Film-Tech Forum ARCHIVE: Sound Question what is X curve?

This is topic Sound Question what is X curve? in forum Film Handlers' Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=1;t=008286

Posted by Cameron Glendinning (Member # 3516) on 10-03-2006, 11:31 PM:

I am currently being told that this X curve is a bad thing, on one side a speaker manufacturer of 4 way active systems and also by a person who designs Cinema Sound Mixing suites. I believe its a 3db roll off from 2 - 3 K, anyone with more details, are there other curves, has anyone heard a system flat? Any comment?

Posted by Gordon McLeod (Member # 33) on 10-04-2006, 08:31 AM:

The x curve is a 3 db rolloff in the room response stardig at 2K
This was determined as to what a test group felt was a natural playback of a live orchestra that was then played back
It was origin was with Bell Labs and Dolby defined it in the 70's
The smaller the room the less of a roll off the larger the room the greater the rolloff
If the room isn't designed to that standard then playback will not be what is intended just like anyone who records in a non standard suite would have a problem
The ourepose of standards is an attempt to get a uniform yardstick

Posted by Robert Minichino (Member # 3495) on 10-04-2006, 10:14 AM:

The X curve (ISO 2969) comes from the way the ear hears the reverberant sound field versus the way something like a time-averaging real-time analyzer hears it. If you set up a large room to have a flat frequency response on an RTA, it will sound too bright as the ears add in the direct sound from the speakers to the reflected sound in the high frequencies, while at lower frequencies they average more like an RTA. The X curve is applied to the playback system, and not the recording, in an attempt to correct for the apparent response of the space.

Posted by Steve Guttag (Member # 268) on 10-04-2006, 04:47 PM:

A key to the discussion on the X-Curve is being left out...that is the X-curve relates to how continious sounds like pink noise will be heard in a reverberant room versus impulse type sounds that is more representative of actual program material.

Since, as others have touched upon, continous sounds will allow the room to build up and have direct and reflected audio reaching the microphone (or ears for that matter), the X-Curve allows one to compensate for the fact that the direct sound will have subtractive audio from the delayed reflections added to it. Thus, the actual level will indeed be less. In theory, if one were to set up an average RT60 room with pink noise and follow the X-Curve and then remeasure the room using short tone bursts so that the direct sound is measured before the reflection...the response should take on that of FLAT. Thus, the X-Curve does not, in fact, roll off the response but actually allow one to have a flat response using conventional test material (pink noise) and test equipment (1/3 Octave RTA).

Note, that with more reverberant rooms one is supposed to roll off more and with less reverberant rooms one is to roll off less...which is all part of ISO-2969 or ANSI/SMPTE 202M (Appendix A6 table A.1). Dolby's Technical Guidelines (available for download here and contains ANSI/SMPTE 202M of its year), Figure 2.3 expands upon that and allows for one to lessen the "roll off" for smaller rooms.

Posted by Christopher Seo (Member # 7) on 10-06-2006, 05:43 PM:

To put it another way: If you place the speakers in a free field (i.e., open space) or an anechoic chamber (i.e., 100% acoustic absorption all around), and EQ them for flat response on an RTA, then take the same sound system and install it in an "average" theatre, or at least the average theatre of the 1970s, then the RTA would show something approximating the X-Curve, but of course most theatres would show some deviation from the mean.

So the X-Curve should be taken as a rough guide rather than an exact specification. It actually seems obsolete at this point. With the theatre acoustics, sound systems, and test equipment available in the 1970s, the improvement offered by the X-Curve standard outweighed its margin of error. But today, test equipment to measure direct sound is readily available, bypassing the guesswork of correcting for steady-state RTA measurements.

Posted by Steve Guttag (Member # 268) on 10-06-2006, 07:26 PM:

There are certainly alternative measurment techniques that have certain forms of validity and superstition assoicated with them...however when it comes to movie theatre room measurement, the current pink noise/RTA measurment has continued to prove to be the most reliable and consistant.

Like all techniques that involve human reading, the person doing the interpretation has a great influence over the results.

However, the statement about the X-Curve applying to theatres from the 1970s is flat out wrong. Furthermore, theatres of today are hardly superior...there are good ones and bad ones...in any era. The crux of them in the middle seem to be little better or worse.

Note, despite much of the critism of THX...they (Tom Holman and others) did extensive research in acoustics in movie theatres. Furthermore, THX rooms are typically built better than the typical movie theatre in terms of RT-60 times and other anomolies. Even with such criteria, guess what...the X-curve applies due to the measurement techniques used.

A movie theatre is not to be completely dead. Nor is it to be an echo chamber...there is a tolerable range given the volume of the room that will have the sound reproduced in a manner that most of the people can hear/understand it properly.

Lastly, the X-Curve is reviewed every 5-years by SMPTE (people like Tom Holman are on the committee, A12) to ensure that a SMPTE standard reflects accurately what is proper to the best of everyone's ability to verify or submit changes.

Posted by Christopher Seo (Member # 7) on 10-08-2006, 03:49 AM:

Isn't it contradictory to say that pink noise/RTA is "reliable and consistent" and then that "the person doing the interpretation has a great influence over the results"? There's a lot of truth to the old joke that no one ever EQs the same room the same way, including studio and Dolby technicians. Part of the problem is the limitations of the 1/3-octave equalizers, but the other part is the ambiguity of the test data.

I'm not arguing that the X-Curve is invalid for today's theatres, when using conventional measurement technology, but rather that the methodology itself is decades old and relatively primitive. There's no point in having a standard that no two people implement in the same way. If more sophisticated test equipment is available to improve precision and reduce ambiguity, why not take advantage of it?

Posted by Steve Guttag (Member # 268) on 10-08-2006, 04:49 PM:

Nope, nothing contradictory there... it is indeed both reliable and consistant and the person doing the job has a great influence over the result....The facts are that there are no "modern" tuning methods that will result in as consistant or reliable results

If one uses multiplexed calibrated mic system...odds are that if one were to tune the room...then another person were to come in and remeasure the room...also with a muliplexed calibrated mic system...the room will still read the same or so similar as to not be out of tolerence.

The facts are...I've done just that.

An arguement can be made about using 1/3-octave EQ versus parametric but...the bulk of the cinema techs can't handle parametric EQ...with 1/3 octave, there is a 1:1 mapping on the analyzer versus EQ. That has the best chance of working with most techs.

When I end up going to DSP based crossovers (I don't necessarily take an analog signal, go through an A/D D/A conversion process unless the benefits outweigh the problems)...I will almost always use parametric EQ in the DSP that negates any need for 1/3 Octave EQ.

However...the mere need of EQ, in my opinion admits failure to some extent. A speaker that is acoustically flat to begin with, in a room that is made properly...will always sound better than the system that require vast amounts of EQ to bring it into spec.

Posted by Richard Hamilton (Member # 321) on 10-08-2006, 05:21 PM:

quote: Steve Guttag
However...the mere need of EQ, in my opinion admits failure to some extent. A speaker that is acoustically flat to begin with, in a room that is made properly...will always sound better than the system that require vast amounts of EQ to bring it into spec.

Exactly. Otherwise you over EQ and try to make the speaker do something it isn't designed to do.

Posted by Gordon McLeod (Member # 33) on 10-08-2006, 05:39 PM:

Christopher said"So the X-Curve should be taken as a rough guide rather than an exact specification. It actually seems obsolete at this point. With the theatre acoustics, sound systems, and test equipment available in the 1970s, the improvement offered by the X-Curve standard outweighed its margin of error. But today, test equipment to measure direct sound is readily available, bypassing the guesswork of correcting for steady-state RTA measurements."

1st it is a standard not a guide and it has a tolerence factor in it
It maybe an old standard but it is still very valid
As to measureing direct sound it is irrelevent as in a large room one is in the reverberent field not the direct field

Also Parametric EQ is vastly superior to 1/3 octave EQ but even 1 octave eq is better than badly done parametric EQ
Many a Starscope has sounded better when someone who understands a parametric eq aligns it

Posted by Jack Ondracek (Member # 1466) on 10-08-2006, 06:50 PM:

So... if a processor was intended only for personal use through a set of high quality headphones, one would use no equalization at all?

Posted by Mark Gulbrandsen (Member # 72) on 10-08-2006, 11:16 PM:

You'd want it set flat if you're going to do critical listening with phones. The response of most headphones already take into consideration most of what was done at Bell Labs regarind the ear's sensitivity at different frequencies. I have EQ'd ramp systems to somewhat of an X curve though and they did sound alot better than they did flat.

Posted by Richard Fowler (Member # 893) on 10-09-2006, 10:59 AM:

Parametrics are especially great some situations such as live sound, where notching with a 1/3 EQ doesn't do the job. X curve is a good standard; most of the problems I see are in very small rooms or dailies sound playback situations where the roll off is not adjusted for the space....or the sound mixer says "run it flat" for playback, the end result ear splitting high frequency [Frown]

Posted by Christopher Seo (Member # 7) on 10-09-2006, 03:24 PM:

Sure, it's a standard, but it's also a rough guide. It's a rough guide because it's not a direct specification of what one wants to achieve, which is flat loudspeaker response. What is the point of having a standard that Dolby technicians themselves tell you cannot be followed to the letter? What is the point of having a standard which allows a tolerance range of 6 - 10 dB?

SMPTE 202M-1998, Section A.11 ("Future work"), lists the following as useful future inclusions to the standard: "a) reverberation time vs. frequency; b) loudspeaker radiation pattern vs. frequency; c) the consequent direct-to-reverberant ratio at listening positions within the space due to a) and b)". Can these factors be accounted for by traditional measurement methods?

Posted by Gordon McLeod (Member # 33) on 10-09-2006, 03:40 PM:

A standard is a standard and the tolerence range is less the 10 db
it is +-3 db which is the norm for rateing most audio devices
and THX is +2 -3 as a norm

Posted by Christopher Seo (Member # 7) on 10-09-2006, 04:00 PM:

SMPTE 202M-1998, Table 1, shows greater than -3 dB tolerance outside the 100 Hz - 8 kHz range. It does not state whether this is simply for the benefit of older loudspeakers.

Posted by Steve Guttag (Member # 268) on 10-09-2006, 05:04 PM:

Christopher, it is not a rough guide but a well defined standard. I hope you don't tune rooms because your responses show that you clearly don't know what you are doing.

For your reference...most speakers have graphs that show closer to +/- 5 dB of range and most manufacturers list their range at the -10dB down points.

In reality...the reponse deliberately opens up outside of the vocal range. How close the HF horn comes to the screen will have a dramatic effect on its HF response and the curve allows for that. In fact, I want to say that it was a more recent inclusion to alter the curve for above 10KHz. The +3 upper limit is uniform from 40Hz to 16KHz. The lower limit does vary based on actual speaker performance. Furthermore, with the optical subwoofer rolling in at about 50Hz - 100Hz (installer settable), the response of the stage channel need not be ruler flat down to 20Hz.

I have always questioned that cinema stage speakers should have their 3dB down point at 40Hz but optical subwoofers should come in at 50Hz.

In any event, if you know how to interpret the X curve and how speakers actually perform...it is pretty straight forward.

You seem to be looking for some new magical sound physics or futuristic sound measurement method that will make it all cut and dried....sorry to say...that day is not here yet nor does it look like it is around the corner.

Posted by Tim Reed (Member # 137) on 10-09-2006, 07:16 PM:

I'm of the less-is-more philosophy. Pull down the peaks, little or no boosts.

Too many people try to use EQ to make the speaker compensate for deficiencies in the room, instead of insuring the room matches acoustically to begin with (be it in design or remodelling/construction). There is only so much you can do before you start getting noticeable phase distortion and coloration.

Btw, when did it stop being called the Academy curve... or are we talking about something that I've completely missed during my time out?

Posted by Robert Minichino (Member # 3495) on 10-09-2006, 08:14 PM:

From my understanding, the X curve is the one with the gentle 3dB/oct roll-off starting at 2k, the Academy curve is the pre-NR one that cuts the highs and lows rather sharply.

Posted by Jarrad Salmon (Member # 3818) on 10-09-2006, 11:49 PM:

quote:
+-3 db which is the norm for rateing most audio devices

Quite right, +-3dB is the standard. Still a 6dB range is pretty big, especially at frequencies in the presence range. The amount of speakers that I have compared side by side in this scale blows me away. You can always hear the difference even if they are showing the same response.

Please try not to focus on minor details like above

I agree that the idea of an x curve is purely a starting point in a rooms EQ. Keep in mind this test is completed with a RTA mic in a specific position in the empty cinema. If you move the mic the response will change, change the temperature, humidity, the amount of people the response will change.

There will never be a perfect tuning method. Just methods to use as a reference point.

Posted by Cameron Glendinning (Member # 3516) on 10-10-2006, 04:07 AM:

Ok, Over the last 2 weeks I have asked a few people with life time worth of experience in the professional audio their opionon on the x curve.

So far on the phone I have talked to

1. John Bernet founder of Lenard Audio. Who were the world pioneers of the high power amp. When the rest of the world was making amps in the 10's of watts he was producing and selling amps in the hundreds of watts.

2.Fred Pickering, Who for decades worked in TV sound at Australias ABC Television. Fred is the sort of person who use to send the sound recordists to the school of the blind as work experience, he was the head of sound there when the now famous Neville Theele was the head of video tape there ect.

3.Richard Prittle has been contacted, designer of the Aussie Monitor Cinema Speaker range and even designed Dolby mixing suites ect

4. I am also keen to talk to a person who mixes movies for a living to see if they actually try to compensate for them ect.

One of the above is a friend of Tom Holmen and ultimatly his opinion will be asked as well.

First to respond is John Bernett who has given permission for this email to be reproduced here on Film-Tech. He has been researching the history of it.

The mysterious x curve

A variation of the x curve was originally described as the Academy curve, which came about as a convoluted entanglement from many different directions.

First we must understand that cinema evolved before radio and Television. Cinema was globally successful as the number one entertainment for 20 years before electronics evolved to enable sound systems to exist. This is the primary reason sound has always taken second place. We must also remember that the cinema industry has had an obsession that any new technology had to be backwards compatible, meaning that it was and still is contained by its past. This not right or wrong but simply evolution.

Sound appeared as a miraculous achievement and was applied when electronic technology was unfolding so rapidly that there wasn't time to for cinema to grow with it. Many of the decisions at that time were done with the best intentions to achieve consistency.

Before sound systems existed the largest auditoriums were cathedrals and the only way to increase the sound level was by controlled reflection of reverberation. Reverberation was therefore essential to maintain consistent sound level but it came at the cost of decreasing articulation and intelligibility. Early acousticians began with an imaginary ideal as a 100% reverberant environment enabling sound to be heard at the same level throughout the auditorium or cathedral.

Any sound absorbed was regarded as a loss described as 'room loss'. The aim was to provide minimum 'room loss' with zero loss of intelligibility which of course is a contradiction. Any reverberation reduces intelligibility. Therefore what evolved was the imagined ideal of correct reverberation. That is, reflective area around the stage of short distances opening up to larger area of longer distances. However this is not the case with cinemas because of the large screen.

It was noticed that the worst effect on destroying intelligibility was hearing distinct echoes reflected from walls and ceiling at large distances apart, compared to reverberation reflected from walls and ceiling at short distances (e.g. bathroom).

Lets compare a small and large reverberant auditorium made from the exactly the same building materials, and we fire a cap gun. Assume a period of 1 second in which sound will travel 1125ft (344m). In the small auditorium average distance between the walls and ceiling is closer, compared to a large auditorium. Therefore in one second the sound from the cap gun will strike the walls and ceiling more often in a small room compared to a larger room. This is described as 'Mean free path' or 'Mean free time'.

Each time sound strikes a wall or ceiling a % will be absorbed and a % will be reflected and so on and so on. Therefore in a small auditorium -more- sound will be absorbed by the walls and ceiling simply because sound strikes the walls and ceiling more often compared to a large auditorium, for the same period of time. However in most auditoriums the high frequencies are more readily absorbed that the low frequencies. Which means that in larger auditorium the echo reverberant sound will appear to be brighter.

The second explanation for why the Academy curve exists is to do with noise masking.

The first sound tracks on film were basic optical and the technology that existed at that time could only enable 2K - 6K max to be obtained. The first problem was the very noisy hiss from the optical track and the crude valve pre-amplifiers. Putting a filter at the output of the pre-amplifier reduced the hiss but also reduced the small amount of high frequencies information available from the film stock. The simplest solution was to boost the high frequencies when recording and allow the noise reduction filter to give a close to flat response again when played back. This became called the 'Academy Characteristic'.

The 3rd explanation for why the Academy curve exists is to do with speaker horns.

The early speaker systems consisted of a 12in or 15in speaker for the low frequencies and a compression driver and horn for the higher frequencies. Crossover at approx 700Hz which mostly consisted of a simple capacitor only. The compression driver and horn were approx 20dB (50 - 100 times) more efficient that the 12in or 15in speaker. The understanding how to correctly equalize both components so as to sound flat was very difficult as only a very few laboratory had the basic and crude test equipment to calibrate and acoustically align the drivers. The majority were aligned by ear and this was always in favor of the upper end of the voice. Also valve amps at that time were mostly 30 Watts or less. Often the horn was left almost at full efficiency so it could be loud enough for lost of people to hear with only a small amplifier power. The Academy curve also helped equalize the extreme level of the horn.

Conclusion.

We have to understand we are now living in a new century. Sound system power and fidelity is now unlimited, but unfortunately the will to act decisively is limited. Digital formats allow for a flat response 20Hz - 20KHz with almost zero noise. Electronic amplifying circuitry is now virtually silent. Auditoriums can now be made close to anechoic with almost zero reverberation and echoes. There is no longer any need for the Academy x curve or any external equalizing curves in the new millennia. We now have the technology and engineering skill and un-limited building resources to make every commercial cinema acoustically free of echoes and reverberation at all frequencies. Traditional cinema now belongs at home as 'Home cinema' where it can be trivialized and de-valued to any level we wish. But for the commercial cinemas survive in the new century they will have to be fully digital and provide the highest level of projected entertainment for mass audience. But to do so the industry has to let go of the past and invest in the technology that will enable it to maintain its position as the leader of mass entertainment

Posted by Gordon McLeod (Member # 33) on 10-10-2006, 02:59 PM:

With all due respect to Mr Bennet he has done some poor research
The X curve and the Academy Curve are two very different items and very different in there function and amounts
The academy curve was a pre emphisis (much like the RIAA premphasis) and had nothing to do with room accoustics
The X curve is the response of the B chain portion of the system

Posted by Steve Guttag (Member # 268) on 10-10-2006, 06:09 PM:

To back up Gordon on this one...yes...the description just given by Mr. Bennet seems to talk about the evolution of the Academy characteristic curve as measured in several large cinemas years ago which lead to the pre-emphasis mentioned by Gordon (some Dolby papers actually show how the sound moves from flat to the Academy curve)...

The X-curve refers to an SMPTE standard on how to tune a room to ACHIEVE A FLAT response. To think the sound is actually being rolled off is a misinterpretation of what is going on. The reason for the range also has to do with not over taxing a driver or the electronics.

The person that brought the tid bit about the X-Curve being for a tuning of the room at a particular mic location just showed everyone that he did not read the document he was providing comment on...actually the document decribes a multiplexed microphone system with spacial temporal averaging so that the bulk of the listening area is properly measured (see A4 through A8). Now the standard does advise on single mic placement but doesn't recommend it.

The goal of the standard is to standardize the measurement method for uniform and predictable results and also provide the best results.

Posted by Cameron Glendinning (Member # 3516) on 10-10-2006, 07:13 PM:

Gordon, Johns research is not finished and he is very interested in learning more, your and other Film Tech members input is very useful to help demystify what appears to others, who do not work in cinema industry, an unusual practice.

Personally in the 90's when I was building speakers with wings/ baffles / modifying cinema speaker systems, I had the opportunity to get our 3 different service providers to each do an eq of the same auditorium, with very different results. So I have a good idea of the acceptable variation within the curves. It is not subtle. The way I deal with it was to only use the company that had provided the best and most natural sounding results.

Having worked in over 50 cinemas so far, I can also say that most auditoriums do not have good acoustics. With Digital Cinema promising uncompressed or IMAX quality audio, should cinemas start copying their dampened / dead auditoriums, Do IMAX use the X curve? What could be done to make it more consistent? if anything?

Posted by Tim Reed (Member # 137) on 10-10-2006, 08:00 PM:

quote: Gordon McLeod
The academy curve was a pre emphisis (much like the RIAA premphasis) and had nothing to do with room accoustics
The X curve is the response of the B chain portion of the system

Forgive my cobwebs, guys. I've spent the last 5 years putting much of this out of my mind - thinking I would never return to service. It is critical that I get this right. Seems I remember the slit-loss correction being called Academy correction (?), but I could've sworn the old CP-50 manuals also referred to the B-chain rolloff at 3dB per octave above 2kHz as the Academy curve. [Confused]

Gordon and Steve... is there anything in recent Dolby manuals that outlines this, perhaps? Thanks! I've got some books to crack.

Posted by Steve Guttag (Member # 268) on 10-10-2006, 08:54 PM:

Tim,

Download the CP50 manual from this site...read chapter 3 for the Acadmey cuve stuff.

Then download the Technical Guidelines manual from this site and you can see ANSI/SMPTE 202M...it isn't the most current but it is going to let you see what the X-Curve is all about. Also the N-Curve which relates back to the Academy curve.

Posted by Tim Reed (Member # 137) on 10-10-2006, 09:19 PM:

Roger that, Steve. Thanks!

Posted by Gordon McLeod (Member # 33) on 10-11-2006, 06:19 AM:

Also there is some matereial in the paper on Quintaphonic on room equalization in a old SMPTE jurnal

Posted by Edward Jurich (Member # 1787) on 10-11-2006, 11:01 AM:

For two or three years I ran summer classic films in an old 2000 seat movie palace. The films ranged from the 40's to the 60's. We had an EQ and when a new feature came in I would pick a middle reel and run sound tests. In order to be able to understand speech and have a good sound, the EQ had different settings for each feature.
The one thing that stayed pretty constant was a big dip between about 200HZ and 2KHZ. If you didn't put that dip in the sound was muffled.
The first couple years we used an old original mono photocell pickup requiring voltage. The third year we put in a modern stereo cell so we could run stereo on West Side Story. Even with the newer pickup, the dip still had to be put in.

Posted by Tim Reed (Member # 137) on 10-11-2006, 07:50 PM:

Gordon, is that online anywhere?

Posted by Gordon McLeod (Member # 33) on 10-11-2006, 08:23 PM:

I don't think so