quote: Harold Hallikainen

---

During the mix, were speakers behind a perforated screen?

---

Yes, absolutely. But not on Foley stages I have been on. If I can track down my buddy that was a lead mixer at Todd-AO and later Sony I can find out fomr him what if any curve was used. I suspect there was an X curve though. They were simulating theater sound after all.

Mark
 

---

Posted by Steve Guttag (Member # 268) on 12-10-2018, 03:18 PM:
 
Harold,

That looks like an interesting paper that I don't have the time, at the moment to read.

How many "cinemas" were used/measured for RT60 to draw the conclusion? I can tell you, from decades of experience in measuring a LOT of rooms, it is all over the place and you can see it alter the response on the HF.

The one RT60 plot I caught while skimming the paper showed a well behaved response in the 250Hz area, which tends to indicate decent insulation above the ceiling. For those that don't, I typically see high RT60 in the 250Hz area. I also see quite a few cinemas with high RT60s in the 2-4KHz area.

Again, I think the paper is fascinating and I'm sure is well worth the read.
 

---

Posted by Harold Hallikainen (Member # 5405) on 12-10-2018, 04:13 PM:
 
Again, thanks for the comments! From http://www.screenexcellence.com/downloads/SMPTE_TC_25CSS_B_1_oct_2014.pdf#page=15 , six auditoriums were tested: Three commercial theaters, one "reference theater," and two dubbing stages. I was there for the tests on dub stage F.

Glenn Leembruggen did an amazing job writing this report. He did all the data analysis (from raw data submitted by the "Zero Dark Thirty" field testing team), graph generation, and report writing. The rest of us provided editorial comments, but it's really Glenn's work.

Harold
 

---

Posted by Marco Giustini (Member # 4544) on 12-10-2018, 04:44 PM:
 
Very interesting indeed, I shall read that too.
 

---

Posted by Chris Erkal (Member # 10646) on 12-11-2018, 03:34 AM:
 
Ok, there is some homework for me to do... Just to get to the same level, allow me a couple of questions:

First of all, I've read this article: https://www.smpte.org/sites/default/files/files/X-Curve%20Is%20Not%20An%20EQ%20Curve.pdf
It states that the x-curve is not applied to the mix, but is a way of measuring/normalizing cinemas "to match the sound heard by the director in the mixing room. "

-> The idea is to simulate a certain ideal room frequency-wise so you can hear how it should sound in the cinema (if the cinema was designed and tuned to match the x-curve), right?

Harold: -> You stated that one could listen flat and then "apply a complementary high frequency boost". I understand that this is a highly individual process as every rooms different etc. But to understand you right: How exactly do I have to understand your proposal with regard of the statement of the x-curve not being an eq curve?

Or more generally asked: If I'd be doing a mix on music monitors without an eq on my monitors in a normal music mixing room, will a complementary high requency boost generally improve or decrease the result?

I will print another test dcp for the small cinema here in Düsseldorf. I think about applying the x-curve in some passages and compare wether it improves the sound or not.
Hopefully I can measure the cinema to find out how it is tuned and how applicable my results are (and then hope for the best).

Best regards,
Chris
 

---

Posted by Harold Hallikainen (Member # 5405) on 12-11-2018, 04:38 PM:
 
As mentioned, not everyone agrees with my views on the X-curve. Looking at Michael Karagosian's paper, we see that its copyright date is 2013, a year before the theater testing report was published, so Michael did not have the latest measurement data.

Michael's paper mentions air attenuation. Note that the level of all frequencies decrease by 6 dB every time you double the distance from the source due to the sound being spread over a larger area. Imagine a light bulb at the center of the sphere. If the light bulb emits 1 watt and the sphere has a radius of 1 meter, the "power density" will be 1 W / surface area of sphere = 1 W / 12.6 m^2 = 79.6 mW / m^2. Doubling the radius to 2 m, the surface area becomes 50.3 m^2, so the power density is 19.9 mw / m^2 or 1/4 that we got at 1 m. 1/4 power is a decrease of 6 dB. So, this is not attenuation, it's just spreading the power around. Further, spreading the power around is independent of frequency.

However, since air is viscous, there is attenuation where some of the sound gets converted to heat. The amount of attenuation depends on the distance (how much air is passed through), the air temperature, and the humidity of the air. A nice table showing this is at http://www.kayelaby.npl.co.uk/general_physics/2_4/2_4_1.html . At 10 kHz, attenuation is between 95 and 280 dB/km depending on humidity. If we average the values for 10 kHz, we get 167 dB/km or 0.167 dB/m.

Looking at 1 kHz, the average attenuation is 6.1 dB/km or .006 dB/m.

So, if we go back 30 m, as suggested by Michael, we expect 1 kHz to be attenuated 0.18 dB, and 10 kHz to be attenuated 5 dB, putting 10 kHz about 5 dB below 1 kHz, just as Michael points out.

30 m is pretty far. And, if this is 2/3 of the way back, as specified for the measurement position in SMPTE RP 2096-1, the auditorium would be 45 meters deep (about 148 feet). Big auditorium!

So, while there is indeed some high frequency attenuation by the air, it is probably pretty small in typical auditoriums. Also, I do not think we want to compensate for this loss in equalization. If we were to compensate for it, those sitting in front of the reference position would have excessive highs. Further, we are used to having less highs with distance, so our brains compensate.
I agree with Michael when he says “Notably, neither reflections nor absorption should be addressed with equalization.” We are interested in the frequency response of the “direct sound.”
“Thirty years ago, when attempting to devise a technique that could easily be deployed in cinemas around the world, it was decided that readily available realtime analyzers could be used if one could take into account the effect of room reverberation. Out of this thinking was born the X-Curve sound measurement method.” However, the B-chain report shows that a modern cinema has very little reflection above 1 kHz, so room reflections do not contribute to an RTA reading above 1 kHz. The B-Chain report also reveals this when it finds that short window impulse response measurements (which reject reflections) closely match long window measurements (equivalent to RTA readings which include reflections). Further, near field measurements (very close to the screen) showed the same X-curve. The difference in the far field is likely due to air attenuation (0.167 dB/m). In many theaters, this would be negligible. Therefore, I believe the ideas that X-curve at a distance results in flat at the screen and that it is necessary to attenuate the highs as measured at a distance to compensate for buildup due to reflections (there is very little high frequency reflection) are not correct.
“Users that do not understand the proper application of X-Curve may sometimes treat it as an equalization curve, attempting to equalize the system until an exact replication of the curve as published in ST 202 is achieved, ignoring the response window described by the text in the standard. This generally leads to excessive equalization, and not necessarily the best possible sound.” I believe we do not want to exactly replicate the curve because it would use excessive filters. Fewer filters sound better than exactly matching the frequency response. We want to, instead, equalize the trend instead of each peak and dip in the response (high frequency peaks and dips are probably due to local reflections from seats, the floor, etc.).
“In addition, the X-Curve measurement method is standardized in SMPTE ST 202 and ISO-2969, and methods for equalizing a room, based on first arrival sound are not standardized, causing temptation to discount the use of first arrival methods in cinema. But, in fact, many skilled technicians prefer to use first arrival methods, and proper use of such methods should lead to equivalent results.” First arrival (impulse response with short time windows to reject reflections) are now standardized in SMPTE RP 2096-1. As noted in the B chain study group report, at high frequencies, there is so little reflection in a typical cinema that short window impulse response and RTA measurements are very nearly equal.
“SMPTE established a new Technology Committee, TC-25CSS Cinema Sound Systems. As the name suggests, members of the committee will be evaluating other measurement methods for cinema, based on first arrival of sound. If one or more alternative cinema sound measurement standards should emerge, one should keep in mind that they are an alternative, and not a replacement, for ST 202. X-Curve, the method that takes into account room reflections, and newer methods based on first arrival of sound, both strive for the same thing: a measurement method that allows the cinema technician to best approximate the sound heard in the mixing room.” At this point, we do have the new recommended practice (RP 2096-1). However, we still are using the X-curve as defined in ST 202 because the HF rolloff is not to compensate for room reflections (as revealed by the B chain study group report), but, instead, is (in my view) de-emphasis provided by the screen. I believe those creating content boost the highs so the content sounds good through the screen on the dub stage, thereby also making it sound good in the theater.
Again, not everyone agrees with the way I see it, but I THINK what I’m saying is based on the latest evidence.

Harold
 

---

Posted by Steve Guttag (Member # 268) on 12-11-2018, 05:18 PM:
 
Harold,

(and I still have not read the 200+ page report) I think "typical cinema" is too broad a statement. You had what? 4 cinemas in your test sample? Are they to represent the broad spectrum of cinema construction throughout the world? I'd dare say you don't even represent the broad size of cinemas let alone construction and RT60 to make such a statement.

If you feel that the screen is the "cause" of the X-Curve, did any of the dub stages have speakers not behind the screen? Did anyone measure the response with the screen removed and did the response magically pop up to flat (all else being equal)?

Thus far, what I've seen of the study, it seems to show that using traditional pink noise yields similar results as transfer function when it comes to evaluating the response of the system. So why bother with the more complicated set up? Sure you get more data, but what is the typical cinema tech going to do with it?

Note, we are in complete agreement on not over-EQing a room and that once we are within the tolerance (+/- 3dB) you are there. Trying to EQ to the perfect line often results in a harsh and poorer quality audio.

-Steve
 

---

Posted by Harold Hallikainen (Member # 5405) on 12-11-2018, 05:30 PM:
 
Thanks Steve. A quick response on "no screen." Dub stage F has a woven screen and has to electrically apply HF attenuation to meet the X-curve. That was the dub stage I was at during the testing.

Harold
 

---

Posted by Steve Guttag (Member # 268) on 12-11-2018, 06:10 PM:
 
I suppose that I'll have the opportunity at home pretty soon to make this evaluation myself. I'm also using a woven screen and will endeavor to measure the room before putting the screen up and repeat after putting the screen up. Mind you, I think my home set up will be on the very low RT60 side. In fact, I'm concerned about it being too low and plan to make RT60 measurements before the wall fabric goes up to ensure it isn't too dead and adjust the insulation on the walls to adjust the RT60 of the room, presuming the fabric over the insulation will not be as significant a contributor. The speakers under test should be relatively flat (they are rated from 20Hz to 40KHz). Close micing them, they are spooky flat from about 30Hz to 16KHz.

I plan to have a Smaart system by then but I do most of my tuning with the D2 analyzing system, which has proven itself time and time again.
 

---

Posted by Chris Erkal (Member # 10646) on 03-04-2019, 01:45 PM:
 
Ok, just really quick on what I ended up with doing:

I had a monitor path in my daw which rolled high frequencies off (with a combination of shelve and high cut filters) as well as the lows creating some x-curve-like feeling. Also there was a high mid boost for a more horn loaded speaker sound.

I checked the mix in a bigger cinema here in Düsseldorf to see if everything was somewhere where it should be. There is a final display before premiere - I'm excited to hear how it went.

Since the monitors at my studio are pretty close and I even had to do some work on headphone (yikes...) there was pretty much imgaination involved. Still it's a big learning process and your brain needs to rewire for the cinema, but I'm glad I went through the whole process: As soon as you dig deeper you face pretty complicated theory - this discussion about the influence of different kinds of screens for example - but you shouldn't get to hung up on the details. As always there is a big amount of intuition your heading for and you don't get that by theory alone.

But Trial and error alone doesn't get you far either. I had some support from re-recording mixer Matthias Lempert who showed me some vital tipps of workflow and answered questions...
 

---

Posted by Carsten Kurz (Member # 5396) on 03-04-2019, 04:46 PM:
 
Steve - why does Harkness publish different attentuation curves for their different perforation patterns? These curves show a significant high frequency rolloff, and quite some difference between standard and micro perf.

- Carsten
 

---

Posted by Marcel Birgelen (Member # 6801) on 03-04-2019, 05:59 PM:
 
The whole idea of the X-Curve, is that it's the response of the room versus a totally flat speaker fed with pink noise. The X-Curve, in the original idea, doesn't include the screen attenuation.

Now, it's indeed striking that the screen attenuation itself shows a similar plot to the X-Curve in many cases. It probably only adds to the confusion and the black art of movie theater tuning.

I know that for many in the industry, even questioning the X-Curve has become somewhat of a sacrilege and in general, it has done more good than bad.

But since we're living in 2019, there has been quite some change in the industry since the 1970's, when the idea behind the X-Curve was created.

Since then, the average auditorium has seen quite some changes:
- Rooms have generally become smaller.
- Where mono was the norm, we now have many speakers to consider, in the same room, some behind the screen and some not.
- Speakers and amplifiers have become more powerful.
- Movies have become louder and so has the need for isolation. All in all, rooms have become more "dead" over the years as a result of increased dampening.
- The advent of stadium seating has not only decreased the volume of the average room even more, it also drastically changed the geometry of the average room.

Measurements across the years have shown that the X-Curve seemed to fit the traditional auditorium quite nicely, but different kind of rooms have vastly different characteristics. The ways we build auditoriums have changed over the years, that might have impacted the validity of the X-Curve.

Personally, I think that every room should be considered on its own. Nowadays, we've got the equipment to measure and adjust accordingly. Also, modern auto-calibration and parametric EQ allow us to do stuff we never could about 40 years ago.

In the end though, you can calibrate all you want, the proof is in the listening. That involves humans and is still a very subjective thing.
 

---

Posted by Chris Erkal (Member # 10646) on 03-20-2019, 08:21 AM:
 
Just a question: Whats your opinion on making little covers for the speaker out of screen-fabric? Wouldn't that improve the quality of the dub stage as it adds a little diffusion and high frequency damping?
 

---

Posted by Carsten Kurz (Member # 5396) on 03-21-2019, 06:50 PM:
 
Well, you could do that, but, how do you make sure you get it right?

If I'd be concerned, I would probably try to find a cinema that allows you to set up your personal monitors in the auditorium and do some nearfield/farfield and response comparisons.

- Carsten
 

---