openjournals / joss-reviews

Reviews for the Journal of Open Source Software
Creative Commons Zero v1.0 Universal
721 stars 38 forks source link

[REVIEW]: Software Design and User Interface of ESPnet-SE++: Speech Enhancement for Robust Speech Processing #5403

Closed editorialbot closed 11 months ago

editorialbot commented 1 year ago

Submitting author: !--author-handle-->@neillu23<!--end-author-handle-- (Yen-Ju Lu) Repository: https://github.com/espnet/espnet Branch with paper.md (empty if default branch): Version: v.202310 Editor: !--editor-->@faroit<!--end-editor-- Reviewers: @joimort, @justusschock Archive: 10.5281/zenodo.10048174

Status

status

Status badge code:

HTML: <a href="https://joss.theoj.org/papers/f518d0b78a29ab6521936834e5d90ef1"><img src="https://joss.theoj.org/papers/f518d0b78a29ab6521936834e5d90ef1/status.svg"></a>
Markdown: [![status](https://joss.theoj.org/papers/f518d0b78a29ab6521936834e5d90ef1/status.svg)](https://joss.theoj.org/papers/f518d0b78a29ab6521936834e5d90ef1)

Reviewers and authors:

Please avoid lengthy details of difficulties in the review thread. Instead, please create a new issue in the target repository and link to those issues (especially acceptance-blockers) by leaving comments in the review thread below. (For completists: if the target issue tracker is also on GitHub, linking the review thread in the issue or vice versa will create corresponding breadcrumb trails in the link target.)

Reviewer instructions & questions

@joimort & @justusschock, your review will be checklist based. Each of you will have a separate checklist that you should update when carrying out your review. First of all you need to run this command in a separate comment to create the checklist:

@editorialbot generate my checklist

The reviewer guidelines are available here: https://joss.readthedocs.io/en/latest/reviewer_guidelines.html. Any questions/concerns please let @faroit know.

Please start on your review when you are able, and be sure to complete your review in the next six weeks, at the very latest

Checklists

📝 Checklist for @justusschock

📝 Checklist for @joimort

editorialbot commented 1 year ago

Hello humans, I'm @editorialbot, a robot that can help you with some common editorial tasks.

For a list of things I can do to help you, just type:

@editorialbot commands

For example, to regenerate the paper pdf after making changes in the paper's md or bib files, type:

@editorialbot generate pdf
editorialbot commented 1 year ago
Software report:

github.com/AlDanial/cloc v 1.88  T=3.74 s (1008.8 files/s, 125748.1 lines/s)
--------------------------------------------------------------------------------
Language                      files          blank        comment           code
--------------------------------------------------------------------------------
Python                         1244          33140          36882         174020
Bourne Shell                   1103          19396          13931          91530
YAML                           1038           3985           5721          40970
Markdown                        269           5849              0          22512
JSON                              3              0              0           8675
Perl                             71            869           1858           5629
MATLAB                           14            411            490           1691
diff                             13             49            406           1075
make                              2             53             19            206
CMake                             1             40             39            169
TeX                               1             22              0            156
Bourne Again Shell                5             23              6             73
Cython                            1             10              3             36
DOS Batch                         1              8              1             27
reStructuredText                  5             25             38             20
--------------------------------------------------------------------------------
SUM:                           3771          63880          59394         346789
--------------------------------------------------------------------------------

gitinspector failed to run statistical information for the repository
editorialbot commented 1 year ago

Wordcount for paper.md is 1359

editorialbot commented 1 year ago
Reference check summary (note 'MISSING' DOIs are suggestions that need verification):

OK DOIs

- None

MISSING DOIs

- 10.1109/slt48900.2021.9383615 may be a valid DOI for title: ESPnet-SE: End-to-end speech enhancement and separation toolkit designed for ASR integration
- 10.21437/interspeech.2022-10727 may be a valid DOI for title: ESPnet-SE++: Speech Enhancement for Robust Speech Recognition, Translation, and Understanding
- 10.1109/icassp40776.2020.9053512 may be a valid DOI for title: ESPnet-TTS: Unified, reproducible, and integratable open source end-to-end text-to-speech toolkit
- 10.18653/v1/2020.acl-demos.34 may be a valid DOI for title: ESPnet-ST: All-in-One Speech Translation Toolkit
- 10.1109/icassp43922.2022.9747674 may be a valid DOI for title: ESPnet-SLU: Advancing spoken language understanding through ESPnet
- 10.21437/interspeech.2018-1456 may be a valid DOI for title: ESPnet: End-to-End Speech Processing Toolkit
- 10.21437/interspeech.2020-1673 may be a valid DOI for title: Asteroid: the PyTorch-based audio source separation toolkit for researchers
- 10.1109/icassp40776.2020.9054266 may be a valid DOI for title: Dual-path rnn: efficient long sequence modeling for time-domain single-channel speech separation
- 10.1109/icassp.2018.8462116 may be a valid DOI for title: Tasnet: time-domain audio separation network for real-time, single-channel speech separation
- 10.1109/taslp.2019.2915167 may be a valid DOI for title: Conv-tasnet: Surpassing ideal time–frequency magnitude masking for speech separation
- 10.1109/icassp.2019.8683855 may be a valid DOI for title: SDR–half-baked or well done?
- 10.1109/icassp.2001.941023 may be a valid DOI for title: Perceptual evaluation of speech quality (PESQ)-a new method for speech quality assessment of telephone networks and codecs
- 10.1109/icassp.2017.7952154 may be a valid DOI for title: Permutation invariant training of deep models for speaker-independent multi-talker speech separation

INVALID DOIs

- None
editorialbot commented 1 year ago

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

justusschock commented 1 year ago

Review checklist for @justusschock

Conflict of interest

Code of Conduct

General checks

Functionality

Documentation

Software paper

justusschock commented 1 year ago

@faroit @neillu23 I completed my review and I recommend the paper for publication. I did not (yet) check the quality of writing box as I the layout of the paper (with the big figures on pages 3-7) doesn't look too good. I'd recommend to either change the formatting there or find another way to present the information. From the wording and language itself, the paper is fine though.

faroit commented 1 year ago

@justusschock thanks for your review. Can you please add an issue to the repo so that we can track it back here?

@neillu23 i agree with @justusschock that this could be improved

faroit commented 1 year ago

@joimort We are typically aiming for reviews to be completed within about 2-4 week. So please let me know if you need any help.

neillu23 commented 1 year ago

@justusschock @faroit I appreciate your review! I plan to create a pull request to make changes to the figure. Should I proceed with it immediately or should I wait for the results of the other review?

joimort commented 1 year ago

Review checklist for @joimort

Conflict of interest

Code of Conduct

General checks

Functionality

Documentation

Software paper

joimort commented 1 year ago

Thanks for your submission, @neillu23! The software and documentation meets all the criteria of the JOSS. Also, the accompanying paper is generally well-written. It provides a comprehensive view of the technology.

However, I do have a few recommendations that could enhance readability and clarity. For instance, the section "ESPNet-SE++ Recipes for SSE and Joint-Task" commences with a screenshot of the folder structure, which can be somewhat confusing for readers. As suggested by @justusschock, it would be better to present this information in a different way. The textual content contained in the figures on pages 2-7 could be transformed into straightforward text, which would be easier to understand. For the exact structure of the folders, perhaps the authors could refer the reader to the respective README files for further clarification.

Figures throughout the manuscript are not numbered, which could potentially lead to confusion. It is unclear whether this is a JOSS requirement, so the authors might want to consult the JOSS guidelines regarding this matter. Specifically, the figures on pages 7 and 9 possess two titles, adding to the ambiguity.

Some other minor imperfections: In terms of the code examples provided on pages 11 and 12, it is unclear why the code is highlighted. A minor formatting issue was noticed in lines 154/155, where a link appears to be improperly formatted due to a line break. Lastly, there are a two points related to citations that need to be addressed. In line 198, the authors reference "Asteroid" but cite it with the arXiv version (published at Interspeech). Similarly, the conference name, Interspeech, is missing in line 217.

These recommendations are intended to further enhance the quality of this already solid manuscript. The ESPnet-SE++ is an exciting piece of software, and this paper does a good job of presenting it to the community.

faroit commented 1 year ago

@joimort thanks for your review. @neillu23 can you include the reviewers suggestions (if you agree), update the paper and report back here when that is done? thanks

neillu23 commented 1 year ago

Thanks to the review from @justusschock and @joimort ! Your feedback is truly appreciated and we'll thoughtfully integrate your suggestions into the paper. Once we have made the revisions, we will send you an updated version and let you know. Additionally, thanks to the assistance from @faroit !

faroit commented 1 year ago

@neillu23 can you let us know about the status of the revision?

neillu23 commented 1 year ago

Hi @faroit , at present, I'm in the process of transforming pictures into text and making a new README.md file to minimize the requirement for explaining folder paths inside the document. I expect to complete these activities and make a PR by the end of this weekend. Many thanks!

neillu23 commented 1 year ago

I have submitted a PR at https://github.com/espnet/espnet/pull/5212 to address the feedback from the reviewers. The modifications made are as follows:

  1. Instead of including configuration figures, we have provided links to the configuration file along with brief explanations.
  2. The figure illustrating the directory structure for the Joint-task has been relocated to a separate readme file "TEMPLATE/enh_asr/README.md."
  3. Add captions for figures and remove redundant titles for the UML diagrams.
  4. The conference details in the citations have been updated.
  5. The formatting problem with the links has been resolved.

I will notify you once the PR is merged. Thank you very much!

neillu23 commented 1 year ago

Hello @faroit , I wanted to let you know that the PR(https://github.com/espnet/espnet/pull/5212) has been successfully merged! I apologize for the delay in completing it.

faroit commented 1 year ago

@neillu23 thats great. Thanks for letting me know.

@justusschock can you let us know about the status of your review? is there something left to do for @neillu23?

justusschock commented 1 year ago

@editorialbot commands

editorialbot commented 1 year ago

Hello @justusschock, here are the things you can ask me to do:


# List all available commands
@editorialbot commands

# Get a list of all editors's GitHub handles
@editorialbot list editors

# Check the references of the paper for missing DOIs
@editorialbot check references

# Perform checks on the repository
@editorialbot check repository

# Adds a checklist for the reviewer using this command
@editorialbot generate my checklist

# Set a value for branch
@editorialbot set joss-paper as branch

# Generates the pdf paper
@editorialbot generate pdf

# Generates a LaTeX preprint file
@editorialbot generate preprint

# Get a link to the complete list of reviewers
@editorialbot list reviewers
justusschock commented 1 year ago

@editorialbot generate pdf

editorialbot commented 1 year ago

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

justusschock commented 1 year ago

@neillu23 thanks for the update!

@neillu23 @faroit all my comments have been addressed. From my side, there's nothing left to do!

faroit commented 1 year ago

@joimort can you please check if there is anything left to do or if your last items from the checklists can be closed?

joimort commented 1 year ago

@faroit @neillu23 Thanks! As far as I'm concerned, there are no remaining tasks!

faroit commented 1 year ago

Post-Review Checklist for Editor and Authors

Additional Author Tasks After Review is Complete

Editor Tasks Prior to Acceptance

faroit commented 1 year ago

@editorialbot check references

editorialbot commented 1 year ago
Reference check summary (note 'MISSING' DOIs are suggestions that need verification):

OK DOIs

- None

MISSING DOIs

- 10.1109/slt48900.2021.9383615 may be a valid DOI for title: ESPnet-SE: End-to-end speech enhancement and separation toolkit designed for ASR integration
- 10.21437/interspeech.2022-10727 may be a valid DOI for title: ESPnet-SE++: Speech Enhancement for Robust Speech Recognition, Translation, and Understanding
- 10.1109/icassp40776.2020.9053512 may be a valid DOI for title: ESPnet-TTS: Unified, reproducible, and integratable open source end-to-end text-to-speech toolkit
- 10.18653/v1/2020.acl-demos.34 may be a valid DOI for title: ESPnet-ST: All-in-One Speech Translation Toolkit
- 10.1109/icassp43922.2022.9747674 may be a valid DOI for title: ESPnet-SLU: Advancing spoken language understanding through ESPnet
- 10.21437/interspeech.2018-1456 may be a valid DOI for title: ESPnet: End-to-End Speech Processing Toolkit
- 10.21437/interspeech.2020-1673 may be a valid DOI for title: Asteroid: the PyTorch-based audio source separation toolkit for researchers
- 10.1109/icassp40776.2020.9054266 may be a valid DOI for title: Dual-path rnn: efficient long sequence modeling for time-domain single-channel speech separation
- 10.1109/icassp.2018.8462116 may be a valid DOI for title: Tasnet: time-domain audio separation network for real-time, single-channel speech separation
- 10.1109/taslp.2019.2915167 may be a valid DOI for title: Conv-tasnet: Surpassing ideal time–frequency magnitude masking for speech separation
- 10.1109/icassp.2019.8683855 may be a valid DOI for title: SDR–half-baked or well done?
- 10.1109/icassp.2001.941023 may be a valid DOI for title: Perceptual evaluation of speech quality (PESQ)-a new method for speech quality assessment of telephone networks and codecs
- 10.1109/icassp.2017.7952154 may be a valid DOI for title: Permutation invariant training of deep models for speaker-independent multi-talker speech separation

INVALID DOIs

- None
faroit commented 1 year ago

hi @neillu23, sorry for my slow summer response time, I was moving to a new home. The paper is close to acceptance, can you please address the minor things mentioned in https://github.com/espnet/espnet/issues/5399 to proceed with the paper?

faroit commented 1 year ago

@neillu23 can you please update us on the status of the submission so that we can move forward?

neillu23 commented 1 year ago

Hi @faroit , thank you for reminding me, I can address the comments over the weekend!

neillu23 commented 1 year ago

I have submitted a PR at https://github.com/espnet/espnet/pull/5418 to address the comments, the change including

  1. Add references to the actual model architecture papers as the paper mentions that espnet-se++ adds "several new models, loss functions, and training recipes".
  2. Add DOIs for the references. This is required for a JOSS publication
  3. Modify title={Onssen: an open-source ...} to title={{ONSSEN}: an open-source ...}

I will notify you when the PR is merged. Thanks again!

arfon commented 1 year ago

@neillu23 @faroit – it looks like that PR is merged. Are we ready to pick up this review again?

neillu23 commented 1 year ago

Hi @faroit, what's the subsequent action we should take?

faroit commented 1 year ago

Post-Review Checklist for Editor and Authors

Additional Author Tasks After Review is Complete

Editor Tasks Prior to Acceptance

faroit commented 1 year ago

@editorialbot check references

faroit commented 1 year ago

@editorialbot generate pdf

editorialbot commented 1 year ago
Reference check summary (note 'MISSING' DOIs are suggestions that need verification):

OK DOIs

- 10.1109/slt48900.2021.9383615 is OK
- 10.1109/icassp.2016.7471631 is OK
- 10.1109/icassp.2017.7952155 is OK
- 10.21437/interspeech.2020-2537 is OK
- 10.1109/taslp.2021.3082318 is OK
- 10.1109/icassp43922.2022.9746372 is OK
- 10.1109/icassp43922.2022.9746171 is OK
- 10.21437/interspeech.2019-1550 is OK
- 10.1109/asru46091.2019.9003849 is OK
- 10.1109/icassp43922.2022.9747146 is OK
- 10.1109/icassp.2018.8462116 is OK
- 10.1109/icassp.2019.8683855 is OK
- 10.1109/icassp39728.2021.9414661 is OK
- 10.1109/icassp43922.2022.9747473 is OK
- 10.21437/interspeech.2022-10727 is OK
- 10.1109/icassp40776.2020.9053512 is OK
- 10.18653/v1/2020.acl-demos.34 is OK
- 10.1109/icassp43922.2022.9747674 is OK
- 10.21437/interspeech.2018-1456 is OK
- 10.1163/1872-9037_afco_asc_1322 is OK
- 10.21437/interspeech.2020-1673 is OK
- 10.15199/48.2016.11.70 is OK
- 10.1109/icassp40776.2020.9054266 is OK
- 10.1109/taslp.2019.2915167 is OK
- 10.1109/tasl.2011.2114881 is OK
- 10.1109/icassp.2001.941023 is OK
- 10.1109/icassp.2017.7952154 is OK
- 10.1109/mcse.2014.80 is OK
- 10.1145/2792745.2792775 is OK

MISSING DOIs

- None

INVALID DOIs

- None
editorialbot commented 1 year ago

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

faroit commented 1 year ago

Hi @faroit, what's the subsequent action we should take?

@neillu23 please check the task list above. Basically you would need to tag a new release on github and notify us here and then archieve the repo on zenodo like described above.

faroit commented 1 year ago

@neillu23 can you update us on the status of the Post-Review checklist?

neillu23 commented 1 year ago

Hey @faroit, ESPnet has released version 202308(v.202308) at https://github.com/espnet/espnet/releases/tag/v.202308. This could be the one we used for our JOSS publication. Although I don't have direct access to the ESPnet GitHub, I can help with uploading the release assets to Zenodo. Would that suffice, or should I reach out to the main account holder to get it archived on Zenodo? Appreciate your guidance!

neillu23 commented 1 year ago

Hello @faroit, we've just released the newest version of ESPnet on Zenodo and completed the tasks on the checklist. Please let me know if there are any updates required. Thanks!

faroit commented 1 year ago

@editorialbot set v.202310 as version

editorialbot commented 1 year ago

Done! version is now v.202310

faroit commented 1 year ago

@editorialbot set 10.5281/zenodo.10048174 as archive

editorialbot commented 1 year ago

Done! archive is now 10.5281/zenodo.10048174

faroit commented 1 year ago

@editorialbot generate pdf

faroit commented 1 year ago

@editorialbot check references