|
|
 
|
|
Author
|
Topic: SMPTE B-Chain Report Shows Standards Not Working
|
Brian McCarty
Film Handler
Posts: 1
From: Clifton Beach, Qld 4879 Australia
Registered: May 2014
|
posted 10-29-2014 09:45 AM
I note that a link has previously been posted to the October 1, 2014 SMPTE B-Chain Frequency and Temporal Response Analysis of Theaters and Dubbing Stages.
I can't stress enough to those running movie theaters how important this report is for the industry. The report has taken over four years since the original work statement was submitted, and the final drafting committee included Ioan Allen, Glenn Leembruggen, Floyd Toole, Philip Newell, and many other leaders in the audio industry.
The reality has been scientifically documented - the Standards called out in S202, RP200, and others are broken. This report, along with other work done in Europe (Philip Newell, et. al.) prove that the so-called calibrations to these Standards are non-existent. Every dubbing room is different - in many cases wildly different. This explains why the tonal balance of every movie is different. And the movie theaters are in similar shape.
The challenge now is to fix the problems that have been created. In my view this overhaul will result in performance Standards which is how the picture Standards work. It should also remove the X-curve and bring cinema audio into line with the rest of the audio community.
I urge the experienced techs here to read and comment on this report. The AES has organized a Hollywood Conference March 6-8, 2014 to spend three days discussing audio issues across the whole entertainment chain. I'm sure many of you will be interested in attending.
http://tinyurl.com/olf7cya
| IP: Logged
|
|
|
|
|
|
|
|
|
|
|
|
David Buckley
Jedi Master Film Handler
Posts: 525
From: Oxford, N. Canterbury, New Zealand
Registered: Aug 2004
|
posted 10-29-2014 09:36 PM
quote: Louis Bornwasser
We are NOT the only people using the X curve. It is common in arenas and worship areas.
I've never heard the term used outside of cinema usage, I've never heard it used in connection with sound reinforcement, and I've certainly never heard a system tech who works arenas use the term.
quote: Marco Giustini
well the X-Curve is an approximation of what a flat direct sound looks like in a reverberant field.
A paper by Michael Karagosian published last year reminds us that "X-Curve Is Not An EQ Curve" (google for it, the link is horrible).
quote: Marco Giustini
The main question has always been: how do I account for different reverberation times?
That paper states:
quote:
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.
In pro audio, RTAs were the tool of choice in the last century, but I doubt anyone in pro audio has used an RTA to equalise a system to a venue this century. RTAs are time and phase blind, and thus give misleading results, especially in a reverberant field.
The X-Curve was a bodge to enable cheap, easy setup using 80s technology. Sound reinforcement technology and measurement technology has moved on significantly in the intervening decades. It's time for the X-Curve to go.
| IP: Logged
|
|
Harold Hallikainen
Jedi Master Film Handler
Posts: 906
From: Denver, CO, USA
Registered: Aug 2009
|
posted 10-30-2014 08:06 AM
I believe SMPTE will need to do more study on this (if AES does not do it first). A few points of interest from the report:
1. The high frequency rolloff was present in "close field" measurements as well as measurements deep in the auditorium. Since reverberation is a much smaller component of the total field close to the speaker, if the rolloff were due to reverberation, I'd expect the rolloff to be less. The report pretty much says the it could not demonstrate the the "reverberant build up effect" used to justify the X-curve.
2. The report is based on impulse response measurements. The IR was generated by convolving the pink noise generator output with the microphone output. The IR was then run through an FFT with various window sizes to generate the frequency response graphs. The main advantage to this (as far as I can determine) is that you can use a short time window on higher frequencies to get the frequency response of the direct sound. Longer time windows or RTA include reverberation in the frequency response measurement. However, the measurements with a 2 second window (about the same as RTA) are nearly identical to those with a shorter window. I think that in a cinema environment (where low reverberation is the norm), RTA and IR give very similar results.
3. Not mentioned in the report, but a small group of participants repeated the original X-curve experiment. A flat near field speaker was placed close to listeners and an RTA (maybe it was IR measurement - I was not there). People were asked to adjust the tone of a distant speaker behind a perforated screen so it sounded like (same timber) as the close speaker. The resulting response showed the HF rolloff. If indeed I understand what was done, it indicates to me that the instruments do not measure what we hear. I'd like to see more work done on this. Did the perforated screen create this effect? Or was it just the distant speaker?
4. My current thinking (and this is subject to change) is that the HF rolloff is due to screen loss. If the screen loss on the dub stage is the same as the screen loss in the theater, and the mixer adjusts the sound to sound good on the dub stage, the mixer is applying a preemphasis, and deemphasis is applied by the screen. But, I think more research in this is needed (at least more is needed so I can understand it).
5. One dub stage of interest in the report is Venue F. I was there for those measurements. It uses a woven screen, which is pretty much acoustically transparent. HF rolloff has to be applied in the electronics since it is not present in the screen.
Harold
| IP: Logged
|
|
Marco Giustini
Film God
Posts: 2713
From: Reading, UK
Registered: Nov 2007
|
posted 10-30-2014 10:46 AM
quote: David Buckley
The X-Curve was a bodge to enable cheap, easy setup using 80s technology. Sound reinforcement technology and measurement technology has moved on significantly in the intervening decades. It's time for the X-Curve to go.
I agree. One of the issue here could be time. I am more than happy to spend as long as it is needed to EQ a sound system, but will the customer be happy with that? We all know way too many cinemas are EQ'd in minutes or not EQ'd at all. I am not saying we don't need a better system, I am just saying that when designing a new system we also need to take into account the real world.
quote: Harold Hallikainen
. My current thinking (and this is subject to change) is that the HF rolloff is due to screen loss. If the screen loss on the dub stage is the same as the screen loss in the theater, and the mixer adjusts the sound to sound good on the dub stage, the mixer is applying a preemphasis, and deemphasis is applied by the screen. But, I think more research in this is needed (at least more is needed so I can understand it).
I don't think the screen has anything to do with that but it can be easily verified: just EQ the system before the screen is put in place and listen. I believe the subjective result will be the same -even though a speaker will sound better with no screen, just because you are not pushing the HF, no comb effect caused by the perforations and no reflections back to the back wall.
| IP: Logged
|
|
Steve Guttag
We forgot the crackers Gromit!!!
Posts: 12814
From: Annapolis, MD
Registered: Dec 1999
|
posted 11-02-2014 04:38 PM
Being a member of the SMPTE standard's community, I'm a bit reserved with discussing an SMPTE document outside of committee. This is not a document I personally had a hand in, I'm sad to say...didn't have the time (but I did contribute to the one on measuring light off of movie screens). I have had my hand in 202 over the years though.
Having gone through it once, I'm a bit troubled by several aspects of it. Not the least of which was that the equipment under test (speakers, crossovers, equalizers...etc. were not identified. While I can understand the desire to not endorse or condemn any brand/system, it is impossible to draw conclusions without completely isolating the thing you are trying to measure or understand. Too many variables are changing from room-to-room to make a valid claim as to what is going on.
In fact, one of the things that they "discovered" and that Harold has touched up is how the screen loss affects things. However, even then, incomplete information is provided to know things like the distance behind the screen each speaker was placed and how that relationship affects the final outcome.
However, the impetuous for the study is to see if there isn't a better way to measure/tune a cinema to have a more uniform outcome in terms of tonality from room to room, particularly from dub-stage to cinema. It is evident that not all cinemas sound the same...not all dub stages sound the same and is that a result of the measurement and tuning techniques used? This is all very valid areas of study to improve our industry.
I do find it interesting that when RTA with a steady-state PN source is used compared to an impulse response, the results are VERY similar...to the point, they are closer than mic-to-mic variations for most systems.
But lets look at the whole "X-curve" thing a bit and how it all came to be in ST202. Just like there is more to measuring light to be 14fL in the center is but a portion of that standard, the X-curve is but a portion of 202. You can't just take that portion out of 202 and ignore the rest.
The X-curve is part of the whole package the presumes that one is in a cinema, with a standard movie screen, with proper RT60 times. That the primary microphone is placed 2/3rds back (and slightly off-center to avoid being in a mode). I don't agree nor have I found that having the primary microphone 2/3 off-center of the seating area to be best but that was chosen to ensure one is not in a mode. It is simple enough to play pink noise and listen to where you are going to place your mic(s) to ensure they are not sitting in a room mode that would drastically alter response that is representative of the room/system. It is also identified in 202 that a minimum of 4 microphone positions are needed to get a representative response of the room with CALIBRATED microphones (and I can't emphasize the calibrated part enough).
So, in essence, what 202 currently states is if you follow its instructions in a typical cinema that is properly made, it will provide a valid means of evaluating and tuning the room using conventional equipment. That remains valid and nothing in the report presented really shows that it isn't the case. However, as the room deviates from being properly made, the valid points that an RTA system is "blind" to time/phase anomalies are true and that one adhering to the X-curve at all costs in all environments will likely tune the system to their own detriment. 202 does provide some very rudimentary corrections to the curve (see table A1) based on approximate seat counts based on the presumption that with larger rooms the RT60 times will grow and with smaller rooms they will shrink. Furthermore, the microphone positions will move further away or closer to the noise source (speakers).
But if you have an atypical room with a very odd shape, RT60 times, speaker locations...202 will apply less and less and you will need other techniques in order to achieve the best possible results. Mind you, you will never achieve ideal results...you can't fix a bad room (or equipment) with mere tuning and measurement...you can only make the most of what you have. 202 was never meant to deal with that sort of room.
Another key goal of a standard like 202 is to provide a uniform method/procedure to achieve a common result. If a dub-stage and a cinema use the same technique, the odds of similar results go way up. Tom Holman and the whole THX process really took this idea to heart. Since THX could control what equipment went into both the dub stage as well as the cinema as well as have a hand in how these rooms were constructed, the ability to achieve a much greater room-to-room uniformity was achievable. And, early on, it worked. The problem, as most of us discovered, they succumbed to dollar and started approving sub-standard rooms (and equipment), which made them, potentially, no better than rolling one's own system.
However, if the goal is to achieve the same tonality between dub-stage and cinema, then THX was the right idea (and reality for a period). They both used identical procedures/equipment/effective responses on many parameters. These are all key. Furthermore, even the equipment used to measure the rooms was also unified with the R2 analyzer (and its predecessor).
This is the type of situation where 202 actually really applied the best. It is irrelevant if the loss due to the screen affected the response because both the dub-stage and Cinema had the same attenuation...the person mixing the film would tuen their effects/dialog...etc so it sounded best to them...since the cinema was tuned identical, that tonality would translate into the cinema perfectly. Think of it like analog noise reduction...so long as the systems are complimentary, it all works.
Let me touch on more more thing I think is key to a good result with 202 as it currently stands...the analyzer. Thus far, I've only personally seen or used two models that really got it right. They are the R2 and D2. Why? Because those analyzers calibrate their microphones and normalize them to mic-1. What does this mean? You've taken the variation between manufacturing tolerances (and aging since you should get them calibrated annually) and zeroed them out...you've taken your measuring device out of the equation of influencing the results. You can take an R2 or D2 system...tune the room...then move the mics all around and it will read essentially the same...if it was within tolerance before, it will be still. It gets a true average of the room by not impressing the inaccuracy of four different mics onto the system. I can always tell when a room was tuned up with an R2 or D2 before me...eventhough I don't know where they placed their mics...when I put my mics down and look at the response...it is going to look like it was just tuned. You normally don't get that with other systems. This actually shows a validity to the whole 202 process...it can achieve consistency. A problem I have with other plexer systems, including the USL that "uses the same equipment" is that there really isn't that calibration factor in there. You can trim the levels to try and normalize them...but who really does that on a room-by-room basis? And the response...they are not acoustically flat microphones and there is no means to really have them read flat or identical from mic-to-mic. The R2 and D2 achieve that.
The test procedures in this article also didn't calibrate the mics...they went off of "spec sheets" for the mic responses. There is a definite flaw in the test procedure. Even though a D2 was present...it would seem like one of its key features was kept from being used. Again, I'm just going off of the test procedures as presented in the document. It seems like common microphones were used and split amongst several analyzers. I will ask Tim Holmes if he was able to "calibrate them for the D2 prior to testing since he was listed as part of the "ZDT team."
Another flaw I perceive in the document is that somehow all speakers seem to be considered equivalent! That is, when trying to achieve uniform tonality, the thing that is making the sound is somehow given a pass in the whole testing procedure. We don't know how much each speaker is contributing to the response we are seeing. Is the theatre with the woven screen achieving a higher HF response due to the better transmissivness of the screen? Does the speaker have a VHF piezo super tweeter (it was in a dub-stage so all bets are off as to what is behind that screen if it isn't identified...you can get some esoteric speakers or custom builds). Was the speaker right up on the screen, was it back a foot? All of these factor into the response of the system. We will never achieve identical tonality with disparate speaker systems used. And that should be a subject of discussion though. If the industry uses a particular level of speaker...should the dub-stages be required to use a like speaker? If they are going to use some super-duper speaker system that costs more than an entire typical cinema sound system, how will we ever achieve similar tonality? And if we can (through some magical tuning and measurement techniques), why would they waste the money on the super-duper speaker?
Oddly enough, before THX, though we had a bit of a varied quality of electronics and cinema construction (everything from old barns of a theatre to mini-multiplex ones)...we did have a high degree of commonality in cinema speakers with Altec Voice of the Theatre. Love em or hate em...they were universally used in cinemas and dub-stages alike. It is one variable removed from the equation. We will never achieve that again. It is unreasonable to expect a QSC to sound like a JBL, EV, Meyer, Klipsch, EAW...etc. When was the last time you heard a prominent speaker manufacturer use a sales pitch that "we sound just like the other prominent speaker manufacturer"?
So it is a bit of a fallacy to think one is going to match a dub-stage's sound if the cinema is not using the same equipment. Heck, even within a single complex I often see a range of speaker models based on if it is in a large room or a "move-over" house. I've noted most speaker companies have adopted a "good, better, best" approach to speakers. That is, they have tiers of speaker qualities. Is it reasonable to expect that the tonal quality is identical from "good" speaker to a "best" speaker even within the same brand?
I can't tell you how many times I've seen a speaker that was just too small for the space it is in. There is just no way you can measure and "tune" that sort of problem out. And that is not addressed by 202. While the work of the study group does not address it, one thing they DO touch on is improving the sound by keeping distortion to a minimum. Improved measuring techniques will allow one to better qualify that sort of problem. But there are all sorts of distortions out there. A distortion is anything that alters the signal from its original...but they are implying that they are interested in things like Harmonic Distortion.
However, if I change nothing more than the crossover point in the system...how does that affect the tonality system? At one point, the industry adopted a 500Hz crossover point for 2-way speakers while sound reinforcement often went for 800Hz and above. There was a reason for that...cinemas wanted dialog intelligibly above all else. 500Hz kept the crossover point out of the dialog range the best and our demands on the HF diaphragms were not going to be anywhere near what a rock concert would need (moving the crossover point up in frequency means that the LF driver bears more of the work than the HF). A problem I find with most 3-way speakers is their timbre and how they handle dialog. Invariably, their crossovers will have a point smack dab in the middle of the dialog range and it is VERY tricky to get two disparate devices to sound the same...particularly when they are separated at a distance. Furthermore, one is often going from a paper cone driver to a metallic compression driver...making common tonality nearly impossible. On the up side, you are not driving the HF driver nearly as hard and can likely go to a smaller system that handle the higher frequencies better. In theory, distortions should go down but it doesn't mean it sounds better and, more often than not, it doesn't!
Another thing I'm finding is that the pre-configured parameters or specified parameters for setting an active crossover, including the delay settings, in the real-world of cinema are not necessarily yield the best results. Better testing techniques can catch such stuff better than that the current 202 processes.
Many here have come across where a bad diaphragm was not discovered until AFTER tuning and a proper listening test was done. Steady state pink noise with an RTA is very poor at finding such problems.
But lets take a bit of a reality check here. A goal of ST202 is to provide a technique, if used, will result in a unified and repeatable method of optimizing a B-chain response. A fair number of people do follow it to some degree. The disparity of results can be traced to those that do not follow it or follow it fully and/or rooms/equipment that not suitable/optimized for cinema sound reproduction.
The level of effort to use a "modern" IR technique would increase the time required as well as the skill set of those that use the equipment in order to take advantage of it. This would tend to lessen the likelihood of it being followed. Furthermore, as it stands, the Study Group still has to actually provide a procedure to follow (so you are seeing something in its developing stages, not in its final form...which I find VERY odd for the SMPTE).
Likely, what will be needed is a multi-tier approach as we did with the light measurement document where it is acknowledged that the time available and accuracy needed in a review room are not the same as a typical cinema. That the tools of a typical technician taking care of a cinema are not going to be as plentiful as those taking care of a review room where color grading may be made.
I foresee that ST202 will NEED to keep the X-curve and the technique/procedures that go hand-in-hand with it as base line. That perhaps a more sophisticated set of measurements be adopted to deal with high-end rooms as well as a means to address problems that may be identified when a "properly" tuned room ends up not sounding good. But remember, in the end...there is NO substitution for getting the room made properly (good acoustics, reverb, noise levels) and using good equipment that actually reproduce the sound properly. And surprise, surprise, the cheapest equipment rarely does that well.
As always...sorry for being a bit long winded. But hey, in the US and other DST countries, you got an extra hour to read it!
| IP: Logged
|
|
|
|
|
|
All times are Central (GMT -6:00)
|
|
Powered by Infopop Corporation
UBB.classicTM
6.3.1.2
The Film-Tech Forums are designed for various members related to the cinema industry to express their opinions, viewpoints and testimonials on various products, services and events based upon speculation, personal knowledge and factual information through use, therefore all views represented here allow no liability upon the publishers of this web site and the owners of said views assume no liability for any ill will resulting from these postings. The posts made here are for educational as well as entertainment purposes and as such anyone viewing this portion of the website must accept these views as statements of the author of that opinion
and agrees to release the authors from any and all liability.
|
Home
Printer-friendly view of this topic