LFE vs Diverted Bass - Is your processor WRONG? (Updated with custom test tones) (1 Viewer)

Arnel · Apr 15, 2002

We all know that the 0.1 channel of a soundtrack is a separate/discrete channel for LFE and being reproduced by our sub. The sub would reproduce only LFE if the rest of the speakers were set to Large. But if any speaker were set to small (say at the 90hz crossover), the sub will reproduce the LFE (0.1) plus the diverted bass (90hz and below) from the speakers set to small.

Now AVIA DVD has a test tone (under evaluation menu), where you can check the level of the diverted bass and the LFE. On my first test, my RS meter drop to about 10db when the tone reaches LFE. The note says that I have to set my LFE to normal or 0db on my system. So my former setting of -10db (this is the mid setting of my Yamaha RX-V800, where I can easily adjust sub level), was now change to 0db (max setting on Dolby, while DTS was up to +10db). The AVIA test again indicates a drop of about 4db on the LFE. I know that the result indicates that I have a poor bass management, since diverted basses were overpowering my LFE. I don’t know were else can I modify to make my bass properly leveled. Please help.

Harold_C · Apr 16, 2002

You should have two different adjustments on the Yamaha, one for the subwoofer level and one for the LFE levels (separate settings for Dolby Digital and DTS).

The Yamaha manual is not terribly clear, so I can't give you specific instructions. However, you want to use the procedure described on page 22 of your manual for setting the level of the subwoofer (and perhaps the Lo Freq test procedure described on page 36). Use these adjustments to set your subwoofer to the same 85 dB output as any other channel on the Avia test disc.

The LFE level procedure described on page 39 is different. This is an LFE adjustment. It must be set to 0 db when you are calibrating your subwoofer levels. It should be left at 0dB for both Dolby Digital and DTS, unless you find that the loudest bass peaks on soundtracks are causing audible distress to your subwoofer or to your ears! Other than using it as a sort of "limiter" to protect your subwoofer, there would be no reason that you should ever set this to less than 0dB for Dolby or DTS -- with the exception of certain DTS music recordings that were mixed incorrectly and require a -10db LFE setting to compensate for DTS's screw up in mixing these recordings. (It's possible that Yamaha is mislabeling the DTS 0 db setting as +10 db -- a quick listen to a DTS soundtrack would confirm.)

As far as the 4 dB discrepancy -- I will have to play with that portion of the AVIA disc and get back to you. The wife is asleep, so I don't think I can do it now! In any case, if you have any of your speakers set to LARGE, this could be the problem as redirected bass would to to the LARGE speakers AND to the subwoofer. In 99 out of 100 systems, you shouldn't have any of your speakers set to LARGE. You really don't want the surrounds and center set to SMALL and the fronts set to LARGE because this will produce better bass from the surrounds and center than the fronts. Not good. Try setting every thing to SMALL and try the test again. Also, make sure that the LFE/BASS output is set to SUBW -- not the MAIN or the default BOTH settings.

Arnel · Apr 16, 2002

Thank you Harold for your quick response. Below are my current settings:

1. All speakers were set to small and my LFE/Bass Out was set only to SWFR – this means that my LFE plus the 90 Hz and below frequencies from L, C, R, LS & RS will be reproduced only by my Subwoofer.

2. LFE Level for DD and DTS is set to 0 db – this is the default. DD range is from -20 to 0 db, while DTS is from -10 to +10 db.

3. SWFR Level is set to 0 db at the Speaker Level Settings.

Under Audio Test Patterns of AVIA DVD, you will find the “Low Frequency Pink Noise, 6 Channel Pan” at Verification/Evaluation Chapter. This is a circulating low frequency test pattern from R, C, L, RS, LS to SWFR/LFE. Since all my speakers were set to small, naturally all sounds were emanating from my SWFR. My RS SPL meter reading is the same all through out the 5 speakers but dropped by about 4db when it reaches the LFE. I don’t know how to compensate for it.

greg_t · Apr 16, 2002

I think you should have your lfe level set at it's highest level for both DD and DTS. So for DD, set at 0, and for dts, +10. The LFE channel is meant to be 10db higher than the other channels. A setting of 0 on DD adds the proper 10db boost, and a setting of +10db will do it for DTS. You then need to recalibrate your overall subwoofer level. The subwoofer level can be the same as the main channels, but many here prefer to run 2-5db hot. For a much better explanation on all this, do a search by name for "Vince Maskeeper". He has several excellent posts on how this all works.

Harold_C · Apr 16, 2002

Arnel:

Thanks. I just wanted to clarify your settings -- Dolby Digital does some various combinations of attenuation on various channels with bass redirection. I wanted to look these up and see if that could be it and all SMALL is simplest configuration. Everything looks exactly right in your settings. I'm puzzled by the 4 dB shift and agree with you that it should be the same levels.

I looked at the documentation for bass management and specifically the bass mixer. I see nothing that would account for the discrepancy.

The only other setting that I can think of that might affect the results might be an LFE peak limiter (which I don't think the Yamaha has) or a dynamic range setting or, conceivably, the dialog normalization. However, I can't really think of why these wouldn't change all channels simulataneously.

The other thing to confirm is to make sure that you don't have a bass control or "loudness" curve activated by accident. I've often run through all the settings on my system and found inadvertant settings memorized -- the result of mashing buttons on the remote in a dark room in the wrong remote mode. However, these shouldn't affect Dolby Digital.

I have found the tones on my Avia disk and want to see if I'm experiencing the same thing you are -- we have the same settings, but different brands of processors. I'll get back with the results.

BTW, there are register settings in the bass management software that would cause the results you've seen. At the input to the bass mixer (with all SMALL speakers), the five main channels are attenuated 10 dB and the LFE is attenuated 5 dB -- this is actually the mechanism the decoder uses to get the extra 10 dB of LFE level during playback. It is possible that Yamaha has used different settings that the Dolby recommendation for these individual channel levels at the bass mixer input. I don't think this is likely, since "ALL SMALL" is one of Dolby's mandatory published configurations, required for certification and detailed in the Dolby chipset documentation. But, anything is possible.

It is also possible that you are seeing some kind of crossover "artifact". For example, if the pink noise test tones are bandwidth limited to 20 - 200 Hz, you would be getting some output from the main speakers as well as the sub. Given that the Radio Shack meter is not "flat" this could account for some higher readings. This would not impact the LFE channel because it goes through a brick wall digital filter at 120 Hz during encoding and is almost always passed through an 80 Hz lo-pass filter as well. If I get the same imbalance as you, I want to try the test with a couple of main channel speakers actually unplugged to eliminate that possibility.

Anyway...all I have for now is speculation and a curiousity about the mystery.

Harold_C · Apr 16, 2002

Arnel:

I'm getting the same sort of thing on my system with those test tones. I dont' know that I could peg it as a 4 dB drop -- I can't get sufficient precision from the Radio Shack to call things that closely, too much needle movement.

I'm goint to defer to the Avia folk on this one.

I do know that, even witha 100 Hz crossover, there is still output from the main speakers on this test. So the test signal extends at least up to 120 Hz or so, perhaps higher. This energy won't be there on the LFE track.

I don't have any real confidence in the Radio Shack SPL meter below 100 Hz. In addition to the needle moving too much to get an accurate reading, there are serious issues decay time issue trying to measure wavelengths that long with simple detector circuits and the weighting built into the mic and the meter itself confuses things.

For exmaple, I noticed that just before the LFE track ended, the needle was rising rapidly. It could very well be that if the LFE test tone is filtered at 80 Hz, the length of time the test tone plays may not be long enough for the SPL meter to get an accurate read. I notice this phenomenom with guitar tuners. The tuner has a delayed response to low E bass string because it's got to process "x" number of cycles before it can lock on.

If this were the case, the SPL meter could be basing it's reading of the main channels based mostly on higher frequency tones (say 100 to 200 Hz). Then, when faced with the LFE tones (sharply filtered in the encoding process above 80 Hz), the meter itself is simply not able to provide a similarly accurate reading.

I think the note on the Avia instructions that "some variations are acceptable" indicates to me that they know this particular test is fraught with measurement error.

I go back to Dolby own tech papers. While they highly recommend the Radio Shack SPL meter for setting levels and channel balancing of the main 5 channels, they flat out state that this type of meter is essentially useless for pink noise measurements below 100 Hz. I think that's right and that's why I would very wary of trying to EQ a subwoofer based on measurements from the Radio Shack SPL.

You can tell that something is funky from the "correction factor" charts published here and there for this meter. In the bass, the correction factors are all over the board, with wildly different "correction" factors for 30 Hz, 40 Hz, 50 Hz, etc. If the measurements are valid, the correction factors should be a fairly smooth curve -- not jumping all over the board.

You can also see the problem inherent with this test by moving the meter 2 feet and repeating it. I bet that you get significant different results.

Arnel · Apr 16, 2002

Harold, Thank you very much for the time you've spent trying to understand my system's problem.

It is also possible that you are seeing some kind of crossover "artifact". For example, if the pink noise test tones are bandwidth limited to 20 - 200 Hz, you would be getting some output from the main speakers as well as the sub. Given that the Radio Shack meter is not "flat" this could account for some higher readings. This would not impact the LFE channel because it goes through a brick wall digital filter at 120 Hz during encoding and is almost always passed through an 80 Hz lo-pass filter as well. If I get the same imbalance as you, I want to try the test with a couple of main channel speakers actually unplugged to eliminate that possibility.

This is a very good assumption. Have you tried unplugging main channel speakers during the test?

I hope we could really pin point were the problem lies. Maybe others can also help us.

Click to expand...

Harold_C · Apr 16, 2002

This is a very good assumption. Have you tried unplugging main channel speakers during the test?

Yes. I just ran it with my main channel amp turned off entirely -- just with the sub amp turned on. It didn't make any difference. However, as I say, my meter is swinging over a 4 to 6 db wide range on all speakers during this test, so I'm not really confident in what the absolute measured values really are.

What I suspect is causing the problem is that the pink noise for this test tone covers the range from 20 to 200 Hz. However, when this signal is encoded on the LFE track, the encoder is running it through a steep 80 Hz filter with an abosolute brick wall filter at 120 Hz. Thus, the actual spectral content for the test tones recorded on the disc in the main channels and the LFE channel is different. Given the Radio Shack meter's unpredictable response to say 60 Hz and 120 Hz, my bet is that this is where the problem lies.

I think that to make this test really valid, you would have to use a real time analyzer and look at a very narrow band well below the crossover point.

Interestingly enough, when I "close-mike" the sub with the meter and run the low frequency sweeps in a main channel and then in the LFE channel, I get exactly the same ouput levels below 100 Hz. This is probably a more valid way of using the meter since it is measuring a limited bandwidth at any given time. The absolute levels probably aren't accurate, but at least the meter inaccuracies will be the same for both. The main channel sweeps start at 200 Hz and go down to 20 Hz. The LFE sweeps start at 100 and go down.

As it stands, I think the test is useful just as a rough confirmation that the LFE channel and the bass redirection is working. Nothing I saw from my system indicates that it is at all useful in making assessments of absolute levels using a Radio Shack meter.

Click to expand...

Arnel · Apr 16, 2002

I'm goint to defer to the Avia folk on this one.

Yes, I think they can really help us. Isn't it Guy Kuo a member of this forum?

Click to expand...

Harold_C · Apr 16, 2002

Having seen that the individual channel bass sweeps perform as they are supposed to (same levels for LFE and individual channels, assumng Avia recorded the LFE channel 10 dB down), my assumption is that the pink noise test tones are just plain wrong on the disc.

There's no reason that one set of tones would work and the other wouldn't -- allowing for some discrepancy inherent in the meter.

Harold_C · Apr 17, 2002

Hey Arnel:

I figured out, in part, why that Avia test was behaving so strangely.

That is a PAN test. It's hard to tell from that one, but if you play with some of the broadband "pan" test tones, you'll see that these don't just jump from one speaker to the next as the speaker calibration tests do. Instead the "panning" tests smoothly and slowly pan the signal from one channel to the next.

The end result of this is that you get just an instant of single channel sound at the start and finish of the pan with all the time in between taken up with various levels of mixing between the two channels. Obviously, this introduces a whole 'nother dimension of room effects and speaker interaction into the equation plus the fact that the absolute levels vary during the pan (depending on the taper of the panning controls).

I believe the fact that you are never looking at a pure "single channel" signal during that test could easily account for the 4 dB discrepancy you were seeing.

The Avia records the LFE test tones 10 dB lower than the individual channel test tones. For the sake of discussion, lets say that the center channel test tone is recorded at -20dB and the same test tone recorded in the LFE channel is recorded at -30dB.

Now, those signals are attenuated at the input of the bass mixer bass management module. The center channel signal is attenuated by 15 dB, so our orignal -20 dB signal is now attenuated to -35 dB. The LFE signal is attenuated by 5 dB, so our original -30 dB signal also ends up at -35dB. All fine and dandy if we are comparing two separate channels back and forth like a light switch. But, who the heck can predict what is happening when you PAN between those two signals (each with different levels of attenuation and each affected by the taper of the panning volume control). On top of that, I'm pretty sure that it is panning from a 20-200Hz signal to a 20-80Hz signal.

Just to see how the panning changes things, watch your SPL meter on the regular 5 channel pink noise panning tests. It will be moving all over the place as the signal is shared by two adjacent speakers.

Bill Kane · Apr 17, 2002

Great dialog you guys.
Arnel, I run RXV-800 and take an interest here. One thing bothered me (and I dont know if it's still in the equation).
It's the balance between AVR sub level and sub amp "volume" setting.

Yamaha SWFR LEVEL ranges from -20 to 0. Many recommend Yamaha Sub Level be in the mid-range of this scale, for playback adjustment capability. Mine ended up at -15.5dB running a SVS 25-31PC w/ driver upgrade.
Adminr Vince Maskeeper belongs to the school that believes the sub amp voltage control ought to be close to MAX. As I ended up, mine is at the halfway mark. Again, this gives me adjustibility should I want it.
My DD LFE is O. DTS LE is +10. Guy Kuo has suggested that DTS +10 is used for DVD tracks while DTS 0 is preferable for DTS music. (I dont have DTS music to test this, tho).
Does this offer any relevance to your Sub Saga?

Click to expand...

Harold_C · Apr 17, 2002

Guy Kuo has suggested that DTS +10 is used for DVD tracks while DTS 0 is preferable for DTS music. (I dont have DTS music to test this, tho).

I don't think this is the result of any inherent need for different bass levels in music and cinema. Rather, when DTS introduced DTS music recordings, they apparently made a rather significant error in their published mixing/encoding guidelines. The result is that all of the early DTS music discs have the LFE channel recorded 10dB too high. To get out of the predicament, they now couch the required -10 dB correction as a DTS Music setting or quasi-feature.

Click to expand...

Arnel · Apr 18, 2002

Having seen that the individual channel bass sweeps perform as they are supposed to (same levels for LFE and individual channels, assumng Avia recorded the LFE channel 10 dB down), my assumption is that the pink noise test tones are just plain wrong on the disc.

Harold, I don't believe this is the case, cause if a calibration disk (which I think has better quality control) was encoded wrong, what else can we expect from a normal DVD (Well, it did happen to Jurasic Park original release). And another thing, the note on AVIA have even specify that the level should be the same through out the 6 channels for us to check our system's bass management.

I was wondering why Dolby did not include level manipulation for 0.1 channel. The Center, RS and LS has level and delay manipulation but none on LFE. They only have SWFR Level which is totally different from LFE level. And maybe DTS is trying to correct this by jacking up their LFE.

I can't believe others were ignoring the issue and don't want to reply on this thread. They may have the answers, I just hope they help us.

Click to expand...

Bill Kane · Apr 18, 2002

I had never used the AVIA evaluation "LF Pink Noise, 6 CHL Pan" and just tested.
First, I recalibrated all speakers again at 75dB incl sub. DD LFE at 0(default). DTS LFE is irrelevant in these settings since AVIA is DD only. (I left it at +10 and will come back later with my Sound&Vision test disc that has a DTS track.)
In the pan test, I too saw a ~5dB meter drop at SWFR. Absent any equalizer in my system, I don't see there's a huge "problem" to concern me. Arnel may feel differently.
Harold wrote:

As it stands, I think the test is useful just as a rough confirmation that the LFE channel and the bass redirection is working. Nothing I saw from my system indicates that it is at all useful in making assessments of absolute levels using a Radio Shack meter.

Comparing one another's settings usually is futile because of room characteristics, sub placement, spkr sensitivity and the lot. But since Arnel and I have the same AVR, here's mine:
AVIA @ 75dB
LF -26dB
SWFR -15.5dB
AVIA @ 85dB
LF -16.5dB
SWFR -15.5dB
SVS sub amp voltage knob at halfway mark.
At least I got my system freshly recalibrated

Click to expand...

Harold_C · Apr 18, 2002

I was wondering why Dolby did not include level manipulation for 0.1 channel. The Center, RS and LS has level and delay manipulation but none on LFE. They only have SWFR Level which is totally different from LFE level. And maybe DTS is trying to correct this by jacking up their LFE.

It's not really necessary. If you get the subwoofer level right, the LFE level takes care of itself. Actually, the relative levels of the redirected bass and the LFE channel in the bass mixer are set differently depending on the speaker configuration. Dolby publishes page after page of configurations for bass levels depending on how many and what type of speakers are configured in the system. It's all automatic from the end-user's standpoint.

The feature that I think it is a very good one is THX's LFE limiter feature. It leaves the overall LFE level unchanged, but you can set an absolute limit at 115 dB (max) or any level below that. This is a great feature if the LFE peaks are making your subwoofer cry out in pain.

Click to expand...

Vince Maskeeper · Apr 19, 2002

I would certainly like to comment on the meat of this discussion- however while I do own Avia and find it useful for some things- I usually use VE for audio calibrations. That said, you each keep referring to comparing a BAss Mgmnt tone to a LFE tone- and as of yet, I had been unaware of a dedicated LFE tone on the Avia disc that was designed with calibration in mind (accurate).
If someone could point out to me exactly what tone you're using, exactly where you found it on the disc- I'd be happy to check into the reason for the problem.

I don't think this is the result of any inherent need for different bass levels in music and cinema. Rather, when DTS introduced DTS music recordings, they apparently made a rather significant error in their published mixing/encoding guidelines. The result is that all of the early DTS music discs have the LFE channel recorded 10dB too high. To get out of the predicament, they now couch the required -10 dB correction as a DTS Music setting or quasi-feature.

Partially Correct, though not entirely accurate.
The LFE channel of multichannel film soundtrack requires a 10db boost on the LFE channel. The LFE is specifically encoded 10db lower than intended- and the decoder adds back in the 10db (assuming your LFE pad is in the proper position).
This standard is not specific to DTS or Dolby, rather is an element of the multichannel source stems as dictated by the SMPTE standard. All multichannel film soundtracks are prepared this way, regardless of digital carrier, and a decoder replaces the 10db.
Music, however, doesn't have this standard. When the basics for multichannel music was created, the extra 10db lfe headroom allowed by the 10db reduction was not included to my knowlegde (although, I have not really seen any decided official specs for 6 channel music mixing). So, I don;t think multichannel music needs the boost. This is true, I believe, for all multichannel music formats in their raw form before encoding.
So, DTS music CDs require the DTS LFE pad to be at the 0 position (on most receivers, since there is no real standard to labels, your results might vary)-- and DTS movie soundtracks require the DTS pad to be at +10. For both the DD and DTS pads, they usually should be in the highest position, regardless of the numbers they use to label them (again, there seems to be no real standard- so some people get worried that DTS is at +10 and DD is at 0-- but 9 times out of 10 this is correct and means proper boost for each... despite different label).
So, it's not correct to say DTS made a mistake on multichannel audio for music-- rather I believe this standard is true for the source of all multichannel music masters- DTS or no.
-V
PS: There is quite a bit of incorrect, partially correct, and misunderstood concepts in this thread- I would urge readers to be careful and do some additional reading if they are actually interesting in how this works.

Click to expand...

Harold_C · Apr 19, 2002

So, it's not correct to say DTS made a mistake on multichannel audio for music-- rather I believe this standard is true for the source of all multichannel music masters- DTS or no.

It's not true for Dolby Digital 5.1 music recordings. In their guidelines for mixing 5.1, they have specific sections on mixing musical recordings, including levels and the use of LFE. Throughout these guidelines, they make no mention of using different LFE levels for music versus film versus TV mixing. In fact, the only place where music really differs in their guidelines is that they do not "suggest" a specific overall recording level as they do for film and TV.

Likewise, (although I can't cite chapter and verse), I do not believe that ALL DTS music releases have the incorrect LFE setting. I think they've corrected the mixing guidelines for later releases. It would only be logical to fix the problem, because having a 10 dB difference in LFE level from one disc to the next is not something you would wish on your worst enemy.

Click to expand...

Vince Maskeeper · Apr 19, 2002

Throughout these guidelines, they make no mention of using different LFE levels for music versus film versus TV mixing. In fact, the only place where music really differs in their guidelines is that they do not "suggest" a specific overall recording level as they do for film and TV.

While they don't suggest a different LFE level literally- do they have any info on setting up monitoring for mixing multichannel music specifically? Again- I didn't even bother to look at Dolby's info, simply because I was not aware that they were one of the formats actually used for serious multichannel music.

-vince

Click to expand...

Richard Greene · Apr 19, 2002

Hey, stop beating up that Radio Shack sound meter!

Quote from my Radio Shack 33-2050 meter owners manual

(copyright 1989):

"With C-weighting, the Sound Level Meter

frequency response is flat from 32 to 10,000Hz. (+/-3dB). Above 10kHz., the frequency response of the Sound Level Meter drops off rather rapidly ..."

------

The meters do read too low below 32Hz. and especially above 10kHz., so ignore those frequencies ... but I believe they meet their specifications from 32Hz. to 10kHz. (and are probably within +/-2db at bass frequencies from 32Hz. to 100Hz. -- the treble has larger deviations from the official C-Weighting curve).

That's a good deal for $40.

Home Theater

Entertainment & Streaming Content

Physical Media

Home Theater Equipment and Hardware

Other Diversions

Bargains and Classifieds

Home Theater Forum

Archives

LFE vs Diverted Bass - Is your processor WRONG? (Updated with custom test tones) (1 Viewer)

Agent

Stunt Coordinator

Agent

Screenwriter

Stunt Coordinator

Stunt Coordinator

Agent

Stunt Coordinator

Agent

Stunt Coordinator

Stunt Coordinator

Screenwriter

Stunt Coordinator

Agent

Screenwriter

Stunt Coordinator

Stunt Coordinator

Stunt Coordinator

Similar threads

Users who are viewing this thread

Sign up for our newsletter

and receive essential news, curated deals, and much more

Forum statistics