Nice study based on relatively limited data - potentially a key publication, opening the possibility of singing as a significant aerosol generation procedure - for WHO and other healthcare institutions to take note of.

Some suggestions to strengthen their argument:

• might be useful to include the now famous Moser et al. plane report on influenza spread amongst confined passengers when the ventilation was off - as poor ventilation is the main predisposing factor and target for intervention here: https://pubmed.ncbi.nlm.nih.gov/463858/

• include this recent modelling/empirical study that suggests that COVID-19 patients:
  https://www.nature.com/articles/s41591-020-0869-5

• the authors assume one index case that was symptomatic - but we now know that many cases in adults may also be asymptomatic - which may be realistic - especially with the above paper.

Can the authors perform some sort of sensitivity analysis to account for the case where there is more than one index case? In a sense, it may not matter much if the airborne virus achieves relatively uniform distribution during the period of the rehearsal. But the authors should discuss this possibility and the impact that it might have on the analysis - just to pre-empt any critique later.

Do we know if the incubation periods for the secondary cases all fit with a single point source exposure (admittedly this can range from 2-14 days)? Do we know the actual distribution of the secondary cases - relative to the presumed index case?

This press article contains critiques of the idea that singing is not a source of aerosols - can the authors formulate a response to these but embed it in the Discussion and relate it to their study? Again, this aims to fend off any future critiques of the study.

I think the authors have revised this paper as best they can with the limited data available to them. No further comments.