Difference Between Avia and DVE - Subwoofer Calibration (1 Viewer)

[Stephen_F]

I was thinking, maybe the easiest way is to compare dates of the files on the DVE DVD. All the files on my DVD have the date 3rd November 2003 (6:34pm), but it should be taken into account that this is the PAL version. The NTSC version may be dated differently.

 

[Jeremy Anderson]

Okay, I figured out how to do a quickie frequency graph. This is the bass calibration tone from Avia (left front speaker):

This is the bass calibration tone from DVE (LFE):

 

[Jeremy Anderson]

August 14, 2003 for all the files on my disc.

The graphs above were done with 1 second of each tone from 20-120Hz, peaks held, using averaging for illustration's sake. The tones were extracted digitally from the DVE and AVIA discs by stripping the AC3 track from the individual chapter and splitting the AC3 stream into its 6 component monaural WAV files. It looks like Avia's calibration tone centers primarily on the 40-80Hz region, while DVE gives what looks like a gradual rise in amplitude as the frequency decreases (which is consistent with Joe Kane's statement that the tone is digitally pre-emphasized for C-weighted meters so that you don't have to do the meter correction). With the Avia graph, reduce the amplitude by 10dB to make them comparable (since Avia is encoded for 85dB at reference and DVE for 75dB), and there is a radical difference between the two tones. Specifically, Avia's tone at 100Hz is down -60 from max (or 45dB), while DVE at 100Hz is at -38 from max (or 67dB). This is because Avia uses tones encoded in the individual channels, and they're taking the common crossover of 80 into account in their bass tone. Also, Avia's tone at 20Hz is down -36dB digitally (or 69dB), and DVE's tone at 20Hz is also at 69dB; the problem with this is that Avia is SUPPOSED to be 10dB higher than DVE (since it is encoded to read 85dB and DVE at 75dB). Therefore, Avia uses a narrower band of frequencies for its calibration, while DVE can use a wider band since its tone is in the LFE channel only and isn't susceptible to any phase/cancellation issues with the other channels. The problem with this is that Avia is emphasizing the frequencies most often associated with room-induced peaks (40-80Hz) and de-emphasizing deep bass (20-25Hz) and bass above 80Hz, meaning the meter is probably reading far higher than it should on any subwoofer that hasn't been equalized for flat in-room response across the frequency range (which, in this example, means that you would be calibrating the subwoofer based on those peaks and UNDERcalibrating for all other frequencies).

If a subwoofer has in-room peaks in the 80-100Hz range, they will show up during a DVE calibration... but not during an Avia calibration (because 80-100Hz bass is far more attenuated due to the narrower region of frequencies used in the tone, again to avoid crossover issues).

So here's my interpretation of all of this:
If the system is equalized to properly reduce peaks to get a flat meter-corrected response across the frequency range (20-120Hz), DVE should calibrate out exactly as advertised. Avia, on the other hand, uses the frequency range where most people have room peaks (40-80Hz), so anyone who isn't equalizing their subwoofer for flat meter-corrected response is very likely calibrating their sub based on room-induced peaks between 40-80Hz. A good comparison for those having the "problem" with DVE is to try the 40-80Hz LFE tone on DVE and see if it matches up more closely to the reading you got from Avia (since the 40-80Hz tone should come closer in frequency range to Avia's subwoofer calibration tone). If it does match up more closely, then the conclusion would be that you have in-room response issues that are throwing off the band-limited pink noise tone.

In short, unless my copy of DVE is different than the others, the subwoofer tone appears to be DIGITALLY correct on the disc. Again, that's just my interpretation, so maybe we can get someone more experienced to weigh in on the matter.

 

[Stephen_F]

Maybe JK should have printed a warning on DVE, "must use a parametric equaliser with these calibration tones"...

 

[Jeremy Anderson]

Well, Stephen_F, if you want to actually "calibrate" it to reference level, you would be equalizing your subwoofer anyway. At the very least, you would map your sub's in-room frequency response and place it to get the flattest possible results. Otherwise, you're really just GUESSING that it's calibrated correctly. That's why I'm glad to see a few subs coming out with parametric equalization built in.

In all fairness, DVE does have a good selection of tones. And despite Avia being far easier to navigate, I tend to get a more cohesive sounding calibration using DVE's tones to set my levels. I'm glad I have both discs.

 

[Stephen_F]

I know what you mean Jeremy and I agree. Assuming this is the problem, not everyone that bought DVE knows that you need to calibrate the sub's in-room frequency response, before you can use the DVE subwoofer extended frequency tone.

IMO most people have been caught out by DVE's tone and it seems (I have no actual proof) that only those with equalisers have not had the problem. The thing is, that people thought they could use the DVE tone, as they used the tone on the other calibration DVDs, like Avia and VE. There was no info from JK to inform people this would be the case and AFAIK there is no info about this on the DVE DVD, the one place you would think should have this info. There has been a lack of communication.

Basically, it looks like people will either need to go to the next level and get an equaliser for their subwoofer or buy another calibration DVD, other than DVE, and at least do a rough calibration so they can hear/feel the subwoofer.

 

[Jeremy Anderson]

I understand your point but on the same note, someone with an in-room frequency response problem could just as easily miscalibrate with Avia. The problem here is the difference between the two tones making people think that DVE is deficient in some way when the problem is that an unequalized subwoofer will calibrate differently with different tones, depending on where the emphasis is. You can't really fault Joe Kane for that.

 

[Stephen_F]

The DVE extended frequencey subwoofer tone makes the problem worse, as can be seen by people getting up to 13db difference between Avia/VE and DVE. The problem is, using DVE, some people cannot even hear their subwoofer, while if they use Avia they can hear the subwoofer. To me, that is a huge difference. I am not saying it is JKs fault, but I thought JK would have included that sort of info on the DVE calibration DVD, considering he changed tones from the original VE DVD.

BTW it is good to thrash this problem out Maybe I will end up buying a Behringer equaliser to see for myself, if I can use the DVE subwoofer tone.

 

[Edward J M]

I analyzed the subwoofer output signal from Avia and DVE with TrueRTA. I set the software to the dBu scale, and to read millivolts. The sound card line voltages have been calibrated with a pro grade digital multi-meter. Master Volume was set at -15 RL since it was early in the morning.

The first graph shown is Avia. Sharply truncated limited band pink noise from 35-80 Hz, with a total input voltage of approximately 34 mV.

The second graph is DVE. Broad band pink noise from 15-100 Hz, with a total input voltage of approximately 31.5 mV. Ironically, the DVE disc calls this tone "LFE 40-80 Hz". My ass.

While the overall voltage input between the Avia and DVE signals is almost identical, the distribution of signal content is radically different.

Anyone with a subwoofer capable of true anechoic extension to 20 Hz (or lower) will experience significant room gain in the typical 2000-3000 ft3 HT room. In my room with the PB2-Ultra, I start to experience room gain at about 30 Hz and it amounts to at least 10 dB at 20 Hz.

Bottom line: The limited bandwidth Avia signal contains no significant content below 35 Hz, and as such sub calibration will not be impacted by room gain. The broad band DVE signal OTOH has strong content down to at least 15 Hz and therefore sub calibration will be significantly impacted by room gain.

My opinion? JKP and DVE should have realized that room gain is a nearly inescapable phenomenon (especially with subs that troll), and should have sharply truncated the subwoofer calibration signal at 35 Hz, similar to Avia.

Unless the enthusiast has spectrum analysis capability, and deliberately chooses to eliminate all room gain with a PEQ, using DVE will invariably result in significant undercalibration of the subwoofer. For this reason, Avia continues to be my recommended DVD for HT/subwoofer calibration.

Regards,

Ed

AVIA SUB TONE:



DVE SUB TONE:

 

[Brian L]

Edward,

Great analysis.

I hate to say it, but while Joe's explanation of the phenomena (as published in WSR and his web site) does mention room issue and that DVE has wider bandwidth, he does not really explain WHY that would be an issue, and in fact, some of what he says still seems wrong to me.

But your graph and commentary shows exactly how and why. I have said it before and will say it again: Joe has done great things for out hobby, but he needs the services of a good tech writer.

The fact that LFE issue ignores the realities of most people's rooms is similar to the fact that on many displays, DVE's test patterns and title screens are chopped on the top and sides a bit. I believe this to be because most users have a bit more oversan on their displays than strictly nesessary.

OK fair enough, but could they not have developed the disc with the idea that MOST users will NOT want to see their screens chopped, rather than making it the fork in the eye that it really is? VE didn't have that problem, nor does Avia.

BGL

---------------------
For reference, here is how Joe explains the LFE problem on the web site:

2) In setting audio levels we’ve discovered that there is a difference between the LFE channel in VE and DVE. Which is correct?

Measuring low bass information turned out to be more complicated than we though when we created Video Essentials back in 1996. In all of the other channels we can trust the metering used in the editing system, but not with the low bass information found in the LFE channel. In VE we used the editing systems’ meter to determine the level in the LFE channel. In DVE we made careful spectral measurements of that audio, applying the C weighting curve that is used in acoustic meters to measure the audio level in that track. The level on DVE is designed to be correct when you are measuring the level with an acoustic meter. The meter should be in the C weighting position with a slow response. The difference between the two discs is 2 dB, with DVE being higher than VE.

3) Several e-mails have reported differences as large as 5 dB between the two discs. Why would it be different than what has been measured between the two DVD mater tapes?

Our measurements were made at the output of the Dolby Digital decoder with C weighting applied. They do not account for anything that might be going on in the amplifier or in the room itself. As much as you might think the amplifier and the room would be consistent when playing the two discs there is a potential of a difference. The spectrum of the noise is slightly wider in DVE than in VE. This was done to be more inclusive of the capability of the channel. There are three places where this could make a difference. The extra energy might be a factor if the cross-over is set above 80 Hz. If the response time of the meter is fast you might be seeing larger peaks from DVE than VE. Last, but not least, is the room bass response at your listening position. If it has not been equalized the slightly larger bandwidth in the LFE signal in DVE could be reaching your listening position at a higher level. Move the meter to the left or right of the listening position and see if it shifts in level. In one of the rooms we tested there was a 30 dB difference in levels of various bass frequencies at the listening position. The addition and proper positioning of a second subwoofer reduced that difference to 3 dB.

 

[Edward J M]

My sub level in the 30-80 Hz region is essentially level with that of the L/R mains in the 80-200 Hz region. So calibrating with Avia works almost perfect for me.

Below 30 Hz, my sub level increases considerably as compared to the 30-200 Hz region. I'll take free room gain below 30 Hz anytime, but it sure screws up the sub calibration of you're using DVE.

I imagine if I calibrated the subwoofer outdoors in my backyard, that I would get almost identical overall SPL levels between Avia and DVE, since the input voltage is almost the same. Unfortunately, most of us don't watch HT in our backyards or in true anechoic chambers.

For those with huge rooms that exhibit minimal room gain, or for those owning "subwoofers" with no genuine extension below 35 Hz, the disparity between Avia and DVE would probably be smaller than I observed in my room.

Again, I see absolutely no need for such a broad band subwoofer signal for calibration purposes. The DVE signal contains an essentially constant amplitude signal to nearly 15 Hz, and indeed still exhibited strong content at 10Hz!

Because it doesn't include room gain, I would imagine the limited bandwidth (35-80 Hz) Avia signal is a much better sub calibration solution than DVE for most rooms, but there could be exceptions.

 

[Jeremy Anderson]

Nice analysis, Edward! Good to see we got pretty similar results. I'm glad we're finally getting some hard data in this debate.

My SVS 20-39CS+ trolls pretty deep, so I guess the calibration being the same between Avia and DVE just means I must have nailed the equalization in my room.

For me personally, I got DVE primarily for the 6.1 tones. My calibration came out slightly different than calibrating with Avia or the internal tones for my right and center surrounds, and ended up sounding more balanced because of it. I also REALLY like the tones between speakers, which helped a lot with speaker aiming and fine-tuning delay times in my room. And I like the color filters in DVE better. But overall, Avia will still get more use in my house. It's nice to have both though.

 

[TimRP]

Ok Ed, if the subwoofer tones are about equal in voltage output, are the L/C/R/Surround levels about equal as well? Doesn't make sense to compare just the sub output...or am I confused in what you are saying/implying?

 

[Edward J M]

Yes, you are 100% correct. I wasn't thinking clearly about that and forgot the DVE tones are supposedly mastered at -30RL, and the Avia tones are mastered at -20RL.

If the input voltage for the DVE sub level tone is nearly the same as for the Avia tone, that would mean the DVE tone is running way too hot, regardless of the spectral distribution. I'll have to check the surround tones now to confirm your suspicions, but I think you are definitely onto something Tim.

If the surround tone input voltages are considerably lower for DVE than for Avia, and the sub tone voltages are nearly the same, we'll have our answer. Thanks for questioning the obvious that we all missed while studying the spectral distribution.

Ed

 

[TimRP]

Should be interesting to know what the other values are. My guess is they will be off by a good amount.

 

[Edward J M]

Master Volume 0.0 (Reference Level)

Avia Left Front Channel: 2.5 V, 85 dB At Seat
DVE Left Front Channel: 1.0 V, 76.5 dB At Seat

Avia Left Front Sub Tone: 0.037 V, 86-87 dB At Seat
DVE Sub Tone: 0.045 V, 92-93 dB At Seat

The voltages and the SPL readings give a complete picture folks. There is no doubt in my mind that the DVE sub tone is mastered way too hot.

Because DVE is mastered at -30RL and Avia is mastered at -20RL, the input voltages from Avia should always be considerably higher than those from DVE. And that is certainly the case with the left front channel. That is definitely not the case with the subwoofer calibration tone.

Spectral distribution issues aside, the voltage for the DVE sub tone should be considerably lower than the voltage for the Avia sub tone. Instead, the DVE tone is actually slightly HIGHER in voltage than the Avia tone at Reference Level. And the higher SPL from the subwoofer proves it.

Even with room gain considered, there is no way to explain away a jump from 76.5 dB on the left surround channel to 92-93 dB on the subwoofer.

Without dragging a subwoofer outside (which I will do eventually but it's raining), I am very confident at this point that the DVE subwoofer tone (on my copy of DVE) is definitely mastered about 13 dB too hot and this is definitely not exclusively due to room gain or other spectral distribution issues, and the subwoofer input voltage is proof.

Regards,

Ed

 

[Jeremy Anderson]

I increased resolution and set the range to 10Hz-120Hz. Here's what I get when I run total power for the sub tones (1 second sample of each):
Avia: -16.61dB total power, peak frequency 46.88Hz, peak amplitude -21.95dB
DVE: -29.88dB total power, peak frequency 35.16Hz, peak amplitude -38.53dB

Again, this is measured directly from digitally extracted .WAV files from the AC3 stream, not passed through a sound card. Here's a graphic to further illustrate how different the tones are (red=DVE, yellow=Avia):

The LFE tone on DVE (title 9, chapter 13) is very obviously mislabeled as 40-80Hz on the accompanying graphic, though the title/chapter list in the case shows this only as "Band Limited Pink Noise - LFE". Title 9, chapter 26 is labeled as 40-80Hz in both places, so I extracted it... and it is identical to the LFE tone. In other words, neither of them are actually 40-80Hz tones.

To answer the above question, I extracted the left front speaker level tones from both DVE and Avia for comparison (20-20kHz range). Here are the results:
Avia: -12.10dB total power, peak frequency 925.78Hz, peak amplitude -30.30dB
DVE: -26.43dB total power, peak frequency 884.77Hz, peak amplitude -45.12dB
Here's a graphical comparison of the two tones (red=DVE, yellow=Avia):

(Edited to fix my screw up on the scale shown.)

 

[ScottCHI]

the yellow rollercoaster looks like it's more fun

 

[Jeremy Anderson]

Here's something interesting:
If you apply C-weighting to the DVE LFE tone and take overall output level from 20-120Hz, you get -30.35dB C, or 74.65dB output at reference. So technically, Joe Kane is correct. On a system with flat in-room response, the DVE tone should calibrate to 75dB on a C-weighted meter (such as the Radio Shack SPL meter).

However, if you apply C-weighting to the Avia tone, you get -17.53dB C, or 87.47dB output at reference. That's right, folks... 2.47dB over reference on an ideal system. Why? Because Avia's tone isn't pre-compensated for C-weighted meters. Therefore, when using Avia, one should be calibrating the subwoofer to 82.53dB to reach reference level. This shouldn't surprise anyone, since the 2-3dB drop to account for meter correction has been common knowledge.

Again, it should be noted that these are numbers directly from the digital domain (the raw AC3 stream), and should represent the best-case scenario for a perfectly calibrated system. So while DVE has its tones incorrectly labeled as 40-80Hz tones, the tones for calibration are still correct in level. Anyone who finds DVE reading hot likely has frequency response issues in their room, which is a setup/hardware problem. In those cases, equalization should be considered to get a more accurate response from the subwoofer. Let's face it... If you've spent the money on a good sub (a la SVS), you should really spend the extra cash to get an EQ. Anything else is like putting regular gas in a high-end sportscar. Sure, it'll run... but not to its fullest potential.

I have my SVS 20-39CS+ EQ'd for flat response +/-3dB from 20-100Hz and I get 82-83dB from Avia and 75dB from DVE at reference level in my room. And I'm doing that with a ART-351 EQ, not a parametric EQ. So get to tweaking, people!

 

[Jeremy Anderson]

Another point of interest:
I ran overall C-weighted level from 20Hz-20kHz on both the left front speaker tone on Avia and the left front speaker tone on DVE. Here are the results:

Avia -12.15dB C (or 92.85dB C).
DVE -26.48dB C (or 78.52dB C).

From these numbers, it looks like Avia's 85dB tone for calibration is actually recorded at 92.85dB, or 7.85dB hot. DVE, on the other hand, is recorded at 3.52dB hot from its quoted 75dB level. That's a difference of 4.33dB in overall level on an ideal system. Based on this (which is the actual digital output), calibrating your subwoofer with Avia (even with the -2.5dB meter correction) would put the subwoofer about 7-8dB hot.

I hope that someone more knowledgeable about acoustics can chime in. From the numbers I'm seeing, DVE is actually the more accurate of the two as far as calibrating for actual reference levels. I wish I had the S&V disc for further comparison.