As promised, this is a quick little guide for people participating in the Sound Design Challenge on how to calibrate the listening level of your monitors (speakers). Meet your friend, the Sound Level Meter!
This device measures the overall loudness of acoustic waves in the space it occupies. It is somewhat directional, and that silver looking nub at the top of the image is the microphone for the device. These come in a variety of makes, sizes, and cost levels. The particular meter I have was bought at Radio Shack a number of years ago for about $30. The price has gone up in the last few years apparently, but the same device can be found here. If the new $50 price tag seems a little steep to you, don’t worry. There are a variety of these devices available on the market, and one of them is bound to fit into your budget. Just take a look at these listings on Amazon; from $25 up to $600.
You’ll need two things to calibrate your monitors, a pink noise file and one of our fancy SPL meters. These meters come with a number of different “weighting” systems built into them. The industry standard is to use a “C-weighted” curve (refer to your meter’s manual to learn how to select weighting), “slow” setting. If your meter does not have that, it’s ok to use “A-weighted” instead; it’s better than nothing. Pink noise, as you may or may not know, is a signal that has equal power across octaves. [ed. Special thanks to Dustin Berta for pointing out my earlier stupidity on this.] In particular, you want a -20dB (full scale) pink noise file. You can get these a number of ways. The quickest way may be through your DAW (such as Pro Tools’ Signal Generator plug-in), but you can also download a useful set of test/calibration files from Blue Sky (this page is great if you really want to get into the nitty gritty of calibration) or purchase a reference CD when they’re occasionally available from AES (I bought one at the Nashville conference last year for $10). What we want to be able to do when calibrating, is measure one channel/speaker at a time. So, route that pink noise to just one of your speakers and you’re ready to start.
First, we need to hold the meter in such a way that minimizes the effect our body has on the measurement. To do this, lay the meter flat on your palm perpendicular to your hand, and hold you arm straight out in front of you. From your listening position, point the meter at the active speaker. This means the speaker is going to be to your side. You should look fairly similar to this image:
Now, I was holding the camera about level with my head; notice that meter is at about the same height. From this position, I can easily see the display readout, but my body will have less effect on the measurement.
At this point I need to mention that there are two primary standards that we deal with for listening level when mixing for film and video. For television, we want the speaker’s output of that -20dB-fs pink noise file to read 78dB-SPL on our meter. For film, the standard is 85dB-SPL. To actually calibrate the monitors to one of these levels follow these steps:
- Load your -20db-fs pink noise file into your DAW.
- The volume on the track for this file (in the DAW) should be at 0/unity. Do not adjust the volume on the track up or down.
- Route the output to one speaker only.
- Hold the meter as previously described, and determine the speaker’s current playback level.
- Adjust the gain/trim for that speaker up or down towards your spec (78dB-SPL for TV, 85dB-SPL for film). If you have “active” speakers make this adjustment on the speaker itself. If you have “passive” speakers make this adjustment at your power amp. [Somesystems may have a function that allows you to make these adjustments for each individual speaker without searching for its trim control, such as the Avid Icon/X-Mon system for Pro Tools that I have at work]
- Repeat steps 4 and 5 until the speaker’s output stays at your preferred playback level.
- Repeat steps 3 through 6 for the remaining speakers in your system [note: if you have a sub woofer, you'll want to get the band limited pink noise file offered on the Blue Sky site.]
That’s all there is to it. Unless you’re in a massive room, it’s not in your best interest to use a film spec playback level (85dB-SPL); and I would suggest you use the television spec instead (78dB-SPL). [Chart with room size and appropriate calibration levels below.] It is strongly encouraged that you calibrate your monitor levels before participating in the Sound Design Challenge. Even if you’re not participating, I’d still recommend it; at least you will be certain that your speakers are all producing the same levels. That means your spatial ideas (stereo, 5.1, etc.) will be more likely to play back correctly on other systems. You should also regularly check you calibrations to make sure nothing has changed. It’s just one of the many steps we need to take to maintain the quality of our mixes.
Update: There’s also some great info in this posting over on the Avid Audio Forum (formerly DUC).
Addendum:What level should you calibrate to for different size rooms!?
This info section comes from a secondary post I wrote nearly two months after writing this initial tutorial. It can be found here. I’m adding this info because it may [should] be useful to some [all] of you. ;)
Essentially, the smaller your room is, the lower your listening level should be. It’s just a simple fact of the way we hear that higher levels in smaller rooms are going to give unnatural (well…maybe supernatural) emphasis to different frequencies. Listening at appropriate levels will help maintain the integrity of your mixes as they move from one system to another, in rooms of varying sizes. It’s the main reason I suggested using 78dB-SPL in the previous tutorial. Unless you’re in a mixing stage that is the size of a movie theatre (and has similar acoustics), 85dB-SPL is too loud for you to be mixing at.
In the interest of keeping this simple I’m going to stick with their basic guidelines for small rooms. They have an expanded spec for really small rooms, but it’s going to be beyond most of our abilities to replicate it. Their suggested listening levels for rooms used in final mixing are:
- Greater than 20,000 cubic feet (566 cubic meters) - 85dB-SPL
- 10,000 to 19,000 cubic feet (283 to 565 cubic meters) - 82dB-SPL
- 5,000 to 9,999 cubic feet (142 to 282 cubic meters) - 80dB-SPL
- 1,500 to 4,999 cubic feet (42 to 141 cubic meters) - 78dB-SPL
- Less than 1,499 cubic feet (41 or less cubic meters) - 76dB-SPL
If you’re not sure how to determine the cubic dimensions of your room, measure the central axis of it’s length, width and height and multiply all of those figures together (length x width X height), and that will give you the cubic feet of your room. I say measure the central axis, because if you’re room is an odd shape, it will at least get you in the ball park. If that ball park is right on the line between two ranges, I’d suggest you round down to the lower level.
So, if my room was 11ft x 13ft x14 ft (these numbers chosen at random), that would yield a cubic dimension of 2002 ft. I’d then want to calibrate my room to 78dB-SPL.
As always, if you’re smarter than I am (which isn’t very hard, I assure you), and you’ve spotted something in my text here that is inaccurate or doesn’t make sense, please let know. I’ll be happy to make changes and credit you for them.
Great article! Is that your board dude?
Thanks. Technically it’s not “my” board, but it is the one I use every day at work. ;)
Fabulous idea! I am a big advocate for standards. This will help the consistency in the coming months.
I would also like to add one thing. I don’t change the levels during the day. As I work I know if a sound is loud or quiet because the control room is a constant. This greatly improves the consistency of each sound I make. I suggest documenting what the settings are on the ampsspeakers.
RoomEQ is also a big help when you want to take calibration to the next level.
That’s an excellent point about volume consistency when working. I’m not sure I agree on the room EQ idea though. I’ve spoken to a bunch of mastering engineers over the years, and (for the most part) they don’t like the idea of EQ-ing a room. They’d much rather people take the time to get to know the sonic characteristics of their rooms as they relate to the outside world. It refines your sense of hearing and overall improves the level of your work. I think I agree with them, it can be dangerous to rely on tools too much. Those sort of applications are much more useful in a live audio environment, where the space can throw monkey wrenches all over the place.
Hi all
Im glad to see we are getting a discussion on loudness.
Im a little puzzled by your no. 2 Shaun. You say that we cant adjust the individual levels on each track, and should only adjust on the speakers. But arent we supposed to fade up or down for each track, in order, for the DAW´s internal meter to read -20 dbfs in the first place?
Best wishes,
Mikkel
No, you should not adjust the volume on the track. The pink noise file you use for calibration should be a -20dB-fs pink noise file. For instance, if you’re going to use the Pro Tools “Signal Generator” plug-in, you would select pink noise, and set it’s level for -20dB, then click on process to create the file. Alternatively, if you download a calibration file (such as those at the Blue Sky link), they will already be -20dB-fs. So, the DAW’s internal meter will read/show -20, because that is the signal strength of the file. The idea is that everything in your monitor path is at nominal/unity. This is the only way to ensure that all of your monitors will be calibrated to the same level. That one trim/gain/volume knob is the only variable in the equation.
Hi Shaun
Thanks alot:)
I dont understand no.2 from the Bluesky pages saying:
Remove all eq and dynamics from the signal path and set all controls to zero / unity gain.
Play the 1kHz Sine Wave, hard assign it to the left channel only, and adjust the output fader so the so the output meter reads -20dBFS
Best wishes,
Mikkel
That’s for calibrating your electronics, particularly if your going through an analog board or if the trims of your I/O haven’t been calibrated. Depending on your system, you may not need to do this.
If I may take part to the conversation, I think Mikkel is getting confused between track and channel. A channel corresponds to a speaker and a track is a playlist. Your hardware setup determines your number of channels (stereo = 2, 5.1 = 6, etc.) but you have a virtually unlimited amount of tracks available.
I have active speakers, so here’s how I’d calibrate (if I had one of these bloody SPL meters!):
– turn my monitors’ gain way down
– crank my MBox output to the max
– play the pink noise (that should peak at -20dBFS)
– progressively crank my monitor’s gain back up until I get an 85 SPL reading.
From what I understand, the part from the Bluesky guide that you’re wondering about helps you set the position of your master fader (if you have one) so it is transparent. They just insist on removing all processing simply because it’d screw the parameters of the equation. In clear, all they tell you is:
– take the file that contains the signal at -20dBFS and don’t process it in any way
– set all your faders to 0dB (unity gain)
– play the file
– you should read -20dBFS on your output. If you don’t, fiddle the master, not individual tracks.
That only means that from the data to your output, everything is clean because output level = input level.
Thanks for helping out.
Fiddling with the master is exactely what Im doing, not the other tracks. Sorry for confusing.
If I dont do this, my meter shows 2.7 db from -20 DBFS, both in the Digi phasescope (which shouldnt be trusted in the first place I know) and on the HR meter from Smassey.
Mikkel
Guys, you shouldn’t be fiddling with your Master Fader either. Everything, and I mean everything, should be at unity. If you have a plug-in meter that’s telling you the file is reading something other than -20dB-fs, then there’s probably something else going on. What is the meter set to display (peak/rms/etc.)? If the audio file is stereo, try splitting it out to mono and using only one channel of the file. See what happens then. If you adjust the Master Fader, you’re introducing another variable that should not be there.
Yes, and I dont think my I/O was calibrated right.
Anyway, Im looking forward to all of us having the same standards.
Mikkel
Ok just tried to set up some different metering.
In VU mode the files shows 0 db, which means its right, right? and everything is working properly.
In RMS mode it is 2.7 db from -20 db.
This is the Digi phasescope.
Regarding the VU meter….not necessarily. ;) All these meters and scales get a little tricky to correlate to one another. I can’t remember if the Digi Phasescope lets you look at peak instead of RMS, does it? If it does, switch over to peak measurement and tell me what it says then.
Thanks for helping out Shaun.
In peak mode it reads -20. So everything is working well.
What confused me more was, that when reading on the master meter I wasnt able to get a reading on the Leq window. It said -inf no matter what I did. When I placed both the HR meter and the Phasescope in the individual channel, it jumped to -20.
Sorry for the confusion, all!
Glad its working now.
Apologies for telling you to fiddle the master, even though I wrote it I find it shocking too…
Pingback: Amplifiers: 720
I read that Justin fiddles his master? (try your monitoring controller, Crookwood has some nice ones) up to 85 dB, but Bob Katz advises to use a 73 dB for a perfect listening signal.
Note also that with 85 dB you are only allowed to work in it for 8 hours a day for EU standardisation, otherwise you are not insured for hearing loss.
And Shaun, The Icon!, naturally you also are on yuotube in Vegas-Mode!
Do you also use that hideous 192 I/O’s?
If so, connect a Grimm CC1 or an Apogee Big Ben as the masterclock, and then listen to your files again!
(I did the test in SAE Amsterdam Studio 1)
Of course, Tony. I’m always in Vegas mode. lol
That’s a good point you bring up regarding 85dB and EU standardization/OSHA (US body), but that’s only if you’re level is consistently 85dB. If you have sustained sounds that get above that (which definitely happens when mixing), then the amount of time you can sit there without damaging your ears decreases significantly.
That’s why I suggest people use the 78dB spec (Bob Katz works primarily in the music industry, so his 73dB suggestion is more aimed for music and mastering work). Heck, I know of a number of professional sound mixers in LA who wear earplugs (the expensive, full fletcher-munson frequency curve type) when mixing the big pictures!
Pingback: Addendum: Monitor Calibration Expanded | Dynamic Interference
hi just been reading all the poat and im having the same problem,i cant get the pink file to reed -20, the sine file does but none of the others? kind regards
Are you referring to your reading on the SPL-meter or in the DAW?
Pingback: Electrically Calibrating Your I/O | Dynamic Interference
Hi,
I appreciate the info and the site. But it bothers me that you listed pink noise as equal energy per frequency spectrum, which almost sounds like the definition of White noise – equal energy per frequency, where as the energy of Pink noise is equal energy per octave.
This really bothers me and makes me second guess any new information this site may offer me.
Thanks
Huh! So I did. How the hell did I make that mistake?!
Thank you very much for pointing that out, Dustin. The correction has been made.
I’ve got some questions about the calibration instructions. You said, ” From your listening position, point the meter at the active speaker.”
Blue Sky says, “SPL should be measured at the mix position, with the SPL meter at arms length, with the microphone at seated ear height, angled at approximately 45 degrees, and pointed at the center point between the left and right speakers.”
I think Blue Sky wants the meter pointed straight between the left and right monitor while I think you want us to point the meter straight at the active speaker. Likely it will make little difference but … well … little things like this just jump out at me :)
The part in the Blue Sky instructions about “angled at ~ 45 degrees” is baffling. Do they mean that the meter shouldn’t be parallel with the floor? I don’t think so but what other angle is there? Rolled at 45 degrees?
Anyway, thanks for the calibration advice. I can’t believe, after all these years, this is the first time I’ve run across instructions for calibrating the monitors.
A lot of SPL use omnidirectional microphone elements. I’ve never noticed a marked difference between pointing them on axis or on centerline between two speakers. On the off chance that your meter doesn’t have an omnidirectional element (and has something along the cardioid lines), pointing at the center speaker only to measure Left, Right and Center might actually weight your speaker balance slightly. So, I just prefer to aim at the speaker I’m actively measuring to ensure consistency across the soundfield. It’s a redundancy thing.
The 45 degrees thing is baflling to me as well. The only thing I can think of is that if you’re holding the device in a similar fashion as I am in the picture provided in the post, rolling it towards you allows you to more easily read the measurement (typically how I have to read the display anyways).
Hello everyone,
I am a novice sound designer, so I still don’t know a lot, but I’m doing my best to learn as much as possible. Anyway, I just finished calibrating my room, and to tell you the truth it’s extremely loud! I played with level and figured out that the most comfortable level for me is 61.2 dBSPL per speaker at -20 dBFS pink noise. That’s in my 1560 cubic feet room. I’m just wondering why do we need it to be so loud? And do we need it to be? My common sense tells me that if my body feels bad about being in such noisy environment, then, probably, it’s a good idea to turn the volume down. :)
My first thought is that either your mixing things too loud, or you’re listening to mastered pop music for reference (which is most definitely deafeninf if you listen at these reference levels). As far as mixing goes, the film standard is -20dBFS pink noise equals 85dB (technically 83dB in each individual channel…it has to do with how they sum) and television is 78dB. Notice that the TV spec is a lower SPL volume than the film, because television is mixed hotter and with less dynamic range than film. Web video is mixed even louder, in which case 78dB reference would be uncomfortable as well. Part of the reason we mix at these levels is for balance in the perception of the frequency spectrum. Start getting much below these, and our perceptions of the spectrum start to get colored. This means that if you mix at the level you mentioned (61.2) you probably won’t be hearing the frequency spectrum of your work accurately; it won’t translate to other systems the way you expect it to. When you do listen on other systems, the bass frequencies may be totally out of whack with what you expected.
Thank you for all the tips. In some time it became obvious what I was doing wrong. I was referencing mastered rock tracks, and that threw things off. It made more sense when I played some movies. Dynamic range is completely different. Now I use the same setting for both film and games, and it works just fine.
With best regards,
Alexander.