@Ed209 I'm not attacking you - I'm disagreeing with you. You have said that you know better than I do what affects my tinnitus and better than another forum member what affects their tinnitus. You have also cited no evidence at all for any of your claims. A seat near the back at the average classical concert is going to be quieter than a seat near the back of a surround sound cinema, where you will never be far away from a speaker. If you disagree with that statement then provide evidence that it's wrong.
There's a New Zealand study which references their cinemas.
Here's some excerpts:
The results of similar surveys carried out by Inhe et al [5], Fiumicelli [6] and Ferguson et al [7] measured LAeq values of 73dB - 85 dB. These values are consistent with the results of this project. Interestingly, none of the surveys measured any movie with an overall LAeq above 85 dB.
Conclusions
Despite claims in media reports, measured noise levels in 17 separate screenings show that even for the duration of the movie,
85 dBA Leq is not exceeded. As an 8-hour exposure, levels are less than 80 dBA LAeq. One event over 140 dB (peak) was recorded, with other results generally well below the threshold at which hearing damage is likely to occur.
In conclusion, it is considered that going to the cinema is unlikely to cause irreparable damage to hearing, nor is it likely to result in short term damage.
https://www.acoustics.org.nz/sites/www.acoustics.org.nz/files/journal/pdfs/Ellerton,_D_NZA2002.pdf
From my measurements, current UK standards are slightly lower than these, but even this study confirms what I already know.
This is also well worth reading as it covers the psychoacoustic side of human perception.
https://www.google.co.uk/amp/s/film-mixing.com/2016/05/02/mixing-for-broadcast-loudness/amp/
Here's an excerpt from a study on an orchestra:
An orchestra model in Odeon
A model of an orchestra was built in Odeon, tuned to match the Queensland data referred to above, with the purpose to study the foreground-background balance, FBB=FGL-BGL, as motivated for above.
The orchestra was divided into 4 sections, strings, woodwinds, brass and percussion, and a single string instrument musician. Each section and the single musician was modelled as surface sources having internal power balance equal to the one calculated from the average of equivalent a-weighted sound power levels, a total of 401 samples from the concert hall in QPAC. Calculated sound power levels per section, normalized to the total sound power of all sections, are presented in Table 2. The "High" and "Low" columns is shown in order to get an idea of how the internal power balance can be different in loud and quiet sessions respectively. The power balance of the sections in the model was tuned to be equal to the values in the "Average" column, namely the average of 401 equivalent levels in Queensland. Related to this normalization, the power level of the single string player is -21dB in the average case. At forte play, the total power level is calculated to be 8dB higher than at the average equivalent level, being 114dB and 106dBA, respectively. Note that brass is much more dominant at tutti forte than in average session periods. While strings, woodwind and percussion are 5dB more powerful than average session period, brass is 13dB more powerful at tutti forte. Forte power levels are from [1], with references to [25] and [27].
In relevance to dynamics, it is worth mentioning that LAeq levels in 60 second periods, near musicians ear, during rehearsal and concert session, varied with a standard deviation of s=7dB as a mean over different instruments [24]. Smallest variation was found in Oboe/French Horn section (s=5dB) and biggest variation in Horn section (s=9dB). During each 60-second period, standard deviation over the orchestra was as high as 7dB in some periods and as low as 3dB in some periods, with a mean value of 4dB. For more than 12% of these periods, the maximum difference between two instruments was 20dB or more.
http://www.akutek.info/Papers/MS_Orchestra_Musician.pdf
Citing studies in this way is largely pointless, though. There are too many variables at play.