Loudness And Dynamics In Cinema Sound - Part 2 - Cinema Survey Results
In the first part of this series on Loudness And Dynamics In Cinema Sound, Eelco Grim gave his permission to share a paper he presented in May 2013 outlining the state of cinema sound levels and dynamics and how the Dolby calibrated system has completely broken down with films having to played back at a much lower level because they have been mixed and mastered to be so much louder. Eelco presented a well-reasoned proposal for a plan that would use the BS1770 loudness standard and require no investment in hardware, just a reprogramming of the Dolby boxes, but here we are 4 years later and nothing much has changed.
Steven Ghouti AFSI AES has worked in film sound for some 23 years, decided to take it upon himself to put out a survey to his fellow audio mixers in the film sound workflow to establish answers to a number of questions regarding the level we mix at, the size of rooms we mix in. Over to you Steven...
In the time I have been working in sound for film, I have seen that dynamics have gradually disappeared and good practices have all but dissolved in what used to be a standard-driven industry. The catch-all phrase “ Dolby was here just last week ” - which used to mean that the theatre you were in was grossly misaligned - is not even used anymore as the advent of Digital Cinema has in some cases turned theatre installation and tuning into a last-minute rushed job that is rarely re-visited.
For years, Re-recording (or Dubbing) Mixers had to contend with pre-show content being mixed very loud (commercials and trailers with a max loudness of 85 Leq, usually mixed up to 8 or 9dB hotter than the movie), which forced projectionists to turn the B chain level down, sometimes not to be turned back up again.
Today, playback levels from DCP servers can be programmed, but the levels often stay well below the once-standard Dolby Level 7 because the actual movies are often mixed too loud, and cinemas prefer to play at lower levels rather than risk being on the receiving end of a complaint. A knee-jerk reflex emerged over time: to mix louder to counter the fact that the cinema might lower the playback level. This led in turn to more complaints and more projectionists lowering the level further. And so Producers/Directors/Mixers feel the need to mix even louder...
This never ending cycle is causing havoc in film sound all over the world. People used to accuse most US-mixed blockbusters of being the loudest films of all, before starting to mix at lower monitor levels themselves and pushing out very loud mixes with reduced dynamic ranges. As you may have read in part 1 of this series, in a 2013 AES paper, Eelco Grimm did a very good job of explaining what was happening to loudness in cinemas.
The reason I set off on this personal endeavour is to be able to get movie playback levels back to Dolby Fader 7, with sensible mixes (sensible does not mean flat, dull, or whatnot, it just means that the movie plays back in an enjoyable manner to a broad audience when the playback is set to 7 in the theatre). Now I know that this might be a pipe dream, but I prefer that we all get our heads together before we get hit by legislation that could impact the way we work, and hobble our artistic intentions.
I happen to passionate about my job, like many people in this business. I do not want to go with the flow, and let cinema-goers end up watching movies mixed so loud that they have the dynamics of an internet podcast played back at 2 on the Cinema Processor. I believe that filmmakers have a responsibility toward their ticket-paying audience to show them a movie made with an attention to detail, dialogue intelligibility, overall levels and dynamics and that those same filmmakers need to be able to be trusted again by cinema owners to deliver content that adheres to a common playback standard.
I fear that if we do not act altogether, cinema mix levels will soon be subjected to stringent Laws and Rules that will be totally contrary to artistic freedom that one has when mixing for film.
I am also part of a workshop with the French sound associations AFSI and ADM. I have presented my results at a CST seminar in June, and Eelco Grimm will be using the results in an upcoming AES presentation. However, as I stated above, even though I am an active member of AFSI and the AES, this survey is not done on their behalf.
So back to the survey, I accept that there is not much science behind the questions. They are based on the points I believe have an impact on the way we mix, and that may be pushing a re-recording mixer to choose a non-standard mix monitoring level. For anyone who didn't take part in the survey, here are the questions I asked together with the reasons behind each question....
What is your MAIN mixing expertise area? - I thought it would be interesting to know if Film Mixing is the main occupation of the Mixers who answered.
What country do you operate in? - This will allow a region by region analysis to see if there are regional differences.
How would you describe the budgets of the films you mostly work on? This question requires an answer on a scale from 1 (very low budget) to 10 (Hollywood blockbuster). I left it as a relative scale, as a US big budget would not be the same as a German one, but what is important is how the Mixers approach mixing levels for different budgets.
If you work primarily for ONE mixing facility, what is the approximative size of the room? - Room size is one the first reasons invoked for mixing at different levels.
If you work primarily for ONE mixing facility, what sort of screen channel speakers do they use? - Studio Monitors (ie. soft dome tweeters) will react very differently to sound vs typical cinema speakers (horn), as with more modern approaches like the Meyer Sound speakers
When mixing feature films for Cinema (not trailers), what is the monitoring level in the mix room during pre-dubs/premixes? - Some Mixers prefer to pre-dub a bit hotter (lower monitoring level) so that the pre-dubs will retain their integrity when mixed with other elements (ie. Music brought in later in the mix)
When mixing feature films for CINEMA (not trailers), what is the Monitoring level in the mix room during Final Mixes? - This is the level for Final Mixing and Printmastering
When mixing, do you offset the Monitoring level to compensate for the size of the mix room / mix stage? - Related to a question 4, but relevant also for freelance Mixers who go to smaller studios to mix.
When mixing for CINEMA at the level stated above, how loud does your Mix feel to you in the room? - It’s important to know if - for a certain monitor level, mixers feel that they are mixing too loud or too soft.
When mixing for CINEMA at the level stated above, have you ever used ear protections while being actively engaged in the mix? - Designed to get the myths out of the way
When mixing for CINEMA at the level stated above, have you ever mixed loud scenes with the DIM button enabled? - Intended as another myth buster
How often do you take breaks in a normal work day? - Not directly related to the mix level, but a good indicator of how we take care (or not) of our most precious tools: our ears
When mixing for CINEMA, do you have an approximative loudness target for dialogue? - This was to see if there was a trend starting about measuring or aiming for a non-broadcast target for dialogue, on top of using one’s ears of course.
If the Monitoring is set at a level other than Dolby Fader 7 (85dB), what is the reasoning behind it? there were ready-made answers as well as free ones.
Do you go to a cinema to watch the movie after release?
If yes, how do you rate the sound level of the movie in the cinema compared to the level of the mix stage? These two questions were out of pure curiosity, as we all know that cinema experiences are as varied as there are cinemas.
I would like to take this opportunity to give big thank you to all the mixers who responded to this survey. We got 387 responses in total from all over the world, and from mixers working on all sorts of movies (from ultra low budget to Hollywood blockbusters). The survey was voluntarily anonymous, so those responders would not feel inhibited in their answers by having their name tied to their monitoring level.
The full set of responses can be found here (some responses have been edited to conform to a common set ie. United States has been edited to USA). Below is summary of the responses using the Awesome Table widget...
A Summary Analysis Of the Survey Results
The results displayed above show the worldwide responses but there are some interesting regional variations. Because the number of respondents outside US & Canada and Western Europe (the UK, France and Germany) are so much smaller, Mike Thornton feels we do need to be careful about any conclusions we draw from those territories.
As you can see in the above table, 35% of Mixers, who responded, mix at 85. It also shows that mixers in US and Canada are more likely to mix to 85dBC. However, the data also shows that 82dBC also seems to be a favourite with 24% of US and Canada responses. Further drilling down into the data and linking high budget rated 7 to 10 (where 10 is really a blockbuster budget) to average mix level, shows the average mix level is for high-budget movies is 85dBC if you disregard one questionable answer. That said, 22% of the US and Canada responders have used ear protection while actively engaged in mixing.
Moving onto the reasons that were given for using a lower monitor level, take a look at this table which is for the US and Canada responders....
The next table shows the responses to the same question for Western Europe, specifically UK, France and Germany...
Next let's take a look at the area of expertise...
There is a significant proportion of responders whose main are of expertise is mixing TV/broadcast, but Mike Thornton asks are they mixing in film rooms? Mike wonders if the size of the room they work in and the type of speakers they use might shed some light on this.
Approximate Size Of Room - 35.8% of responders mix in what could be called a very small room when compared to theatrical standards, and 26.9 % in a small mix room (120 to 250 cubic meters, my own room falls into this category, at 200). The first group would clearly not have been approved by Dolby at the time when Printmastering for 35mm required a Dolby License, and some if not most of the second group would also be in that situation. It would appear that with the advent of Digital Cinema, people are now able (for good or for worse) to mix for Cinema in very small rooms.
Screen Centre Speakers - Out of the 359 responses to the screen channel question, over 50% reported using studio monitors. If one looks at the very small room responders, this figure jumps to 81.9%. This is, of course, natural as cinema speakers tend to be too aggressive in smaller listening environments, but shows that many films are being mixed on systems that bear no resemblance to the equipment used in cinemas.
It may be that the area of expertise, room size and screen speakers are all linked and Mike wonders if TV/broadcast mixers have skewed the film workflow results, which was the primary reason for Steven to create the survey.
Steven's Conclusion And Proposed Solutions
It must be remembered that this survey, though quite successful at covering a good deal of countries, remains limited in the number of responses it received. As such the responses are only a sample and might be subject to some bias because someone concerned with mix and playback levels is more likely to answer than someone who is not.
With that caveat, there are a number of key elements that stand out...
If big budget movies are still mainly mixed in large mix rooms, at a level not too far off 85, many, many small-to-medium budget movies are mixed in small rooms, with generally lower monitor levels (in part to compensate for room size).
Many of the aforementioned mixes take place using studio monitors which do not necessarily translate properly to cinemas in terms of dynamic response. Many theatres will end up exhibiting a mix that was deemed pleasing in the mix room, but once played back on average cinema speakers will turn out to be quite aggressive sounding to the audience.
There are suggestions to implement a system close to EBU R-128 and its various applications. Personally, I disagree with this, as that kind of system imposes a target value for overall loudness for a project. I would argue that for a TV product that needs to fit into a broadcast workflow, with content often used as background sound in a living room that might work, but not in an area where many projects are driven by artistic sense rather than by Loudness Meter numbers.
Cinema is a storytelling media. Many would agree that one key element to telling a story is dialogue. With that in mind, I propose the following (very broad) guidelines for ensuring that films are displayed as intended in cinemas:
A Dialogue Loudness measurement, akin to Dialnorm, would be the anchor for the main playback level. The cinema would have a written playback value (still ideally 7), or metadata for automated cinema playback systems for each movie.
The overall content would be rated depending on overall dynamics so that the audience can be warned that some or many loud passages may occur. These dynamics would be relative to the Dialogue measurement.
At no moment would filmmakers have to hit any form of target, be it Dialogue Loudness or Overall Dynamics, leaving full creative freedom.
With this system, all films would be displayed with the dialogue at what could be considered the proper level for the audience. Films could still have wide dynamic ranges, which is one of the reasons we love mixing for the film in the first place. In the long run, dynamics could come back into movies, with mixers and directors coming back to peak and valley soundscapes, always keeping in mind that a mix needs to show some form of robustness in any case to survive the wilds of cinema background noise like air conditioning and audiences.
Mike Thornton's Observations
Steven and I are in complete agreement on a desire to have a system that works for everyone, enabling the audience to experience the film as the director conceived it and heard it in the dubbing theatre, at a level that the audience is comfortable with. When it needs to be loud it should be loud as long as it isn't unpleasantly loud for extended periods, and to have the dynamic range to be able to make it quiet when it needs to be quiet.
However, I have to take issue with Steven's observation that the using a BS1770 based system similar to the EBU R128 specification would not work "in an area where many projects are driven by artistic sense rather than by Loudness Meter numbers". I think those working on TV drama productions would say that if anything the move to loudness normalisation from peak normalisation has been a very liberating process and that loudness workflows, far from restricting their creativity, have opened up creative opportunities and speaking personally I have definitely found both a liberating and supportive workflow for broadcast content and could see it working very well in a cinema workflow.
The scenarios Eelco described in part 1 of this series show that dynamics on many film mixes have been squeezed and there has been a fight to make the mix louder and louder and again. Steven and I are in complete agreement on problem, after all it is why Steven took the trouble to set up this survey to try and get some data on what is going on. Thank you Steven for doing this and I trust that your survey will help in the proces to rresolve the loudnes and dynamics of cinema sound.