Calibrating your monitors: audio and video

Posted on July 8th, 2007 in Film technique by emre

Have you ever wondered why your project looks and sounds wrong on someone else’s equipment? This happens because their output devices are configured haphazardly and behave differently natively. So what can you do about it? Calibrate your equipment to a common standard, of course. In this brief article I will show what can be done on a small budget.

Video

I use a pair of 19″ CRTs for all my editing. If I wanted to go all out I would add a television monitor to the mix. I do not feel an urgent need to do this as I know my CRTs are showing me accurate colors. I routinely calibrate and profile both of my monitors with Colorvision OptiCAL. The way it works is that I attach a small hardware device (called a colorimeter) to the screen and run the calibration utility. The utility generates a series of colors, and the colorimeter measures them so the utility can determine the difference between the measured value and the proper value. The result is an ICC profile.
Microsoft Color Control Panel
To take advantage of the ICC profile I use Microsoft’s Color Control Panel to load the LUT when Windows starts up. Once the applet is installed, you need to place wincolor.exe in your Startup folder and run it as follows:
"C:\Program Files\Microsoft Color Control Panel\WinColor.exe" /L
This will ensure that your profile gets loaded when you start Windows. Macintosh users will naturally use ColorSync. Isn’t it amazing how similar MCCP and ColorSync look? I wish Microsoft would copy Safari too, and implement color management in Internet Explorer so we could take it for granted. Maybe in five years from now, when the next version of Windows is released…

Once the LUT is loaded, you need to tell your editing application about it. If Windows had been designed with color management in mind, applications would do this behind the scenes with a few API calls, but it was not designed that way, so here we are wasting time configuring our applications. I use Sony Vegas, so I go Options>Preferences>Preview Device and select Windows Secondary Display from the drop-down list.

Sony Vegas' Preview Device options

Let’s take a closer look at the configuration above. Basically I instructed Vegas to show the full-screen video on my second (right) monitor at its current resolution. The virtue of this is that it avoids constant resolution changes that would occur whenever you start and stop previewing, which is not good for CRTs. If you use an LCD, you will of course want to use its native resolution at all times. For obvious reasons, you want your video to fill the screen rather than be too small (SD) or too large (HD), so you tick Scale output to fit display. Also you don’t want to get a headache looking at interlaced footage so you tick Apply deinterlace filter. However, the most interesting option, as far as this article is concerned, is the Use color management option, which you should also tick. This tells Vegas to take the video from its native color space (e.g., NTSC DV) and convert it to your monitor’s color space, as defined by the Monitor color profile. If you use Vegas’ native DV codec, tick the Use Studio RGB option too. This tells Vegas that a luma setting of 16 actually corresponds to RGB (0,0,0) and a luma setting of 235 corresponds to (255,255,255). In other words, it expands the luma scale, but this is only true of certain DV codecs. This is what the help file has to say:

DV Codec Studio RGB Setting
Sony On
Microsoft Off
Matrox Off
Mainconcept Off
Canopus On
Quicktime Off

Audio

Mixing in surround is de rigeur so that is what I assume. There is not much to configure with a pair of stereo speakers anyway. With regard to the speakers, there are two options: self-powered and amplifier-powered; I have the latter, for historical reasons. This means that my sound card interfaces with my receiver. Again, there are two options: digital and analog. For the former, there are once again two options: coaxial and optical. It does not matter which you use, since digital is digital; my m-audio sound card uses coaxial. To digitally transmit audio the audio has to be encoded (to define what exactly the ones and zeros mean). The two standards are S/PDIF (consumer) and AES/EBU (professional). These standards support PCM stereo, AC3, and DTS. Unfortunately, there is no consumer standard for PCM surround, except for the ones defined by HD, and you certainly can’t get that at the consumer production level yet. Unfortunately this means that you have to work in analog…

Six channels means three pairs of stereo cables: front left & right, surround left & right, center & LFE. Connect these from your sound card to your receiver, then go into Vegas’ Audio Device options. If you are lucky (or you chose well), your sound card should have an ASIO driver, for reduced latency (processing delay). My preferences look like this:

Audio Device preferences in Sony Vegas

Once you have properly connected your sound card to your receiver, you still have to properly position the speakers and calibrate them. This is pretty complicated stuff, and expensive to do right; the room’s shape and materials have to be taken into consideration. That is what recording studios are for. I assume people reading this do not have the means for that, but there is still a lot that can be done.

First, let us consider the speaker positions. I used Dolby’s Surround Mixing Manual as my guide. Sections 2.2 and 2.3 tell you all you need to know, so I will not repeat them here. Ideally you want all your speakers to be identical, positioned at eye level, so you can pan the sound from one to another without noticing any difference. If the Dolby manual isn’t enough for you, Wendy Carlos goes into excrutiating detail.

Once you have properly positioned the speakers, you will want to calibrate them to the correct sound level, and the way to do that is with a sound level meter (cf. 2.3.9). I picked up an Extech 407735 off eBay for $70. The idea is to get each speaker to generate a particular sound level when fed a particular signal. Dolby calls for an 85dBC (front) or 82dBC (rear) SPL, slow response with -20dBFS pink noise. The LFE is more tricky since sound level meters are not said to be as accurate at such low frequencies but I ignored the warnings, for lack of a better option. Of course, there is the possibility that your SPL meter might need to be calibrated but let us not get bogged down :) You can use these audio files for your own calibration. I use band-limited pink noise in order to reduce room interaction effects, as explained here.

If you really want to polish it off you can equalize your monitors to linearize their frequency response with a spectrum analyzer.

Leave a Reply

You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>