The No.1 Website for Pro Audio
Please explain the "extra 10dB headroom" for LFE channel
Old 15th January 2013
  #1
Gear Nut
Please explain the "extra 10dB headroom" for LFE channel

I'm just about getting my head around 5.1 surround but would be grateful if someone could clarify the "extra 10dB headroom" that an LFE channel has. Does this mean that during PLAYBACK, playback systems will take the signal from the LFE channel and boost it by 10dB compared to the satellite and redirected bases (thus allowing the LFE level on the delivered media to be 10dB lower)?
Presumably, having calibrated the physical subwoofers to whatever your required SPL (85dB for example) this extra headroom business doesn't just mean turning the gain on the subwoofer amplifier up 10dB - otherwise this would also boost the redirected/bass managed signal from the surrounds.

I'm using the Waves plugin surround bundle which includes software bass management. One of these, the Waves 360 Surround manager plugin is intended to be engaged for monitoring only - and disabled during printing of the audio to file. The LFE channel in this plugin dialog has an option in a drop down (-10, 0 or + 10). Presumably when monitoring, I should choose the +10dB option - correct?

Seems like you could end up with wildly varying levels of bass if you don't get this right!

Any insights?

Cheers
Old 16th January 2013
  #2
Gear Maniac
 

Quote:
Presumably, having calibrated the physical subwoofers to whatever your required SPL (85dB for example) this extra headroom business doesn't just mean turning the gain on the subwoofer amplifier up 10dB - otherwise this would also boost the redirected/bass managed signal from the surrounds.
Correct! They are treated separately. LFE channel is a sound source. Redirected bass is just a workaround for monitor speaker deficiency. Separate problems, separate treatment. The difference in gain is most conveniently treated within your monitor speaker controller (big knob) that can create a redirected bass signal upon demand, and combine it with LFE signal (with +10db applied to the LFE signal before combining), and sends the combo to a subwoofer speaker(s).

Always apply +10db gain to the LFE before it hits the subwoofer, and before it is added to redirected bass from the other channels.
Old 16th January 2013
  #3
Old 16th January 2013
  #4
Gear Addict
 

I must ask as well, what are you mixing?

This +10 is more for theaters, and not much else.
Old 16th January 2013
  #5
Lives for gear
 
minister's Avatar
Quote:
Originally Posted by postprosound View Post
I must ask as well, what are you mixing?

This +10 is more for theaters, and not much else.
Actually, Richard has it correct in post #2.

It is true that in theatrical mixing, the subwoofer is aligned +10 in-band gain above the screen channels. This assumes no bass redirect. The LFE is not +10, the subwoofer is.

In DVD/consumer home theater, the channels are the same all around (no -3dB for the surrounds) but the LFE channel will be +10.
Old 16th January 2013
  #6
Gear Addict
 

Quote:
Originally Posted by minister View Post
Actually, Richard has it correct in post #2.

It is true that in theatrical mixing, the subwoofer is aligned +10 in-band gain above the screen channels. This assumes no bass redirect. The LFE is not +10, the subwoofer is.

In DVD/consumer home theater, the channels are the same all around (no -3dB for the surrounds) but the LFE channel will be +10.
Good catch! I thought I had edited that before I posted it.
Old 16th January 2013
  #7
Lives for gear
 
mikevarela's Avatar
 

Tom,

just to clarify.

Home setup, non theatrical.

calibrate 5.0 normally, no drop in Ls Rs - all at some db, say 85 for brevity. Now what for the sub, also to 85 with same full range pink as the rest using say a radioshack full band spl meter?

where does the +10 come in in this situation?


thanks man, appreciate the help
Old 16th January 2013
  #8
Lives for gear
 

You really need a RTA meter to set the level of a sub. A standard SPL meter will not work as well.
The way cinemas/dubstages are calibrated by dolby is to replay dolby pink noise at reference level. Calibrate to 85 dBC slow. Look at the RTA and switch the feed to the LFE on/off. The pink noise peak level in the LFE frequencies should then display + 10dB in average compared to the regular pink noise level.
With simple SPL meters this will result in a SPL reading that will vary quite a bit! I have seen differences in readings from 88-95 (I guess it can be be even worse)... And it's mostly at the lower end if that swing. So if you calibrate that way, your replay level of the subwoofer will likely be to hot and your mastered level will be to low.
The inaccuracies of measuring low frequencies are simply to great just using a SPL meter.
The good news is that wit a decent flat measurement mic yôu can use any good spectrum analyser to check this, but learning how to interpret those readings will take a little time.
Old 16th January 2013
  #9
Lives for gear
 
mikevarela's Avatar
 

Thanks Erik. I hear what your saying. Full band pink measured at 85dbspl with RTA changing sub volume up until visually you see +10 db in the low (say 85hz and lower) region.

makes sense, though I've used some RTA software and damn does the low end look messy sometimes.

but... what about a smaller studio for non theatrical material. Say TV or Web mixing? Nearfeild style.
Old 17th January 2013
  #10
Gear Nut
Quote:
Originally Posted by postprosound View Post
I must ask as well, what are you mixing?

This +10 is more for theaters, and not much else.
I'm not mixing anything yet - I'm just getting my brain around it all!

The Film Tech info was very helpful in understanding this - thanks for that.

So, if I've understood correctly, turning the recorded level of the LFE channel down 10dB (and boosting it 10dB during playback) is an historical artefact from when the sound track was printed optically onto the celluloid - Correct?

So, this artefact still applies in the digital domain when producing sound for film shown/played back digitally in cinemas - Correct?

But, for films going to DVD or BluRay, or for HD TV, we don't record the LFE channel 10dB lower because it is not going to be boosted 10dB during playback - Correct?

Last edited by amathie; 17th January 2013 at 01:14 AM.. Reason: mistake in original
Old 17th January 2013
  #11
Gear Nut
The Film Tech notes on Calibrating the Sony Cinema gear and subwoofer level suggest that the LFE channel 10dB boost applied to playback should be done acoustically rather than electrically or in the digital (workstation) domain.

Sorry for this basic question: If I am playing back a mono bandwidth limited pink noise file from Nuendo (-20dbfs 40-80Hz - like one of the calibration files provided by Blue Sky) and I've calibrated the LFE subwoofer output to give say 85dB SPL at the listening position, if I change the output fader in Nuendo to boost by 10 dB will I expect the SPL to change by 10dB at the listening position?

In other words, is there a linear relationship between the digital level in a DAW and the acoustic level measured in SPL at the listening position?

This would appear to be rather central to whether or not I can apply various calibration changes in my Waves Surround Manager plugins, for example swapping between + 10dB LFE boost for film playback during monitoring and
no boost for TV work.
Old 17th January 2013
  #12
Gear Maniac
 

Simple answer without pithi-nes, no. dBFS doesn't equate to dBSPL.
You will have acoustic anomalies in your room which will excite at different levels and in different frequencies preventing a linear relationship.
Wouldn't you think?
What you hear is the combination of your DAW output, your speakers, your room and where you sit, right?
Old 17th January 2013
  #13
Gear Maniac
 

This is why the term is "room" calibration - not system calibration.
Old 17th January 2013
  #14
Gear Maniac
 

It does not matter one pip what you are mixing for. The LFE channel is ALWAYS boosted by 10db during playback, in theater, in home, in office, in swimming pool, in your SUV, in your control room, and everywhere else. Very simple rule to follow. There seems to be plenty of internet "knowledge" that is contradictory, which is wrong.


Quote:
But, for films going to DVD or BluRay, or for HD TV, we don't record the LFE channel 10dB lower because it is not going to be boosted 10dB during playback - Correct?
Not correct.
Quote:

In other words, is there a linear relationship between the digital level in a DAW and the acoustic level measured in SPL at the listening position?
Depends on what you are actually asking. Relatively speaking yes. If you set an arbitrary playback level and make a measurement of the LFE, then raise the electrical signal playback level by 5db, your acoustic measurement will increase 5db also (assuming no amplifier/speaker overload). Acoustically, though, the way low frequencies from any speaker loads a room will affect overall speaker alignment and response linearity, and your measurement position will greatly affect your measurements. Alignment without good measurement tools is hopeless (RTA with 1/6 octave resolution would be an absolute minimum I think, preferably better) Experience and lots of experimentation, and ability to compromise, help to achieve good outcomes in smaller rooms
Old 17th January 2013
  #15
Lives for gear
 
ggegan's Avatar
Just a note: When you calibrate the sub using an RTA to boost the sub bass freqs 10dB above the full frequency front speakers, you need to low pass the pink noise at at 125Hz.

BTW, rfnoise is correct. The entire purpose of the LFE channel is to reproduce low frequencies at very high level. My understanding is that the original reason for the 10dB boost was because the oxide magnetic stripe on film release prints were not capable of recording bass frequencies at the intended level at unity gain without crapping out, and the solution was to record the sub material at a lower level that the mag stripe could handle and then boost the output 10dB in order to attain the intended levels without distortion.
Old 18th January 2013
  #16
Lives for gear
 
mikevarela's Avatar
 

RF - Gary --- Cheers

Got the point on all LFE calibration being the same. thanks for that. So, low pass at 125hz and send the signal through the LFE, measure with RTA and then boost volume knob on ACTUAL sub till a reading of +10 reads back in the RTA graph. so if everything else is reading 85, this should read about 95?
Old 19th January 2013
  #17
Gear Maniac
 

No, forget about 85 or 95 when talking about the LFE and/or subwoofer by itself.

Here's what should happen. This assumes PERFECT speakers in a PERFECT room. I'm a bit long-winded but bear with me.

Play full bandwidth pink noise into one full range speaker. Set an appropriate playback level. (Most people choose a C weighted meter and adjust volume to 85dbC, but the actual level is not relevant to what I am about to describe.) Once the full range speaker is set at a level, use an RTA frequency display, (Do not use C weighting of an RTA. It must have NO weighting). Usually you see a bunch of vertical bars, each showing a fraction of an octave of acoustic energy, low freq's on the left, higher on the right. Each band, or frequency slice, should be at the same height, the same value. Lets say each slice is showing 50db, for example. Remeber that value.

Now, turn off the full range speaker and send the exact same pink noise signal, at the same electrical level, into your perfect subwoofer. Look at the RTA, you'll notice that only the bands on the left side of the screen are reading because the sub speaker does not reproduce high frequencies. In a standard setup there will be almost nothing above 160hz. Adjust the perfect sub's level until each band below 125 hz reads at 50db (or whatever the perfect full range speaker was reading). Since the sub and room are perfect, each band will already show the same value. In a real room, each one is probably different and adjustments should be made.The perfect sub is now set to the proper relative relationship to the perfect fullrange speaker. Your perfect speakers calibration is done.

If you have a monitor controller than does proper bass management, when you play music through the perfect speakers, and flip bass mangement on and off, there should be NO CHANGE in the amount of bass you hear in your perfect room. With bass management off, all sound goes through the fullrange speaker. With bass management on, the bass range is taken out of the full range speaker and is put into the sub woofer. Since the sub is adjusted to play the same bass energy into the room as the fullrange speaker is, you will hear the same thing. Get it?

Now, once the above is done, any signal from the LFE channel must be boosted 10db before going to the subwoofer. This means you can make very strong bass signals from the LFE channel, stronger than you could with a full range speaker.

If you boost the LFE channel by turning up the subwoofer, rather than applying gain to the LFE channel, the LFE channel will sound correctly but you are also boosting the bass-managed bass from your monitor controller, which will make the bass-managed bass way too strong.

I've ignored the challenges of using real speakers in real rooms, obviously. The above is for concept.

Problems come when a person refuses to use an RTA and instead relies on a single SPL meter, and uses SPL readings they discovered on the web, which are probably wrong to begin with and are certainly going to be wrong in YOUR room.

Hope this helps. Sorry for my big bag of words.
Old 19th January 2013
  #18
Lives for gear
 
ggegan's Avatar
While it is true that you boost the LFE level before it gets to the sub, it should be pointed out that the bass managed low frequencies that are routed to the sub will have a different EQ curve than the LFE. The BM lows generally are rolled off at 3dB/8va at 50Hz ( the 3dB down freq), and the the extreme lows are high passed. This is a spec of the X-Curve, and while for small rooms without long reverb build-up you don't need to do the X-Curve high freq roll-off, I do use the low freq roll-off because I have found that translates well to large rooms. The LFE, however, should not have that low end roll-off, but should be flat from about 80Hz down to somewhere around 30Hz, albeit 10dB hotter on an RTA than the un-rolled-off main speaker baseline. At least this is how I apply EQ for my room.

Specs for very small rooms are subject to some debate. Dolby's small room spec is for stages that are considerably larger than many of us work in, so it doesn't always apply 100%. None of what I say is "scripture", but based on logic and experience. When it comes to small room translation to large theaters, there is plenty of room for personal preference.
Old 19th January 2013
  #19
Gear Maniac
 

Quote:
it should be pointed out that the bass managed low frequencies that are routed to the sub will have a different EQ curve than the LFE
the above confuses and complicates the basic question and is rarely found in practice

Quote:
and while for small rooms without long reverb build-up you don't need to do the X-Curve high freq roll-off, I do use the low freq roll-off because I have found that translates well to large rooms.
I know you to be knowledgeable and experienced enough to arrive at such a workaround for your particular circumstance and taste. It works for you apparently but is contrary to all recommended practice. A slightly reduced bass response (a couple of db down below 50hz, which is practically nothing compared to room modal problems) was common in theater systems during the 70s when measurements were made (upon which the xcurve was based). The low end dip of the original x curve is generally thought to be a result of deficient speakers systems in use at the time and was, even then, considered optional for a theater mixing room environment to emulate. It was not then, nor now, suggested for small rooms. Linear low response is part of AES and ATSC recommended practices. As I said, you have a setup technique that apparently works for you but it is not common practice and should not be a starting point for anyone.
Old 19th January 2013
  #20
Gear Maniac
 

Sorry, ggegan, I don't want to argue. Reading back what you wrote, I see you did say it was your personal preference to do that.
Old 19th January 2013
  #21
Lives for gear
 
ggegan's Avatar
Quote:
Originally Posted by rfnoise View Post
Sorry, ggegan, I don't want to argue. Reading back what you wrote, I see you did say it was your personal preference to do that.
Yes, I go back and forth between my small room and large Hollywood feature dub stages all the time. If I EQ my main speakers flat in the lows, the work I do in my studio inevitably sounds thin on a large feature dub stage.

BTW, every large feature dub stage I have ever worked on since the switch from the Academy Curve has been EQed to the X-Curve. Please show me a graph of the X-Curve that doesn't include the low frequency roll-off in the lows, because I have never seen one.
Old 22nd January 2013
  #22
Gear Maniac
 

Quote:
Please show me a graph of the X-Curve that doesn't include the low frequency roll-off in the lows, because I have never seen one.
The X-curve is what it is. I’m not trying to re-write history.

Quote:
The low end dip of the original x curve is generally thought to be a result of deficient speakers systems in use at the time and was, even then, considered optional for a theater mixing room environment to emulate.
My memory failed me and I misspoke, which is my comeupance for not having the necessary document in hand before posting. The first part of my statement is true, the second is not. The graph that I mentally pictured as I posted is from ISO 2969/1987, with a text note that the low end curve is optional for a small room. It is not optional for a large room adhering to the X-curve. Apology offered!

I stand behind my comment that your technique is not standard for small room alignment. AES, ATSC, and Dolby have each published recommendations for surround mixing with desired sound field graphs showing flat line curves below 2khz. I am thinking ITU has something similar but I haven’t located it, so maybe my memeory fails me again. The primary reason I commented about your post is that your statements clouded my attempts to explain a basic concept of subwoofer calibration that I thought would benefit anyone with questions similar to those of the OP.
Old 25th January 2013
  #23
Lives for gear
 
mikevarela's Avatar
 

Thanks guys. This makes a lot more sense now.

No adjustment on the actual speaker. Makes sense when switching between BM and LFE.

Are you adding the 10db to the LFE channel via wide volume boost - think automation level up 10 db or by using some plugin like gain or input of an eq plug? I'm asking because if you send material there (LFE channel) via surround panner LFE fader, then who's accounting for the 10 db boost?

RF - I appreciate the thorough explanation. It helped a lot.
Old 25th January 2013
  #24
Gear Maniac
 

Quote:
by using some plugin like gain or input of an eq plug?
Those can work. Just make sure you don't print the final mix outputs that way. The gain happens only for playback. It would be done on a surround buss that is specifically for speaker control. In PT you could use a mono master fader instead, applied to the LFE channel of a surround buss, which would not introduce any latency or require any additional processing power. Still, that buss is only for listening not for printing the final mix and you need to make sure that 10db gain does not causing signal clipping. Usually, this is not a problem but could be.

Quote:
I'm asking because if you send material there (LFE channel) via surround panner LFE fader, then who's accounting for the 10 db boost?
The gain happens after this point. You often don't have to push the LFE fader very high to get strong sub action.

I use a monitor controller that can gain up the LFE channel by a separate user-specified value (which I have set at +10). This is independent of any bass redirection and other speaker trims. So, I don't have to worry about doing it within a workstation.
Old 26th January 2013
  #25
Lives for gear
 
mikevarela's Avatar
 

totally get it now.

you're the man.
Old 29th January 2013
  #26
Here for the gear
 

Quote:
Originally Posted by ggegan View Post
While it is true that you boost the LFE level before it gets to the sub, it should be pointed out that the bass managed low frequencies that are routed to the sub will have a different EQ curve than the LFE. The BM lows generally are rolled off at 3dB/8va at 50Hz ( the 3dB down freq), and the the extreme lows are high passed. This is a spec of the X-Curve, and while for small rooms without long reverb build-up you don't need to do the X-Curve high freq roll-off, I do use the low freq roll-off because I have found that translates well to large rooms. The LFE, however, should not have that low end roll-off, but should be flat from about 80Hz down to somewhere around 30Hz, albeit 10dB hotter on an RTA than the un-rolled-off main speaker baseline. At least this is how I apply EQ for my room.

Specs for very small rooms are subject to some debate. Dolby's small room spec is for stages that are considerably larger than many of us work in, so it doesn't always apply 100%. None of what I say is "scripture", but based on logic and experience. When it comes to small room translation to large theaters, there is plenty of room for personal preference.
thank you Gary,
this cleared up many questions that I have been wondering about.
Old 29th January 2013
  #27
Lives for gear
 
ggegan's Avatar
Quote:
Originally Posted by JAlexander66 View Post
thank you Gary,
this cleared up many questions that I have been wondering about.
My post specifically refers to how I EQ speakers for a very small room (say 1500-2500 cu' with the speakers anywhere from 1.5 to 3 meters from the mix position) for predubbing features or other projects that will be final mixed on a large dub stage or exhibited in a large theater with speakers that are EQed to the X-Curve.
Old 27th January 2015
  #28
Gear Nut
 
Dragomir's Avatar
 

Quote:
Originally Posted by ggegan View Post
My post specifically refers to how I EQ speakers for a very small room (say 1500-2500 cu' with the speakers anywhere from 1.5 to 3 meters from the mix position) for predubbing features or other projects that will be final mixed on a large dub stage or exhibited in a large theater with speakers that are EQed to the X-Curve.
Could you please describe, step by step, your procedure of calibration, eq-ing, delay setting and LFE +10dB boost??!?!! Valuable knowledge much appreciated!
Old 28th January 2015
  #29
Lives for gear
 
ggegan's Avatar
Quote:
Originally Posted by Dragomir View Post
Could you please describe, step by step, your procedure of calibration, eq-ing, delay setting and LFE +10dB boost??!?!! Valuable knowledge much appreciated!
There is a GS sticky thread that explains calibration very well.

B chain delays and EQ is too complicated of a subject to try to cover in a forum, so I'm not going to try to do a step by step how-to.

In very small rooms it is possible to avoid using delays in the front speakers by making sure that they are all the same distance from mix position by placing them in an arc. If the surrounds are placed at the exact same distance as the front speakers acoustic phasing may sometimes occur with certain high noise content sounds when they are sent simultaneously to the front speakers and the surrounds. The phasing can be avoided by delaying the surrounds so that the surround audio arrives 8-10msec after the front audio. This will alter the imaging a bit, but I find that preferable to putting up with the phasing.


With regard to EQ I will just say that in order to do it correctly you need to be able to analyze what is going on in the room acoustically before you start messing with EQ, because there are several issues that EQ can't fix and that need to be dealt with in other ways. You also have to realize that it is impossible to achieve a perfect curve and that there are some compromises that have to be made in order to get the best results. If you don't know what you are doing you can wind up doing some very radical EQ in order to try get a certain frequency range perfectly flat when in practice the anomaly you are looking at on the RTA is almost inaudible, so in trying to make things perfect all you have done is screw things up.

Last edited by ggegan; 29th January 2015 at 01:38 AM..
Old 29th January 2015
  #30
Gear Nut
 
Dragomir's Avatar
 

Thank you, Gary!

Considering the acoustics are fine, the order of operations would be as follows (?):

1. LCR and surrounds calibration using an SPL meter and pink noise (full bandwidth or limited bandwidth?)

2. SW calibration using an RTA and pink noise (LPF @ 125Hz - brickwall filter)

3. Set the LFE to +10 dB (for monitoring purposes only)

4. Set the delays for the LCR

5. Eq each channel individualy (except for the LFE) to the x-curve (what about the surrounds - they suppose to be bandwidth limited...?)
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump