Monitor Calibration Quandry

[georgia]

if you are delivering music for audio editorial and mix for broadcast... just make sure the Monitors are matched and set to whatever level you want.... Music is different than broadcast / film post. Your music is going to be edited and mixed by a re-recording engineer or mixer in a calibrated room.... What you do is all about making good music.

I applaud you for delivering at -20.... I hate it when I get music slammed to 0 ...or delivered at -30 or less. The first is too hot and I just end up dropping it to around -20 anyway.... Delivering at -30 seems a bit low in my opinion and at that point if I have to raise the level, i'm dragging the noise floor up with it.

If you are mixing and listening to music at 79 db with monitors 2 feet from your face, it's gonna be subjectively loud, especially if you aren't used to it.

If you are mixing the entire project, then you need to assure it plays well and meets broadcast QC, as well as your client's approval. In this case I would calibrate with both broadcast EQ and level calibration. A lot depends on the room, the distance from the monitors, and what you do in the room.

cheers
geo

[jackmack]

Thanks so much for this Georgia. I come from a music background and while I've spent the past few years honing my "tech chops" I sometimes still get a little overawed on the GS forums. Everything you say a) makes sense to me and b) reassures my inner non-tech child. I've just seen your sticky on the forum and will be sure to read as much as I can (before I get frightened). Thanks again.

[MixMasterM]

I usually monitor at 74/76 dB in my small edit room with near field monitors. That seems to be a good level for me (personal preference and might not suit everyone). My edits and/or mixes translate pretty well onto a large 2000 sq foot mix stage for final mixing at unity. I think if you set your room to a comfortable level and deliver music in the -20 ballpark, you should be in pretty good shape for the re-recording mixers....and if you say they haven't complained, then your room calibration is probably pretty good as is.

[nathand]

Originally Posted by georgia

In this case I would calibrate with both broadcast EQ...

Care to elaborate. Obviously I'm aware of Processing in the Broadcast Chain (Compression, EQ, etc) but I was not aware of any sort of "standard practices" being used/required. Is this "broadcast EQ" published anywhere? It's really hard to imagine all networks using the same settings.

[georgia]

There is a Broadcast Curve, similar to the Film Curve.
Both have calibration setups. The broadcast curve is wider than the X-curve for film monitoring.

If you dig through the Room Calibration thread, Speaker calibration, and my thread, as well as the web, there is more on this.

cheers
geo

[nathand]

Originally Posted by georgia

There is a Broadcast Curve, similar to the Film Curve.
Both have calibration setups. The broadcast curve is wider than the X-curve for film monitoring.

If you dig through the Room Calibration thread, Speaker calibration, and my thread, as well as the web, there is more on this.

cheers
geo

Looked around a bit and all i could find were your own mentions of there being a "broadcast curve" (one from a post 10 years ago). Nothing published by a network. Nothing from the FCC (the only body I could imagine that would be able to institute a practice that all broadcasters would have to adhere to). None of my Network spec sheets define an EQ Curve used during broadcast.

Again I'm not questioning whether or not content is EQ'd, of course it is, but just if every broadcaster is doing the same thing thereby making it easy for a mixer to apply a single curve that would be applicable in all situations. The closest thing I could find is in section 10 of the ATSC Recommended Practices 2009 where it "recommends" mixing through an EQ that is totally flat all the way up to 4kHz where it rolls off to -6 at 20kHz. Would you call that the "Broadcast Curve"?

I'll look around some more...

[dr.sound]

That is called a "Modified X Curve" or even "Small Room X Curve".
Personally I wouldn't waste my time with it if your room is that small.

[tpad]

...I've recently tried calibrating my monitors to 79dB using the pink noise, -20dB, C, slow setting etc and I find it to be prohibitively loud....

Eric Benjamin at Dolby Labs did a study of preferred listening levels 8 years ago, and found that the mean preferred TV listening level is 58 dBSPL (65 dB for home theater buffs). This gives you an idea of what the population in general thinks are "normal" listening levels. So, if you think that 79 dBSPL is prohibitively loud, you're right... it is!

[dr.sound]

First, how far away are you from your speakers? How many speakers?

[jackmack]

I'm about four feet away from each speaker and they are again four feet from one another.

[dr.sound]

I don't sit 4 feet away from the TV when I watch. Nobody I know sits that close.
If you sit that close then sure, lower the monitors.
All my speaker calibration articles like this sticky on the DUC
tell you how far away suggest monitor speakers should be:
NEW UPDATED Room Calibration for Film and TV Post - Avid Audio Forums

[tpad]

If you're doing nearfield monitoring, then 4 feet is about the right distance. It doesn't matter whether you're 4 feet or 6 feet from the monitor if the playback level has been calibrated at the listening position. And the old .... it doesn't sound as loud at distance because of room diffusion is bogus! It's the same sound pressure level!

The basic problem is that a suggested 79 dBSPL calibration for smaller room monitoring is too high and probably needs to be changed. If you work the numbers out, you can see why.

The original spec of 85 dBC @ -20 dbFS was intended for MOTION PICTURE THEATER mixing and monitoring, where the sound was associated with a full size screen. Due to human perception quirks, if the visual image is larger than normal than the sound also needs to be larger than normal to maintain the perspective. Dolby intended dialog average level to be -31 dBFS, which would mean that dialog would be 11 dB below 85 dBC or 74 dBSPL. Apparently, everyone seems to think that this works (no argument).

Totally different ballgame when you get into a smaller venue, and based on Dolby's own study that I cited above, the preferred listening level in a home theater environment is 65 dBSPL, almost 10 dB below the cinema level discussed above (i.e. half as loud).

If you apply the same -31 dBFS dialog criteria to 79 dBSPL setup criteria, you get 68 dBSPL for dialog level, which is higher than what the Benjamin study claims is preferred. So 79 dB probably needs to be lowered 4 or 5 dB at least.

Further compounding the problem is the ATSC/CALM spec which specifies average speech levels to be -24 dBFS (not -31 dBFS ala Dolby), which would mean that average dialog levels in a 79 dB system would playout at 75 dBSPL, which is obviously way too loud.

[GregorioM]

Originally Posted by tpad

And the old .... it doesn't sound as loud at distance because of room diffusion is bogus! It's the same sound pressure level!

You really need to look at the field of psychoacoustics. Loudness is a human perception and not something which can be accurately measured (with SPL or any other scale). SPL is only one of the factors the brain uses to determine loudness and there are a number of additional factors which is why "the old distance" argument is true and in fact it's your statement which is bogus.

Originally Posted by tpad

Further compounding the problem is the ATSC/CALM spec which specifies average speech levels to be -24 dBFS (not -31 dBFS ala Dolby), which would mean that average dialog levels in a 79 dB system would playout at 75 dBSPL, which is obviously way too loud.

Neither the ATSC A/85 nor the EBU R128 or spec are based on the dBFS (or dBSPL) scale, they use the LKFS and LUFS scales respectively. In everyday use, you can't directly compare dBFS values and LKFS (or LUFS) values.

G

[tpad]

....You really need to look at the field of psychoacoustics.....

I've been "looking" into psychoacoustics long before Gearslutz's existence and probably before your's too!

To repeat the basic myth: if you calibrate your speakers to one of the standardized values (like 79 dBC), and the calibrated level sounds "too loud", then all you need to do is to move back in the room and the "level" won't sound as loud.

Of course if you do move backwards, then you DO need to recalibrate the speakers to develop the same previously established SPL. Having been in plenty of smaller rooms and done this sort of thing, I know from personal experience that there is no significant drop in loudness impact, providing that you are developing the same basic SPL. Quite to the contrary, if you move backwards into a large standing wave peak/zone, loudness impact could actually go up, not down.

In very large acoustic spaces, like in cinema sized movie theaters, where mid and HF directional horn loudspeakers are being employed, there is a noticeable difference in impact between standing up very close to the speakers (directly in the beam of the horns) and moving far back into to room, due acoustic absorption and diffusion of the horn's output combined with the smoothing effects of the room's reverberation. This does not appear the case with OP jackmack, since from what he posted, it appears that he is listening in a smaller environment. As Floyd Toole acknowledges in his book, smaller listening environments are considered to be non-Sabine (not significantly reverberant) and do not follow the acoustical rules that are commonly attributed to the cinema-sized reverberant environments.

LKFS is in fact referenced to dBFS (0 LKFS = 0 dBFS at 400 Hz). That's what the "FS" designation refers to. dBFS is the basic unit and LKFS is the derivative, apparent loudness unit, both referring to the full scale digital output of the system.

[brew]

Originally Posted by jackmack

I've recently tried calibrating my monitors to 79dB using the pink noise, -20dB, C, slow setting etc and I find it to be prohibitively loud. For a start, the working dial and fx I receive, clips when at unity and previous music mixes I have done seem overly loud when played at this calibration. Am I doing something wrong and at the risk of sounding naive, what difference does it make what the speakers are calibrated to if I can deliver a final mix at a suitable output level despite having monitored at lower levels?

Yeah, much of what you get is going to be loud bc chances are they weren't mixed with the same attention to monitor levels, so their electrical levels are compressed. Most people have their monitor pots too low and record too hot.

If you deliver a suitable level mix, obviously it does not matter at what SPL you monitored. However, the idea is that it is easier to deliver said suitable mix if you have considered the context of monitor loudness, since it is part of the system. The level that you monitor can have a dramatic effect on the average electrical level of your mix.

OT - the same measured SPL can sound subjectively louder in a smaller space, and yes LKFS is referenced to dBFS.

[jackmack]

Thanks to everybody for all this very useful advice by the way.

[tpad]

jack, you never did say what your monitoring arrangement was - nearfield I presume?

I've measured my own personal listening level at home when I sit down to watch a DVD or BluRay, and it's surprising how close it is to the cited Benjamin study 58 dBSPL level. He got it right on the money!

One reason I don't do movie theater much anymore is that the previews (aka trailers) are so friggin loud and piercing that my ears are ringing by the time the main feature starts. Don't run into this problem at home.

[jackmack]

Originally Posted by tpad

jack, you never did say what your monitoring arrangement was - nearfield I presume?

I've measured my own personal listening level at home when I sit down to watch a DVD or BluRay, and it's surprising how close it is to the cited Benjamin study 58 dBSPL level. He got it right on the money!

One reason I don't do movie theater much anymore is that the previews (aka trailers) are so friggin loud and piercing that my ears are ringing by the time the main feature starts. Don't run into this problem at home.

Yes Tpad, two nearfields, Yamaha Hs50Ms, placed about 4 feet from listening point, no subwoofer. Mid-sized room (bigger than average bedroom, smaller than average living room) and with partial acoustic treatment.

[GregorioM]

Originally Posted by tpad

Of course if you do move backwards, then you DO need to recalibrate the speakers to develop the same previously established SPL. Having been in plenty of smaller rooms and done this sort of thing, I know from personal experience that there is no significant drop in loudness impact, providing that you are developing the same basic SPL. Quite to the contrary, if you move backwards into a large standing wave peak/zone, loudness impact could actually go up, not down.

If you move into an area of the room where a standing wave is summing, the SPL meter will register the increase, requiring a lowering of the monitoring level to maintain 79dBSPL (or whatever your calibration level). This will result in the apparent loudness going down, not up. This is my experience at least. I too have been in plenty of smaller rooms and quite a few larger rooms. Again, in my experience particularly in smaller rooms the results of moving position are variable/unpredictable but with larger rooms or when comparing small rooms with large rooms, in general I find the old adage to be accurate rather than a myth.

Originally Posted by tpad

LKFS is in fact referenced to dBFS (0 LKFS = 0 dBFS at 400 Hz). That's what the "FS" designation refers to. dBFS is the basic unit and LKFS is the derivative, apparent loudness unit, both referring to the full scale digital output of the system.

Which is why I said: "In everyday use, you can't directly compare dBFS values and LKFS (or LUFS) values." Because in "everyday use" we are dealing with sounds covering a wide frequency spectrum, not only 400Hz sine waves. Also, when dealing with full scale, there are again differences when dealing with real sound (rather than a 400Hz sine wave) because with dBFS the maximum peak level (0dBFS) is the sample value, whereas with A85 or R128 maximum peak level (0dBTP) is the inter-sample value. So "full scale" has a different meaning with LKFS than with dBFS and in practical use there is no direct correlation between the two.

G

[Branko]

Originally Posted by tpad

LKFS is in fact referenced to dBFS (0 LKFS = 0 dBFS at 400 Hz). That's what the "FS" designation refers to. dBFS is the basic unit and LKFS is the derivative, apparent loudness unit, both referring to the full scale digital output of the system.

It is very important to say that, although Loudness measurement is referenced to signal level @ FS, there is no possible comparison between the two, because of the K-weighting curve applied to loudness measurement.

Back to the original question, I would suggest 73 dbC at the listening position.

[minister]

Originally Posted by tpad

To repeat the basic myth: if you calibrate your speakers to one of the standardized values (like 79 dBC), and the calibrated level sounds "too loud", then all you need to do is to move back in the room and the "level" won't sound as loud.

I don't know what myth that is.

The phenomenon, which is quite real, is setting an SPL of 85 in a room of 3000 Cubic feet AT THE MIX POSITION. Then comparing the intensity to the same setting of 85 dBC AT THE MIX POSITION in a room of over 16,000 Cubic Feet. Has nothing to do with horn-loaded monitoring. The intensity will be greater in the smaller room.

The Eric Benjamin study is referenced in the ITU BS 1770 papers. It think they were well aware of it. It is cited frequently in those kinds of papers.

LKFS. You forgot the K-weighting......

[Branko]

Originally Posted by tpad

In very large acoustic spaces, like in cinema sized movie theaters, where mid and HF directional horn loudspeakers are being employed, there is a noticeable difference in impact between standing up very close to the speakers (directly in the beam of the horns) and moving far back into to room, due acoustic absorption and diffusion of the horn's output combined with the smoothing effects of the room's reverberation.

In large theatres, if you're sitting close to the screen, you'll be actually way off speakers' axis, because the speakers are usually mounted high, so that the HF horn mouth sits at 2/3 height of the screen. It is the guys seated at the 2/3 length of the room that will be on axis, in a properly designed and built room. So, the closer you sit to the screen the bigger the loss due to off-axis response, the further you move away, the bigger the loss caused by distance.

[tpad]

...The intensity will be greater in the smaller room...

That's the myth. 85 dBSPL at the mix position is 85 dBSPL, no matter what the size of the room. If it wasn't the same, the sound level meter would give a different reading (that's kind of a no brainer!). Don't use the word INTENSITY, since in acoustics, INTENSITY means sound power, and sound power does not necessarily track in the same manner as the sound pressure level (SPL). The ear is a pressure transducer and directly responds to the SPL, not to the INTENSITY or sound power level.

To all of you LKFS "fans" here, LKFS IS NOT LOUDNESS, although that is what those promoting its use liberally keep referring to it as. Basically, it is nothing more than an RMS program level measurement with some frequency weighting added in front. Loudness is measured in SONES. As far as I know, the International community in general has not agreed to accept LKFS as a substitute/replacement for the psychoacoustical meaning and definition of the SONE.

Branko, you're overly generalizing what happens at 2/3 theater. You should come over here sometime and take a survey of movie theater architecture. They're all over the map! Large stadium sized theaters are very steep, and the look angle into the speakers at particular seating locations is very different from the more common, "flat", suburban styled theaters, where the loudspeaker beam goes right over your head. In any case, I wouldn't sit up close in either style of construction, irrespective of the look angle.

This discussion (as usual) has deviated greatly from what the OP was originally asking. What he wanted to know was if the stated calibration level sounded unusually loud. And, the answer is YES. Primarily because the "standard" calibration procedure that he was following DOES NOT accurately reflect personal preference for sound playback levels in homes or theaters. Because it is a standardized technique, a lot of first-time practitioners are surprised at their initial encounter, presumably because they are expecting "normal" listening levels.

At the time the calibration procedure was originally conceived, there was no Benjamin study to go by and so the industry sort of arbitrarily picked a level and went with it. Holman was the guy that came up with 83 dBC as being the original calibration level, and this was based on listening tests in a theater with a "professional" audience. When asked whether anyone though it was too loud, nobody objected and they just sort of went with it. Maybe the industry needs to go back and revisit some of their assumptions.

[minister]

Originally Posted by tpad

...The intensity will be greater in the smaller room...

That's the myth. 85 dBSPL at the mix position is 85 dBSPL, no matter what the size of the room. If it wasn't the same, the sound level meter would give a different reading (that's kind of a no brainer!).

Oh, it's a brainer alright! The brute SLM is just like our ears? Really? As Archie Bunker says, yer missing the pernt. Yes, the SLM will measure 85 dBC, but your EARS (brain) will tell you it sounds louder in a smaller room. Has to do with fancy stuff like increased ratio of transient response to room reverberation in a smaller room as compared to a MUCH larger room. There are other high-falutin' reasons for this real phenomenon.

As to calibration history and Holman and 83.... He was not there when they started the first alignement tests, he was later. It was Ioan Allen, Max Bell, and some other Dolby guys messing around in the Leicester Square Theater in London in the early '70's. Holman didn't work for Dolby then (he'd only received his Bachelors in '68 and his first job was with Advent Corp.). They adjust levels until people stopped complaining and felt the levels were comfortable. They measured the alignement levels determined that 0 VU was 85 dBC. What came LATER was Holman becoming involved in it and pointing out that the RMS levels of the pink that they were using were not reading the same on all VU's. VU's can have as much as 2dB of variation when reading Pink. Also, more fancy things like variations in crest factor and gaussian distribution was not consistent with the noise generators; the length of the cable affected the level; etc. It has an impact on the question of what was the real RMS value of the PINK. Was it -20dBRMS? You see, they started out with less sophisticated instruments. So Holman claimed that the reading was really 83 because of the RMS values of the source PINK, not because he thought everything should align 2dB quieter. It is an important distinction.

Here is an interesting sidebar from RP200-2002 (NOTE: it says 18 below 100%, not 20):

A.6 Historical note
Previous versions of this recommended practice were technically correct in describing the test signal level as being equivalent to 60%
modulation on an analog photographic soundtrack, and 18 dB below 100% modulation on a digital soundtrack, when measured with a
true rms meter. However, field experience shows that practically all users employ average responding meters for measuring level of
noise in day-to-day work, including VU meters to IEC 60268-17. This recommended practice recognizes the widely used conventional
practice rather than the more technically accurate use of rms derived metering, due to the readily availability of VU and other average
meters. These changes represent an effective level difference of approximately 0.6 dB — motion picture theatres set up according to
older versions of this recommended practice will play 0.6 dB quieter than theatres set-up according to procedures described in this
practice.

[tpad]

...Yes, the SLM will measure 85 dBC, but your EARS (brain) will tell you it sounds louder in a smaller room. Has to do with fancy stuff like increased ratio of transient response to room reverberation in a smaller room as compared to a MUCH larger room. There are other high-falutin' reasons for this real phenomenon....

Basically what I was I trying to explain earlier when I was rudely interrupted by the kibitzers. I agree with you that 85 dBC probably doesn't have the same psychological impact in the large room as compared with the same measured SPL in the small room, which just underscores what I have been saying all along .... that broadband RMS sound level measurements don't correlate all that well with our perception of loudness. Nobody wants to hear this because broadband RMS in the basis of the standard sound level meter, Dolby's dialnorm measurements and of course the newer LKFS spinoff. Head Acoustics in Germany are quite knowledgeable in this area and have written a number of white papers on the subject.

I'm glad your memory is better than mine, because I had completely forgotten that issue with the 2 dB discrepancy between VU and true RMS. Most of today's audio practitioners don't realize that the original standard (Weston) V.I. was not average responding but had a power law of approximately 1.4, due to the copper oxide rectifiers. So it was roughly half way between average and true RMS. As time marched on, the meter manufacturers cheaped-out and I wouldn't be surprised if today's meters no longer exhibit the same power law. Always wondered if the 2 dB difference they were seeing was with the standard or the newer/cheaper meters.

INTENSITY is in units of Watts per square-meter, and unlike SPL which is scalar, intensity is a vector quantity. Intensity is the product of pressure and volume velocity. In a true diffuse acoustic environment, the volume velocity is completely random, and so the long term average product of the two components is ZERO, meaning that intensity is also ZERO. If you've ever been in a diffuse acoustic environment (reverb chamber), it sure isn't intuitive that the intensity is really zero!

[matt82aust]

FWIW - I am most comfortable mixing at 70dBc. It is a small room, doing TV work. I would bump it up to 75 if it was going to a larger stage, but for TV work, 70 is where I am happy at and work best.

If generating pink using Pro Tools.. don't forget to check the RMS or Peak settings... Will make a difference!