Investigation of the Appropriate Reverberation Time in Learning Spaces for Elderly People Using Speech Intelligibility Tests
Round 1
Reviewer 1 Report (Previous Reviewer 3)
The paper analyses the influence of reverberation time on message intelligibility in classrooms for elderly people. For this purpose, the authors carry out different speech intelligibility tests comparing the scores obtained by elderly people, whose hearing deterioration is due to natural causes of age, with young people without hearing problems. The authors highlight the need to adapt the design and regulations of this type of architectural spaces according to the age of the students.
The manuscript addresses a useful issue, providing sufficient evidence for more restrictive standards in the design of educational spaces for the elderly people. However, some parts of the document should be improved or clarified.
· Throughout the text, mention is made of rooms A-L, C and D. Clarify.
· Line 127. The authors mention that the tests are done at "a university in Cheongju." Which one?
· Why include two rooms if, despite having different capacities - which affects its acoustic behaviour - only one will be used?
· It is somewhat confusing to use RTmid and RT60 indiscriminately.
· All references should be included in “References” section (table 15).
Author Response
Dear Reviewer,
We appreciate for your valuable comments on our manuscript.
Responses to comments can be found in the attachment.
Please see the attachment.
Authors
Author Response File: 
Author Response.pdf
Reviewer 2 Report (New Reviewer)
The speech intelligibility by elderly people is studied in the paper. The general idea and methods are very clear. In my opinion some aspects should be explained.
First of all, as noted in lines 35-38, a lecture room should be comfortable for a speaker as well. But this point is not considered.
The second point related with the first one is connected with the formal result. As stated in the paper, The reverberation time should be shorter than 0.8 s. Therefore, anechoic conditions seem to be suitable. But very short RT, for example 0.1-0.2 s, causes non-uniform sound pressure distributions, so people seating at the backside could meet low SPL. Beyond that the speaker in an anechoic room has to strain her/his voice. So, I guess the anechoic conditions are not good for lecture rooms without a sound amplification. Why don't you limit the low values of the reverberation time?
You considered only uniform absorption distribution in the room and only one point in the audience. Is it enough to generalize obtained results?
The impulse response in Figure 3 shows that the time delay between the direct sound and the most intensive reflection is about 40 ms. But the ceiling and walls have the same acoustic properties, so in the considered room the most intensive reflection is given by the ceiling and its delay should be shorter. How can you explain the 40ms delay? By the way, the time range 0-0.5 s would be better in this figure.
Author Response
Dear Reviewer,
We appreciate for your valuable comments on our manuscript.
Responses to comments can be found in the attachment.
Please see the attachment.
Authors
Author Response File: Author Response.pdf
Reviewer 3 Report (New Reviewer)
Dear authors,
The introduction should be improved, there is also C50 for speech understanding.
Explain better how you created the Odeon model, why did you choose this type of detail, the tables and chairs could be simplified, did you carry out any sensitivity tests?
How did you introduce scattering (tables and chairs)?
How did you choose the absorption coefficients?
Have you compared the simulation results to a real case?
Have you performed the calibration of the theoretical model?
Have you done the virtual listening tests, did you compare them to those of other researches?
Because you didn't test in face-to-face classes with pupils (see Sukaj, et al. Use of masks inside the classrooms, INTER-NOISE 2021).
You should better explain how you performed the listening tests and discuss the results better.
If you have compared your results with those of other researches.
Thanks for your patience
Author Response
Dear Reviewer,
We appreciate for your valuable comments on our manuscript.
Responses to comments can be found in the attachment.
Please see the attachment.
Authors
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report (New Reviewer)
In my opinion the revised paper can be published.
Reviewer 3 Report (New Reviewer)
accept
This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.
Round 1
Reviewer 1 Report
The article buildings-1904442 investigated the speech intelligibility of elderly group for various reverberation time. The main findings are two fold: the test cores is lower with increasing RT for all subjects and the elderly people performs worse than the normal hearing people - both are widely known in the community. The authors suggested that the RT should not exceed 0.8 s, which as a guideline is cannot be generalized. This study can be regarded as a case study so has a quite limited novelty as a paper.
Author Response
Dear Reviewer,
We appreciate for your valuable comments on our manuscript.
Responses to comments can be found in the attachment.
Please see the attachment.
Authors
Author Response File: Author Response.pdf
Reviewer 2 Report
The article deals with a very interesting topic, although the inconsistency of some of the results obtained casts doubt on the methodology used. The results of the PBW test are implausible, and no explanation is provided for the reasons for this anomaly.
Apart from this, which is the main reason why the rejection of the article is proposed, there are many more deficiencies in the text, such as:
- English language and style are not good enough for a scientific publication.
- The nomenclature used to designate the reverberation time should be unified. Throughout the text, RT60 and RT30 are used, while in table 6 it is denoted as T30. Something similar happens with the Deutlichkeit parameter, in table 6 it is denoted as D50 and is expressed in natural units and in the text it is denoted as D and is expressed as a percentage. I personally recommend using the terminology from ISO 3382-1.
- In table 7, "PB word list in Korean" cannot be considered as equipment. "TASCAM HD-P2" is not a "digital stereo audio simulation software". The expression "(polar pattern)" does not provide any information.
- Figure 4 is conceptually incorrect.
- Although it does not have an important consequence for this investigation, the CAD model used has an excessive degree of geometric detail for simulations with a GA software.
- It is not clear how the emission level of the sound source is established in Odeon. When the text says that "the output level of the sound source in the virtual sound field was set to 72dB(A)", is it a sound power level value or is it the sound pressure level at a given point?
- It is also not clear how the level produced by the headphones is controlled or how the background level is measured in them. More information on this should be provided.
- The title of Section 3.2 is not related to the content of that Section.
- etc.
Author Response
Dear Reviewer,
We appreciate for your valuable comments on our manuscript.
Responses to comments can be found in the attachment.
Please see the attachment.
Authors
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper analyses the influence of reverberation time on message intelligibility in classrooms for elderly people. For this purpose, the authors carry out different speech intelligibility tests comparing the scores obtained by elderly people, whose hearing deterioration is due to natural causes of age, with young people without hearing problems. The authors highlight the need to adapt the design and regulations of this type of architectural spaces according to the age of the students.
The manuscript addresses a useful issue, providing sufficient evidence for more restrictive standards in the design of educational spaces for the elderly people. However, some parts of the document should be improved or clarified.
· Line 42-44. Some references should be included.
· Figure 1. Larger figure and font size. Clarify legend in caption (although it is implicit in the previous paragraph).
· Section 2, first paragraph. Only RT60 is mentioned, but the paper contains, for example, STI. It could be indicated in this paragraph.
· Section 2.1. Nothing is indicated for class C. Why?
· Because the tables and figures appear before they are annotated in the text, it is somewhat confusing to understand them. For example, table 5 part 2.
· Lines 161 to 178. It seems that concepts are repeated several times.
· Line 170-171. The authors use the same absorption coefficients for all frequencies. Can this significantly affect the result in a real environment? Make some clarification of the effect.
· Line 201. “2) Outputting announcer voice from virtual sound using ‘persons’ sound source.” What is the purpose of this second voice? Clarified.
· Figure 4. Improve the scheme (size and quality).
· Lines 245 to 252. Better clarify the position of source and listener (distance between them).
· Section 3.1. Why hasn't it been tried with a woman's voice? would it change the results? Their spectral content is different, so it could mean an improvement or a worsening of intelligibility.
· Section 3.2. Are the different word groups balanced?
· Line 311 to 315/ Figure 7. Rename the subfigures with a, b, c for better understanding. In the text, reference them as 7a, 7b or 7c.
· Lines 321 to 329. Since we are talking about a preliminary trial, it may be interesting to include somewhat more detailed results.
· Table 10. Move after the text where referenced.
· Section 3.4. Explain the test in more detail. Do listeners have to write down the words? How do you know if they make a mistake?
· Section 3.4. All test procedure should have been randomized in order to avoid some bias.
· Figure 8. Change numbering 1 and 2 to ‘a’ and ‘b’ and modify in text. Reference as 8a and 8b.
· Figure 9. Improve the figure (size and quality).
· Table 12. Include during the paragraph the age range and remove the table, it does not provide much information.
· Section 4.1. / Table 13. Better clarify the use of two different rooms and their influence on the results.
· Line 401. Capital letter after [37]: "...[37]. Since ..."
· Section 4.2. Explain in a simplified form the procedure for obtaining Intellibility Test Score.
· Figure 10 and 11. Review size and quality.
· Lines 416 to 418. Starting from the 0.4 s reference? Clarify.
· Figure 12, 13 and lines 430 to 433. Evidently there was some problem with the results at 1.2 s. Indicate possible causes and comment in the text (not only in the conclusions).
· Lines 438 to 440. Demonstrate and justify the change to 32.4%. How is calculated?
· Line 453. A dot lost before “Therefore”.
· Figure 13 number repeated. The second ones should be moved after the lines 458 to 462.
Author Response
Dear Reviewer,
We appreciate for your valuable comments on our manuscript.
Responses to comments can be found in the attachment.
Please see the attachment.
Authors
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Thanks for the revision
Even after this minor revision, I think the scholarly novelty is lacking as a full paper, so I will defer a decision to the handling editor.
Reviewer 2 Report
I thank the authors for the effort made to modify the article. In my opinion, the changes made have significantly improved the text and have resolved most of the shortcomings of the original version.
However, in my opinion, the results of the PBW test are totally implausible and suggest that some serious methodological flaw has occurred. For this reason, and since the authors say that the PBW test is much less important than the CVC, I think these results should be removed from the article.
