Fader adjustments from trailer to feature

[Carsten Kurz]

I doubt that distortion is the only factor here. Most movies we get if played at 7 are already too loud when the max SPL is not nearly reached. There is a natural perception of proper dialogue level for an audience. If your ears bleed already when two people on screen have a normal conversation, there is something wrong. Whenever we were able to play near 7 on our system, it was very high profile movies (count out Christopher Nolan here). Unfortunately, we play most international movies dubbed, and the dubbing often does not make things better. I recently had to create a new level macro 3.8 on our server in order to actually get a trailer to play below 4. Never thought I needed that. There is no simple or one size fits all solution, of course. An audience that has come to watch a Tarantino movie will happily enjoy his blasting gun shots. But there's always that person that came to the movies first time again since years and without any idea what to expect. Event though we play preshow at 4.5, I usually tell people to wait for the main feature when they already come out during the trailer program.

We actually do store a box of cheap disposable Honeywell hearing protection earplugs at the ticket booth for those very sensitive people. We actually give them out for free when someone complains in a friendly way, indicating that he or she actually suffers. There are movies, though, that we simply will not turn down in volume even if someone complains.

[Steve Guttag]

I did not mean to imply that speakers and distortion are the only problems but they are significantly contributing to it. If Christopher Nolan wants his movies mixed obnoxiously loud, with unintelligible dialog, no sound system will fix that (nor should it). But I've noticed this over numerous theatres...the screens with better acoustics and better speakers get to play louder without complaints.

What I like about LEQ(m) is that it takes into account the duration of the content when assigning a number. So, if you have a 2-hour feature (who are we kidding, 3-hour or maybe 3.5-hour) can have that big explosion but trailers and Ads can't just be all-loud all of the time.

What if we could, at the theatre level, in real-time, measure the content on the first pass/test-screening, determine its LEQ and then apply an offset? Then, you could set your volume once and it would be right for every title, providing you are good with having dynamics. We can already apply compressors. Smart was doing that with the "Afterburner" that Gordon McLeod seems to put into as many systems as possible...it was a straight up 2:1 compressor above reference...so untouched below 85dBc but coming down like a hammer above 85dBc. As I recall, it treated the screen channels together, the subwoofer separately and I forget if it bothered with the surrounds. Something like that could also be used for Sensory Friendly shows though most just turn the volume WAY down and turn on open captions. With DSP processors, like Q-SYS, there are a world of opportunities out there.

Using just the volume control is a blunt instrument in that it affects things not causing the problem...hence I was considering an asymmetrical fader that has a more progressive attenuation to Left and Right or, perhaps, apply a compressor to Left and Right but let Center and Surrounds come through unscathed. Despite subwoofers playing loud, I don't think people in the auditorium complain about them as much as in neighboring auditoriums.

[Mike Blakesley]

But volume is also very subjective.

That's for sure. It also depends on the sound mix for a particular movie. I'm not one of those cranky old guys who doesn't like things to be loud -- I love a good loud sound, but there is still such a thing as "too loud." But on some movies, it is the contents of the sound as much as the volume of it that can be bothersome.

[Randy Stankey]

But first there are trailers, and although it's "legally required" to meet TASA standards for the MPA, even when you're legal there, they are still obscenely loud. I went in person to a two TASA meetings in the early 2000s and was trying to get trailer levels to continue going down, but was shot down. well, ask me about this at some point; long story..

I'm going to have to pre-apologize. My reply is going to seem vitriolic. It's not aimed at you, Larry. It's aimed at those people at your above-mentioned meetings who shot down your arguments...

I don't give a flying frog leap what some idiot at some meeting of some organization whose name sounds like alphabet soup says. Those letters that come in film cans, saying "Turn your fader up to 7" all get thrown in the trash without even being read. I am going to set **MY** fader to whatever level that I believe best serves my paying customers.

Those people have never been to my theater. They have never met my customers. They have no idea what goes on in movie theaters, yet they sit around in meetings and pontificate about some bullshit reasons why they think they need to make movies so loud that they make paying customers go deaf!

They need to pull their fucking heads out of their fucking asses! If I turned my sound system up to the levels they say, my theater would be out of business by the end of the year because there wouldn't be any customers, anymore. If theaters go out of business, they go out of business, too. Right?

There's an old joke that ends, "Let me turn it up!"

If they want me to turn something up, my answer is, "Here! Let me turn THIS up!"

You can tell them that I said that, too!

[Mike Blakesley]

Yes. Much better advice in those "letters" would be: Please pre-screen the movie in your theater and set the volume appropriately for this feature. Not all features are recorded at the same volume levels (despite industry efforts to make that happen), and your audience deserves hearing it at the very best volume setting in your particular auditorium.

[Ryan Gallagher]

Everything old is new again. Just gonna put this here and run away.

IMG_6520.jpg

[Carsten Kurz]

A new record for us today - 'Grand Prix of Europe', the German trailer for a Warner animated children movie with a PG rating, computes to an LEQ(m) of 86.5. Someone is clearly nuts there.

- Carsten

[Carsten Kurz]

But first there are trailers, and although it's "legally required" to meet TASA standards for the MPA, even when you're legal there, they are still obscenely loud. I went in person to a two TASA meetings in the early 2000s and was trying to get trailer levels to continue going down, but was shot down. well, ask me about this at some point; long story..

The trouble seems to be that some people seem to think that LEQ(m) 85 TASA is a requirement. And read that as a justification to level EVERY trailer to LEQ(m) 85. But TASA requirement is that no trailer loudness should exceed that value. It is a max/complaint loudness, not a target requirement.


'At any given time, the TASA Standard identifies a specific Leq_m number as a recommended upper volume limit for trailers in order to bring the audio level of trailers closer to the features they precede. The current upper volume limit is Leq_m 85.'

https://www.tasatrailers.org/whatis.html

[Frank Cox]

a recommended upper volume limit for trailers in order to bring the audio level of trailers closer to the features they precede.

This is obviously wrong since the audio level of trailers is always far louder than the audio level of the features they precede. Which is why everyone started using macros to turn the volume down during the trailers.

Now it's likely an insoluble problem since if someone sends out a trailer today with a reasonable volume level it'll be too quiet in most theatres.

[Stefan Scholz]

It is about hardware and speakers as well, about acoustical treatment and baffle walls.
It's not about measured SPL or calculated LEQm values.
Rooms do have a certain absorption capability, if it is exceeded, the sound gets extremely unpleasant. Take a single dual 18 in woofer cabinet in a room with a 21 ft screen, and around 70 seats. You're reaching the SPL required on the analyzer. Still it is bass without any impact. Adding the second cabinet not only reduces amp power by 50%, clustering them and using a baffle around, suddenly you can feel the radiation with your body, not only measure some SPL value. Using a third cabinet, and it starts to become fun.
Some older 2" hf drivers used to have a very low linear range, around 1 Watt. Rated at 50 W (thermal before destruction) anything than 108 dB in 1 meter led to harsh distortions.
Lack of full frequency capability in front speakers, choosing a small low frequency cabinet, again, we eventually reach the required SPL, but still this will always sound too loud, as low end frequencies contain a lot of energy. If these are not measured, the setting to 85 dBC slow on the meter ends in higher SPL than wanted in usable bands.
The 7.0 rule had it's background. 70 dB at 7.0 fader is easy to remember. Most techs I saw set for 85 dBC slow on their SPL meter, and they never thought about the 70 dB for each 1/3 octave bands. Full frequency added, with rolloff, this eventually becomes 85 dBC. Think of a siren (smoke alarm). Single small frequency, and it's 85 dBC on the meter

At one of my early trainings, the guy said, especially low frequency woofers, bring as many as you can place. It's not about achievable SPL, it's about distortions. You want these to just reproduce the low end, up to 100 Hz, so you can't locate them. The moment they distort, there are overtones, mix product, and they can be localized by the hearing mechanism. Plus, added benefit, the more radiating area there is, the easier they actually move air in the room. For a single cone, the air is like a block of concrete that you try to shift with a finger. You are not needing fast excursion and low area. You are needing low velocity and large area, so it's either a lot of cones with low drive, or an acoustical transformer, that transforms fast movement low area into slow movement and high area (horn, transmission line, etc.

I have seen photos of even IMAX rooms under construction, which, for the sheer size looked like they have bough a ToysЯus setup.

I am tired of the "movie too loud" discussion. Ok, it is the case, if you're screening European TV movies, which account for most of the arthouse stuff here, with it's compressed and mastered on 8' bass monitor speakers is intended for TV, and not theatre.
But definitvely, any major brand name film can even be played in excess of 7 on the fader, and even adds extra fun. It's all about following the long ago published guidelines of dimensioning a speaker system in a venue.

[Steve Guttag]

Stefan, you are right and wrong at the same time. Yes, it is about the distortion (harmonic, in particular). The SPL meter, like the color temperature meter assign a single number but do not describe what they are measuring.

One of the problems with those that use the Radio Shack or like SPL meter is their own body and the microphone affect the reading Yeah, at one point, you measured 85dBc with a "good" mic/SPL meter and set the RS to match but it does not describe what that meter was hearing. How good is the filter? What is that mic's frequency response? If it is extra sensitive at 2KHz it will apply a bias to that frequency. Yet, the industry, after spending a period of time with an analyzer where they "trust" the RTA readings suddenly think the $29 meter with more unknowns and suffers from the proximity of the user should be the reference?

Compression drivers are tricky to get right and not sound harsh. Ribbon tweeters sound decent...until pushed. Horn/waveguide technology has improved a bit over the decades. As mentioned before, I've found more bad sounding speakers than good. A good 2-way beats a moderate 3-way, just about every time Of course I don't see much in the way of good 2-ways anymore. The JBL 4675 was a very good 2-way. It could be a bit harsh on the high end but it could be tuned to be pleasant without being pushed. I'm even good with most of the vintage Altec stuff. Most of the modern 3-ways would have been better as 2-ways. Getting three drivers to play as a team is tough. The crossover points on a 3-way make it tend to cut across the dialog range, which makes it more tricky. Having it transition from a cone to a compression driver adds into that even more. Somehow, QSC got it right on their SC-423 (3-way) and SC-424 (4-way). The SC-424 is magical in that the VHF range cuts through the screen effortlessly. It can get loud without being painful. The new Meyer Astrya 140 (3-way) also can play loud without being painful (except in the wallet). The JBL M2 is an incredible speaker (2-way) for the small screening room. It's one of the best I've ever heard and again, it can get loud without being painful.

You bring up a good point about the relationship between 85dBc and the level of each frequency band measured. 85dBc is only meaningful if ALL of the frequencies within the pass band are being properly measured. If your speaker, for instance, can't play below 100Hz, you don't boost the gain to get an 85dBc with the remaining response. The passband between 100Hz and 2KHz should be about 72dB, per ⅓-octave and then decrease as per the "X-Curve" as defined in ST202 (ISO-2969). If the response does not, then the SPL isn't valid either.

As to subwoofers...there is no generalization of how many cabinets make a system sound a particular way. The right amount of subwoofer is the right amount of subwoofer. The size of the room will limit how low the system can play. If you want to play down to 20Hz...your room has to be nearly 57-feet long to hold that wave. I made a subwoofer (customized clone of the JBL SUB 18)...that plays down to 18Hz. I had to go over 60-feet into the yard just to measure it.

How many sub cabinets are needed really depends on the subwoofer and the room it is playing in. It is almost always better to add a subwoofer system than to increase amplifier power. That 2nd subwoofer is normally a 4-fold improvement (doubles the SPL while cutting the impedance in half so the amplifier gets a better transfer function of nearly double). Geometrically, this can get out of control though. Most of my systems are 2-4 sub cabinets with the very largest theatre having 8 of them. I'm a fan of the JBL 4645C, 4642A, 5628 (give it enough power an enjoy), 5749 (it is the LF section of the 5742 but it is the 4642A box with the 2242HPL drivers...it plays deeper and louder than the 4642A). From QSC, the 7218 is may staple.

Yeah, baffle walls make a difference. It puts the soundwaves where people are listening while keeping them from going where it is wasted. You get more of your money going to the ears. A baffle wall is to a speaker what a reflector is to a light bulb. Driving your speakers less hard almost always will lower your distortion too.

[Harold Hallikainen]

Yes, it is about the distortion (harmonic, in particular).

Of course, distortion is all just nonlinearity, and nonlinearity causes intermod and harmonics (perhaps harmonic distortion can be thought of intermod distortion with the fundamental mixing with itself to cause sum and difference frequencies, getting us the second harmonic and DC. Higher harmonics would be the mixing of lower harmonics).

Anyway, I heard a very impressive demonstration of masking in a SMPTE meeting. A song was played and passed through a distortion block that added crossover distortion (such as from an improperly biased class B push-pull amplifier). The distortion was VERY audible. However, adding a low pass filter at the highest major content frequency vastly improved the sound substantially. Harmonics landing outside the spectrum where there was major content was very audible, while harmonics and intermod within the spectrum occupied by the content was masked by the content. Intermod also generated distortion products below the spectrum of the content itself, so adding a high pass filter at the low end of the content further improved the sound. So, as Steve points out, harmonics generated by distortion are often more audible than intermod products since the harmonics are not masked by the content.

Because of masking and other effects, it difficult to put a number on a system saying "this will sound better than that." We also have user preferences to consider. A user may prefer a different sound than a more accurate reproduction. We have seen this with experiments where people adjusted the equalization (away from X curve) to what they thought sounded best. Though they liked the sound, it was not the intended sound. People may also prefer a more distorted sound. I think it's fairly common to purposely add even harmonics to music to make it sound "fuller." But, I think we are after accuracy of reproduction instead of individual preferences (with the sound ideally reproducing the preferences embedded by the sound mixer). So, we MIGHT just put a number on maximum deviation from linear, but with that, a worse rating may sound better than a better rating due to masking.

[Larry Blake]

Hi all:

Thank so much for all of your feedback and comments. Much appreciated. To reply in part:

— As Mike B. said, indeed, a good thing to put in said letters-in-prints/DCPs: Please pre-screen and set to what you hear. Indeed, back in 1996 I met and became buddies with the great projectionist, Paul Rayton, when setting up for a movie I had mixed (The Pallbearer) and he gave it a good, long listen for fader level.

Obviously, you often don’t have time to audition each movie, but to the degree that you can, and don’t just unilaterally keep all features low because nobody’s complained, I can speak for me and all of my colleagues who don’t work for abusive directors (and, BTW, I point fingers only at Messrs. Bay and Nolan, and not at the very talented people who have worked for them over the years) that we’d very much appreciate a chance to be at 7.0/85 dBc/slow or as close to it as possible.

— Re: sound system quality. I agree with my buddy Steve G. and do think that this is something of a real factor in this discussion. I also agree with Steve that it’s bullshit when Mr. Nolan says that the reason that his movies have had problems in theaters, including dialogue intelligibility, is that he mixes for good rooms. Bullshit. I heard Tenet at a Dolby Cinema theater in New Orleans and Dunkirk at the Universal City IMAX, and they were abusive beyond belief, even though the sound systems could keep up with it. But I digress.

Another way to look at this is that it’s my strict understanding (and please anyone correct me on this as needed) that Dolby Cinema and IMAX theaters do, indeed, play at 7.0, pretty much across the board.

Which is to say that, given that they have, I believe, better than the average bear sound systems, they get very little complaints that I know of. Maybe it goes with “expectations” that the filmgoers bring into the theaters.

— As to TASA standards for trailers, in the U.S. at least, they are legislated by the MPA, and effectively go from first sound to last sound (with some fudge time), and the average of the program. And, yes, all trailers, Merchant-Ivory or Nolan, are mixed to push it to the max.

As implied by my original posting, all I can say to marketing trailer people at studios is boo-fucking-hoo if you all play their trailers at 3.5 and they think it’s too low. People are coming their for the movies, not the trailers.

— As to the question that do we indeed mix at 7.0/85, the answer is absolutely yes. Certainly at every mix stage I’ve worked in (L.A. and N.Y.), about 20, and my own in New Orleans. Again, you can’t second guess this standard, and I’ve heard of techs making 7.0 “82,” for what reason I can’t figure out.

Every year at the Academy’s Goldwyn theater, after it has been carefully tweaked by the theater standards braintrust, there is an evening when we can all bring a seven-minute clip of movies we did, and there are comments on the system alignment in general and the fader level specifically. Always surprised and happy to see that not only are today's mixes in general at sane levels, but the average dialogue level is rock-solid among most movies.

To me, from our end behind the faders, is the key factor is the difference between average dialogue and the loudest effects and music. Not an original observation, f’sure, but that’s the key to a mix playing the real world AND on TV, without problems. Remember, when everything was recorded and mixed to optical, there was perforce a v. limited dynamic range, and whaddaya you know, those old movies work just fine on TV even though they were mixed for 3,000-seats barns.

— Re: letters from filmmakers and requests: I know that it’s happened where filmmakers request that you go over 7.0. They need both a lesson in film sound, and maybe need to mix at a better-calibrated room. “Go to 8”: bullshit.

Re: 8; Back in 1994 I put on a memorial tribute to the great mixer Murray Spivack at the Academy’s Goldwyn theater, and we played the Academy’s own 70mm print (I think an original from 1965) of The Sound of Music. No tones, of course, and we tweaked frequency response and level by ear. Fader setting we ended up at? 8!!! I think the flux level of mag in those days was in the 90 nwb/m range, no?

Speaking of Murray, here’s a link to a talk I did at the Academy Museum in 2022 about Murray and his life, focusing on King Kong:

https://www.youtube.com/watch?v=lFzdeJi_U2I

Blah blah blah. I’ll sign off now. Let me know what fader settings you come up with for Spinal Tap 2.

[Steve Guttag]

1983 was a pivotal year for striped magnetic sound (Return of the Jedi). Flux density was raised to 185 nW/m as well as the track width was widened to 70 mils (.070-inches) from 50-mils (4-track got this too though I forget what the surround track dimensions were...originally it was 38-mils. As for level, it was all over the place...which is why every release had to make their own noise and tone. Different sounding labs had their own methods.

Even after 1983, when standards were set and both Dolby and Teccon made/offered calibrated test films, the release prints wouldn't match. When running older prints, when the calibrations films are long-gone, if you are lucky, you might get a "slate tone" on a leader to give you a clue as to the level/balance of recorder. If the film doesn't have Dolby NR (either A or SR), then it doesn't matter really. You are going to set the volume to patron preference anyway. In the case of the Sound of Music, it's one of my most favorite mixes.

With Dolby prints, one needs to listen for NR mistracking. I had to do that with a Purple Rain 35mm 4-track Dolby-A print.

As to setting up 82dBc for fader 7...yeah, I've done that (a lot)...never in a screening room or for an event where the fader setting means something. This is to avoid people getting down in the 3.5 range for anything and to move the feature level towards the sweet spot of the fader. Once you go below 4.0 (-10dB), you are are getting into the part of the fader's range that is very aggressive. It really depends on the venue (and its equipment). The realities are, and it is an estimate because I have not done real research on it, 90% of the movie going public would prefer the sound turned down 3dB (or more). Nothing stops someone from bumping it above 7 if the movie is too quiet.

As for Nolan...during the throws of C19, when Tenet was released...I was commissioning a 10-plex and they were testing the theatres getting ready for the big opening around Labor Day of 2020 and I noticed that the subwoofers never seemed to turn off. So, I recorded the power level of the subwoofers running Tenet (at 7). The subs in their big theatre have 3000-watts available each. They would get up to around 2800-watts each and there were numerous peaks over 2000-watts each. This is the subwoofer demands on that movie and note how long they were playing at extreme levels:

image.png

That is a power versus time graph. The subs start at the beginning of the movie and with brief breaks they play all of the way through. Talk about obnoxious. Most movies, the subs play some here, some there...maybe a big explosion where you'll get a peak and there may be a handful of such events for the entire movie. This one just beat you up for the whole movie.

[Marco Giustini]

I’m not a sound designer but your ears will get used and ‘dampen’ such soundtracks where it’s loud-loud-loud all the time.
the best and most impressive tracks IMHO are those where you have a small selection of scenes where the dynamic is fully used and the LFE is being pushed really hard. Because you haven’t been hammered with SPL and low frequencies for the previous 40 minutes, you experience the full dynamic and you get that ‘wow’ feeling.

Fair enough, on some movies where its ’destruction, destruction, destruction’ all the time, it must be difficult too keep the levels down lol

on LFE count, Stefan, setting the reference level won’t tell you much about the peak level. Drivers have power compression because of the law of physics, the moment you’re pushing your LFE drivers past 50%, you inevitably introduce compressions and also lose on linearity - with VLF getting behind on other frequencies. So yes, more is better but as Steve says, if you do the proper math, the right amount of subs is the right amount of subs.
the truth is that very seldom - at least with systems I’ve worked with - do I encounter the ‘right amount of subs’. The LFE channel is very often severely underspec’d

on SPL not being valid unless the frequency response is also valid, I’m glad we’re on the same page, Steve. I don’t even look at the SPL numbers when doing levels - or I look but without paying too much attention particularly if my frequency response shows some bumps. If anything, I feel the latest A-weighted references are a much better choice.