NEW UPDATED Room Calibration for Film and TV Post

[Eric Seaberg]

Thanks... I'll have to think about how to route that in ProDrools... not tonight, though.

I'm trying to come up with a solution based on what I own on DVD-A and SACD and how other people are mixing... even though this is a DVD Video project, it's still music.

I have James Taylor's concert at the Beacon, NYC which was mixed 4.0. I understand the reasoning for NOT wanting a center channel, since most people spend NOTHING on their surround boxes, however, I'm not going to compensate for that. I have to assume if someone switches to the 5.1 mix, they'll have an adequate system. If not, they can listen to the LT/RT version.

I guess we can't guesstimate what EVERYONE is going to have... kinda like the Auratone/NS10 concept, eh?


Thanks for the reply.

[abluesky]

What I did on a techno-DVD that I mixed and mastered in 5.1 (and after some experimenting with low-level consumer setups), is to cut everything below say 60Hz on the main channels and send that to the Lfe-channel. That way you can be sure that the bass-management (if applied at all) will not add bass to the Lfe and that there's plenty to go on in a decent non-managed 5.1 system. Hope that helps.

Hello All:

I hope nobody minds me chiming in...

...but this is most likely not a good idea. First of all, the LFE Channel should really only be used when you run out of LF headroom in the main 5 channels and is not to be used as a "bass" channel. This is because when formats such as Dolby Digital fold down to less than 5.1 channels they typically throw out the LFE channel. Also, bass-management is a function of the playback / monitoring system (think of it as a type of summation / crossover network) and by sending everything below 60Hz to the LFE channel, you are ensuring that the summation between SAT and SUB will be less than ideal and will sound very disconnected. If bass-management is perceived to be adding too much bass to the subwoofer(s) (not LFE channel), it is only because you are either not monitoring via a bass-managed / full-range monitoring system, or because one of the systems isn't properly calibrated (monitoring / or consumer system).

Remember there are standards and practices available via SMPTE, Dolby, THX etc for film and TV applications. Attempting to re-write these standards / rules for music, or trying to compensate for "the worst case" consumer setup, will only hurt those consumers that do setup their systems correctly (and there many of them).

For more information, please visit the following links:

What is bass-management?

Tips on using the LFE Channel

More Info

I hope that helps...

Cheers!

[c.evans]

Bravo Bluesky!

I agree with sticking to the standards already in place. I tried to go it on my own with the LFE issue and I have come full circle back to the proper calibration. I have found that if I mix in 5.1 the way I like it to sound on my High End studio monitors that are calibrated according to Dolby spec it translates well to most consumer systems that I have heard.

We can't account for all home systems that aren't set up properly, but I have found that keeping the bass in music mainly in the L-R and setting the balance like I would on a stereo mix seems to translate pretty well.

Thanks, Chad

[Arno Peeters]

Thanks for the info: I did a lot of reading on bassmanagement already and compared lots of well done mixes (DTS, SACD) and came up with the conclusion that if and what to put in the LFE-channel seems more of a creative question, rather than a technical one. Of course it's true that you should not exclusively mix to the LFE, but what I figured is that in a folddown situation in a consumer environment, most stereo-speakers will not go below 60Hz and the alternative is that things get summed. In this particular case (the techno-DVD... boom-boom-boom) I've been testing it a lot, also in a another properly bass-managed studio and on some home-cinema-sets, and it all worked out well. And just to be safe, I included three menu-options: 'small range speakers', 'full-range speakers' and a stereo-option on the DVD.

Main point here is that you have to make a choice if you want to take into account the many poorly set up computer-speakers (all satellites in a row...) , cheap home-cinema-sets (with 'bass boost') and unbalanced systems (surrounds behind the couch), or that you want to mix according to the specs... I just tried to find a compromise.

On the other hand: I feel like I'm defending myself a little here, so it means I'm also still in doubt: I mix in a non-bassmanaged setup without a realtime en/decoder and just have to trust my meters, ears and instinct.
Thanks for all tips and techs: couldn't do much without this and the 'Calibrating'-tread!

[Eric Seaberg]

Thanks for the replies!! I'm still tweaking between the calibrated room I mix in and my home system, comparing it to DVD-A and SACDs I have. I have very little going to the LFE channel and are using a Genelec bass-managed sub in the conrtolroom... I'm getting more comfortable with it.


Great info, you guys!!

[minister]

Hello,

I had a discussion with another mixer over the weekend about LFE calibration. I will distinguish it from SUB because some of you use a Bass Managed system like the Blue Sky's.

There are several references in this this thread about the proper way to Calibrate LFE to be 10dB in-band over the Center channel using an RTA. Indeed, an RTA is the only true way to achieve accurate an accurate reading. If you calibrate your Center to 85 SPL, then calibrate your LFE to be +10 above that. ok, so far so good.

Yet, many of us do not have an RTA. an SPL meter CAN be used. However, due to the SPL meter ballistics, weighting, etc., you do not want to see 95, but a number lower than that, such as 89 SPL. It gets a little confusing because in at least 2 different Dolby Documents, they mention 3 different possibilities of what that number might be on an SPL meter vis á vis an RTA reading of 10dB above Center.

from One of the Links here, the 5.1-Channel Music Production Guidelines

If an RTA is not available, the LFE channel level can be set using the more commonly available SPL meter. Generally, when using wideband pink noise, the SPL reading (C-weighted, slow) should be approximately 4 to 5 dB above the main channels (that is, if the Center channel reads 85 dBC, then the LFE channel should read 89–90 dBC). Note that the SPL meter is taking a wideband measurement (not band-limited to 25–120 Hz) and therefore returns a lower value than the same
measurement on a band-limited RTA.

from Dolby 5.1-Channel Production Guidelines

If an RTA is not available, an SPL meter may be used to approximate the level. When the level
is correct , most meters will read around 90-91 dB SPL C-weighted slow for the LFE channel. The difference in level is because there is no energy being reproduced for the frequencies above 120 Hz (80 Hz for consumer applications).

As you can see, it can be 89, or 90 or 91 dBC as read on an SPL meter. How could this be 2 dB discrepancy be?

Upon further investigation I discovered (was told by a reputable source) that 91dBC is the correct number for a *good* SPL meter. Lesser meters, such as the Radio Shack, may have problems interpreting the power of frequencies that low so will show 89 dBC.

The proper way to find your LFE level is to hit your +10 above Center mark as shown on an RTA and then pull out your SPL meter and see what that reads. If you use the Radio Shack, you will most likely see 89.

Will the 2dB make a difference in LFE content? Maybe not much. But, I hope, at least, this helps readers understand LFE calibration a bit better.

[Arno Peeters]

Thanks Minister: this indeed helps.
I'm wondering: what RTA's are we talking about? Handheld devices or do some of you use software for analysis?

I tried doing it with several software packages and plugins (ETF, Fuzzmeasure, Inspector XL, TrooTrace etc.) but I was never sure how to set the mic-gain on the preamp/mixer for a proper reading: I read a suggestion somewhere to actually measure the electrical output of the mic first....?! (I use a Behringer ECM8000).

Can someone shed some light on this?

[Richard Fairbanks]

You need an acoustic calibrator for software without an integrated microphone and preamp, and really should have one for any measurement. They are designed to deliver a known acoustic power to your measurement microphone's capsule. It is not important what actual signal level the microphone puts out because you adjust your meter (preamp/software or whatever) to read exactly what the calibrator is pushing into your microphone. Calibrators come in many flavors. You may have noticed that real measurement microphones come in one of about three physical diameters: 1 inch, 1/2 inch, or 1/4 inch. These are standards. Calibrators are designed to accept microphones of at least one of those sizes. Many calibrators include adapters for more than one microphone diameter. If a measurement microphone is sized differently than one of those three, the manufacturer will most certainly have an adapter to make it fit one of the standard sizes. In the old days, electrical induction direct to the capsule was popular because no one could really trust acoustic methods. These days actual acoustic generators are more common for us normal folks. Several calibrators deliver 94db or 104db (typically at 1Khz), to your microphone. It is a sine tone, usually with a specified amount of distortion (1% or so). Some mechanical models, piston phones, deliver more like 114db (or more), usually at a lower frequency but also a sine tone. Piston phones are very accurate and have been the lab grade standard for a long time, as long as you can measure barometric pressure and temperature and apply the proper compensation from a chart. Many electronic designs nearly match them in accuracy and automatically compensate for atmospheric variations. Basically, you plug your microphone into the calibrator (it is an air-tight fit) and adjust your SPL meter to read the same SPL as the calibrator is pushing into it (plus or minus a small compensation value for different microphones). It is best to do this before every important measurement, to be certain the meter or microphone has not drifted out of calibration, and the calibrator should be checked yearly or more often by a good lab.

Once you set the level of your microphone/meter using a calibrator, you are on your own. If the microphone does not have a flat frequency response (or flat enough for your needs) you need to know the deviations and compensate for them. This is beyond most people's resources, obviously. It certainly is beyond mine. You and I have to trust that the microphone and meter are relatively flat and accurate. With really expensive measurement microphones you will get a calibration certificate that shows you its electrical output and frequency response errors. You can apply those errors to your readings. They may change over time and need to be checked about once a year.

I own a Behringer ECM8000. It is not what I use but it ain't bad! Just don't believe anything it tells you above 8Khz. The one I own looses sensitivity above that frequency. I'd tell you how much but I don't really have an accurate answer. I don't own an acoustic lab but I know mine is several db low by 12Khz. It does not noticeably affect SPL readings, there is relatively little acoustic power in the upper-most octave, but you canNOT rely on it for speaker eq in the higher frequencies.

[abluesky]

Hello,

I had a discussion with another mixer over the weekend about LFE calibration. I will distinguish it from SUB because some of you use a Bass Managed system like the Blue Sky's.

There are several references in this this thread about the proper way to Calibrate LFE to be 10dB in-band over the Center channel using an RTA. Indeed, an RTA is the only true way to achieve accurate an accurate reading. If you calibrate your Center to 85 SPL, then calibrate your LFE to be +10 above that. ok, so far so good.

Yet, many of us do not have an RTA. an SPL meter CAN be used. However, due to the SPL meter ballistics, weighting, etc., you do not want to see 95, but a number lower than that, such as 89 SPL. It gets a little confusing because in at least 2 different Dolby Documents, they mention 3 different possibilities of what that number might be on an SPL meter vis á vis an RTA reading of 10dB above Center.

from One of the Links here, the 5.1-Channel Music Production Guidelines

from Dolby 5.1-Channel Production Guidelines

As you can see, it can be 89, or 90 or 91 dBC as read on an SPL meter. How could this be 2 dB discrepancy be?

Upon further investigation I discovered (was told by a reputable source) that 91dBC is the correct number for a *good* SPL meter. Lesser meters, such as the Radio Shack, may have problems interpreting the power of frequencies that low so will show 89 dBC.

The proper way to find your LFE level is to hit your +10 above Center mark as shown on an RTA and then pull out your SPL meter and see what that reads. If you use the Radio Shack, you will most likely see 89.

Will the 2dB make a difference in LFE content? Maybe not much. But, I hope, at least, this helps readers understand LFE calibration a bit better.

Hello Minister:

There are couple of other things going on when you calibrate the LFE channel as a separate channel. If you are using full-bandwidth pink noise as your test signal, the SPL meter will bounce around a lot, meaning you will have to do a mental average. Additionally, the SPL measurement of the LFE channel, which should be 10dB above the main channel level, is not a the same bandwidth as a main channel, so you it will read lower in level on the SPL meter when compared to adding 10dB to a main channel (95dBC). Additionally, it also depends on the response of the subwoofer in the room. So, generally we say it should measure about 89dBC, using full-bandwidth pink noise.

If you are using our bandwidth limited LF test file, which is level compensated for the limited bandwidth, then the LFE channel will measure about 95dBc (10dB above the bass-manged subwoofer channel).

As side note to this: In our systems we don't calibrate the LFE channel as a separate source, we just add 10dB of electrical gain to the LFE channel, as is specified by Dolby / DTS etc. Since the subwoofer is calibrated, the end result is the same.

Below is a link to an image, which shows the response of the bass-managed subwoofer, along with the SAT and the LFE channel, if you were using 1/3 octave RTA, along with the SPL meter readings.

http://www.abluesky.com/user/images/...bassmngt_2.jpg


Enjoy!

[minister]

Hi Pascal,

thanks as always for sharing your knowledge here. i have read your documentation and downloaded your files.

you'll notice that i said 'in-band". 'In-band gain' i understand to be the relative level within the bandpass of your intended measurement as it is displayed on your Real-Time Analyzer's display. if pink noise and a sound-level meter were used to set equal levels between one loudspeaker extending to 125 Hz and another extending to, say, 500 Hz, the sound pressure levels in the pass band of each loudspeaker would be significantly different.

i think we are on the same page here.

but, you said what was my original point of all this:

So, generally we say it should measure about 89dBC, using full-bandwidth pink noise.

Fine. and if we define, as Dolby does, the LFE to be 3Hz to 120Hz, then they have 3 possible numbers. you said, generally 89dBC. Dolby says that too. and they have also said 90-91. the 2dB may not change the world or a mix to the extent that it does not translate at all. after all, there are more to translation issues room-->theater or dub stage-->theater... yet i wanted to bring up the issue for those reading the documentation. maybe it is not a big deal. but the discussion of it may help us all understand LFE calibration a bit better...because, frankly, it is a little tricky when you are confronting it for the first time or you don't have an experienced engineer guiding you.

surely, Pascal, you have made some excellent efforts, on your site and on the forums, to explain to users the proper procedures and issues for aligning speakers, so let me know what insight of yours i am missing here.