranges of Tursiops (table). We used frequency modulated pure tones randomly generated from XXX distribution. For this analysis, harmonics were not included. For details on the generation system see supplemental info. 92- replace ‘sounds’ with ‘tonals’ throughout to refer to your generated signals. ‘sounds’ is too vague for a bioacoustic audience 119- this should be it’s own section (array setup, see above) 120- Consider rewording for clarity: 'The 128 generated signals were played at each of the 14 hydrophone locations corresponding to 7 horizontal positions and two depths, xx m and yy m (Figure). Horizontal hydrophone spacing was approximately XXX between adjacent locations. See linked data below.' Talking about the cross doesn't add much. Also nix the imperial measurements. Let’s momentarily pretend we are from a civilized country. 123-130- Unclear Name and thank the assistants in the acknowledgments. 144- ‘Various calibrations’? Define what was calibrated. Explicit details for well established procedures are not needed but I vehemently disagree that defining what they did is ‘outside of the scope of the study’ - First reference to relevant work not included in the manuscript. 145 – Replace ‘collected at’ with ‘sampled at’ as in sample frequency 148- Replaced ‘involved’ with ‘used’ 148 – What does ‘Standard Passive acoustic monitoring system’ mean? Does it mean custom matlab scripts were used to autonomously record sound from the hydrophones? As far as I’m aware there are no ‘standard’ methods for PAM in matlab. Clarify. 135- Why is there so much information about the visual recording system? The need for this section is lost on me. Please clarify how the system fits into your study. Knowledge of the hydrophone and speaker locations within a tank shouldn’t require an advanced visual system unless there is considerable and undocumented flow preventing the speakers from remaining stationary. Clarify. 153- This section is the heart of what the authors did and is really difficult to get through. The jargon is very heavy and multiple concepts are introduced in the same sentences. The authors need to walk your readers through this more carefully. There is clearly a lot going on and a lot of work to be carefully discussed. Do justice to it by conveying it to a broader audience. 158- ‘digested’ is the wrong word here. Pick something more accurate. 15-- It's not a, 'so-called' feature set. It’s a feature set. Also, this is part of the confusion. It's not immediately clear what is going on here. Consider rewording: XXXX features were created from each 4-channel recording of the simulated tonals by converting the wave forms to YYY…. 161- GCC-PHAT isn’t explained but TDOA is. The authors need to spend more time on the former and less on the latter. 161 – Currently unpublished -> put it in the supplemental information Second reference to work not included in the manuscript. 163- Remove ‘briefly’. A ‘brief’ explanation of TDOA consists of , ‘TDOA is the delay in the arrival time of a signal between multiple hydrophones where hydrophones further from the source receive the signal later than those closer to the source’. Feel free to use this. 168- Where ‘elsewhere’? Either this concept is important and warrants explaining and citing or it should be removed. Third reference to work not included in the manuscr 137-173 a single sentence with multiple interjections. Lost. Still unsure how you are getting the 120 features from the 16 hydrophones. Hydrophone schematic would help. 176- ‘snipits’- this is reference above and should be defined in your data acquisition. 176-178- this bit is actually quite good and understandable. 179- Would be nice to know what the geometry is… 180 – this section isn’t clear again. Do you mean that the cross correlation time did or did not include second through n-th arrivals? 187- So the only information going into the regression tree is TDOA, cross correlation and GCC-PHAT? Be explicit here to give your readers a break. Remove what didn’t work or add it in the supplemental information. 187 – What do you mean ‘processing’ for each whistle? I thought the snipits were processed for the feature set, not the other way around. 187- This is the first reference to the labeled data. In the ‘feature set’ or earlier in the methods state that you are using supervised learning and your targets and label sets represent the potential source location and XYZ coordinates. 189- Typical phrases in ML to refer to data used to train and test the data are ‘Traning data’, ‘Test Data’, and/or ‘Validation data’. The wording here is awkward 191- what was ‘novel’ about the whistle? Again stick to the same term for the sounds you generated. I suggest, as above, ‘tonal’ and only use ‘whistle’ to refer to the actuall, real life, whistles coming from the animals. 191- This isn’t true, the authors could batch generate but never the less the sentance should go. Consider rewording ‘We chose the Bierman random forest for the classification task due to its ability to reduce the feature space and address whilst performing multi-class classification. 193-196- This is a single sentence with two interjections. Difficult to read consider rewording. The Bierman random forest is a multi-class classifier with a built-in resistance to overfitting through XXXXX. Additionally, the classifier performs feature reduction through YYY. 200- CART acronym without explanation. Write it out the first time. 204 – This remark can easily be read as condescending, especially when taken in context with the various other references to work not included in the paper. 211 – unclear what ‘additional’ models are 221- The localization portion of your study needs it’s own section. 221- replace ‘training sounds’ with ‘snipits’ or ‘tonals’ depending on whether you are referring to the sounds you generated or the sounds you extracted. Consistency. 121-123- Are you referring to the features extracted from each recording? I’m very lost. 222- Three SVM models? In the start of the to-be-created localization section, please provide a brief 1 or 2 sentence overview to SVM and how you use it here. 223- grid point? Be specific throughout. 227- This justification should be in the start of the section about the localization 229- Employed- I hope you paid it a living wage. Replace with ‘used’. 230 – Is the ‘standard’ approach what is referenced in your citation? Or just some details? Clarify. 231- Is this what you mean by grid-point? Note how much later the definition is than it’s first use. If not clarify. 231-236- This could be simplified. Consider: We divided the available pool space into 6cm grid squares representing all potential source locations. For each grid corner, we calculated the expected TDOA of a source at that location to each of the 16 hydrophones. 233-236- imprecise wording. I know what the authors mean but others may not and all struggle. 237- I assume it’s a single value for soundspeed. State the calculated soundspeed or provide a figure of the soundspeed profile of the pool. 242- state the purpose of the procedure at at the start of the paragraph, not the end. Results 248- Remove, ‘described above’. Else replace it with a section reference. 252- Not sure what is meant by this. 254- Completely lost. By array you mean the 4 connected hydrophones? The descriptions of when the authors use all 16 hydrophones and when arrays are treated separately isn’t clear to the reader. 248-255- 100% accuracy while increasing the model size strongly suggests overfitting 257-258- This should be in the methods 264- Comma after ‘again’ 265- first mention of kriging. Methods. 235- replace ‘test sound’ coordinates with ‘speaker’ or ‘sound source’ coordinates 270- The error looks considerably worse in the Z direction in comparison to the X or Y. This is useful information that I hope is brought up in the discussion 271-277 single sentence 217-281 – This would be useful as a table with rows being the axes and columns as the models Figure 4- Bar graph or histogram. Remove ‘x-ticks’ denote histogram edges. This is confusing, matlab speak that isn’t useful for non-matlab users. The figure itself is hard to parse. The variables are the models, the different axes, and the histogram bins. The comparison the authors should be making is the different models (pannals) for each axis. So, I suggest the authors make each axes a different panel and the colors should represent the models. This will make the model comparison significantly easier to see. Figure 5- I don’t see a lot of value added by this figure. Figure 5- caption – Remove ‘discussed in the text’. Much of the rest of the caption should be in the methods. 289- removed ‘so-called’, 289-290 – methods. 294- replace ‘ask’ with ‘determine’ 298-299- Major point, don’t bury it at the end of the results Figure 6- Unit labels for colorbars. Replace hydrohpne ‘label’ with hydrophone ‘number’ provide figure of hydrophone layout with each hydrophone number. Figure 6 label – ‘common panals’ what? First reference. Describe better in ‘array setup’ section to be added TDOA features do not seem very useful here. Highlight in results. Discussion 301-303-Run on 307- I presume by ‘sound sources’ you are referring to free-swimming animals? If so say it. 307-310- run on 309-311- unintelligible. Reword. 312 – Are you sure recordings is the word you want? 4-channel arrays/panels? It should be arrays but this needs to be made clear throughout. 314- EP already defined (or should have been) 317- sound ‘source’ originated 321-326- run on 331-333- run on. Start with , ‘Also, it is reassuring that a….’. Try to always put the subject next to the verb or risk sounding like Yoda, do you. 334- Which question? There was no stated question 340- It’s not so much the length of the dolphin that’s important it’s the width. The sound source in a dolphin is near the front of the animal. So, the authors need to highlight that this method shows promise when animals heads are greater than a meter apart which may, or may not, be tenable. 355- Just add the citation after elsewhere remove the rest. 356-Move ‘referring to figure 5’ to later in the sentence. Just reference it as (fig 5) I think you should replace (EP-Wall) and EP dimensions with X, Y, and Z or lat, lon, depth. Something more intuitive will allow for readers to better understand the results 370- time-of-flight? Used in abstract and elsewhere. Undefined. 371- Reword to say amplitude was not include rather than it was removed. Highlight just TDOA and other methods that were included. Consider, ‘In this work TDOA and Cross correlation values were used to discriminate between source locations. Direct or relative amplitude was not included in the feature set’. Or something like that. Acknowledgments This is an online journal without word limits. Name and graciously thank the two-dozen people who helped you. Don’t be lazy. ********** 6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: No Reviewer #2: No [NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.] While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step. |
Revision 1 |
PONE-D-19-33956R1 Machine Source Localization of Tursiops truncatus Whistle-like Sounds in a Reverberant Aquatic Environment PLOS ONE Dear Dr. Woodward, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. We would appreciate receiving your revised manuscript by Apr 16 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file. If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols Please include the following items when submitting your revised manuscript:
Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out. We look forward to receiving your revised manuscript. Kind regards, Haru Matsumoto Academic Editor PLOS ONE Additional Editor Comments (if provided): Dear Dr. Woodward, I believe that it is the first application of ML to the localization problem of the sound sources in the water. Your research is unique and results are interesting. You have improved the manuscript by taking suggestions from the two reviewers. All the errors pointed out by the reviewers' were corrected. The experimental setup and results are more clear now. However, the main criticism of the two reviewers is the intelligibility of the manuscript, which is still a problem. For example, the abstract and intro are very confusing and difficult to follow even for us in underwater acoustics and marine mammal fields. Please keep in mind that often the abstract and the results are the sections that readers read first. If they are hard to follow, readers would not read the rest and would not cite your research. The reviewers did the job they were asked and I am not going to send the manuscript back to them for further review. I have suggested in my previous comment to invite someone outside of ML or AI field (preferably someone with underwater acoustic background) to read the manuscript carefully and improve intelligibility. Have you done that? You can include him/her as a last co-author. If you disagree with my decision, another option for you is to request PLOS ONE to change the science editor, which I do not mind at all. But please keep in mind that it would be the same long process again of finding new reviewers and revising. One error that I just noticed is in Fig. 2. The unit you use, dB/Hz is not correct. It should be dB relative to the standard unit (Pa/sqrt(Hz) for sound). Also it is not clear if it is the actual sound, electrical signal or simulation. You have used SQ26-08 hydrophone from Cetacean Research, which is a calibrated hydrophone. But the dB values are too small for sound or electrical power. Regardless, the spectral intensity is already normalized by a unit frequency (uPa/sqrt(Hz) for sound) when you do FFT. Also, it may be interesting for readers if you can include one of the hydrophone channel signals and spectrograms in the reverberated environment of the tank. Regards, Haru Matsumoto [Note: HTML markup is below. Please do not edit.] Reviewers' comments: [NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.] While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step. |
Revision 2 |
Learning to localize sounds in a highly reverberant environment: machine-learning tracking of dolphin whistle-like sounds in a pool PONE-D-19-33956R2 Dear Dr. Magnasco, We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements. Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication. An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org. If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org. Kind regards, Haru Matsumoto Academic Editor PLOS ONE Additional Editor Comments (optional): Congratualtions! Your paper has significantly improved. I appreciate all the efforts you put in. Reviewers' comments: |
Formally Accepted |
PONE-D-19-33956R2 Learning to localize sounds in a highly reverberant environment: machine-learning tracking of dolphin whistle-like sounds in a pool Dear Dr. Magnasco: I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department. If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org. If we can help with anything else, please email us at plosone@plos.org. Thank you for submitting your work to PLOS ONE and supporting open access. Kind regards, PLOS ONE Editorial Office Staff on behalf of Dr. Haru Matsumoto Academic Editor PLOS ONE |
Open letter on the publication of peer review reports
PLOS recognizes the benefits of transparency in the peer review process. Therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. Reviewers remain anonymous, unless they choose to reveal their names.
We encourage other journals to join us in this initiative. We hope that our action inspires the community, including researchers, research funders, and research institutions, to recognize the benefits of published peer review reports for all parts of the research system.
Learn more at ASAPbio .