Subwoofer calibration issue (1 Viewer)

[Harold_C]

I was reading a technical white paper from Dolby outlining the proper way to monitor and mix 5.1 channel recordings in the studio.

There is a fairly detailed section on calibrating levels in the studio...mostly stuff that is pretty familiar ground using pink noise and the Radio Shack SPL meter.

But, when they got to the section on calibrating the sub, they said that the Radio Shack SPL meter is not really accurate using bandwidth limited pink noise for setting subwoofer levels. They recommended using 1/3 octave Real Time Analyzers instead. However, they stated that the Radio Shack SPL meter could give a reasonable approximation for setting sub levels if you adjust the sub level some 4 to 5 dB lower than the meter tells you to set it.

In looking at some other materials, it appears this is indeed the case -- that the Radio Shack SPL reads 3 to 7 dB too low below 125 Hz. So, an indicated 75 dB on the meter is actually a subwoofer level up around 80 dB -- some five dB higher than it is supposed to be.

I spent some time testing this "theory" in my own system tonight and I think it is definitely true. From my normal Avia calibrated sub levels, I backed off 4dB on the sub setting.

On the wonderfully recorded DTS 5.1 Paul Simon in Paris concert DVD, the sound improved rather dramatically with the lower sub setting. Voices were much better, imaging was much better. I think that the sub had, indeed, been 3 or 4 dB too hot and that it was masking the midbass harmonics.

The bass in the recording -- kick drums and superb electric bass -- was much improved. Interestingly enough, by turning the sub down, there was MORE apparent low bass extension ... again, I suspect because the excessive bass levels were screwing up the harmonics.

Has anyone else played around with setting a good subwoofer 3 or 4 dB below what the Avia or Dolby pink noise test tones would indicate on the Radio Shack SPL?

As a side note, if you want to hear mindblowing 5.1 channel sound from a concert, get this "Paul Simon in Concert" DVD recorded live in Paris in 2000. It has both DTS and Dolby Digital 5.1 channel soundtracks mixed by an outfit in LA called "5.1 Sound". These people clearly know what the heck they are doing because the mix is just astonishing. Perfect use of the center channel and surrounds -- the speakers completely disappear into one huge cohesive soundfield.

I cannot believe that it is live recording (and from a big arena venue, no less) -- it sounds like a polished studio recording. It is a huge (and extremely tight) band -- three guitars, three drummers, horn section, two keyboards. There is a whole bunch of stuff happening at all times and you can pick out every note of every instrument. Amazing. The clarity and dynamic range of the recording is beyond anything you normally hear on a CD.

The use of the 5.1 format on this one really shows how it is supposed to be done. No gimmicks, just unbelievable sound. This is the way music SHOULD be mixed. I've been listening to live records and videos since Live at Leeds and Get Yer Ya Ya's out 35 years ago. Somebody finally got one absolutely perfect.

 

[Bhagi Katbamna]

Many people run the sub 5-6dB too hot. I always have like it 1-3dB less than the level on the main channels(as per the RS meter). The one instrument that really makes the bass response even for me is the parametric eq without which I have a 17-18dB peak at 40 Hz in my room.

 

[Fred Pfeifer]

You may choose to use a handy compensation chart when calibrating your sub using the Radio Shack Analog Sound Pressure Level meter. Go to this link to view the chart: http://www.svsubwoofers.com/faq_rscomp.htm
Using the comensation chart is also a good way to determine how flat your sub's response is in your listening room if you have a way to output test frequencies that are at the chart's frequencies. You could use the Autosound 2000 Bass CD #101 to do this: http://www.autosound2000.com/detail....ODUCT_ID=CD101
If there are large peaks and/or losses that can't be flattened out with changing the sub's room placement, you may wish to incorporate an equalizer into your system to flatten out the response, such as the Art 351 equalizer: http://www.svsubwoofers.com/ampeq_art_351.htm
Some people like to verify their speaker level settings using this technology, but it's not required. The bottom line is to calibrate your sub to whatever sounds best to you. It appears that you have already done that.
Good listening.

 

[Harold_C]

I don't really believe in trying to equalize low bass response using narrow band parametric EQ's. Trying to accurately measure bass response is fraught with peril, because the available measuring equipment is not well suited to the task. Localized room effects can produce wildly different response plots just by moving the microphone a foot or two. This leads to a very high degree of likelihood that you will end up measuring things that aren't really there. It's like trying to equalize the inside of a large subwoofer enclosure. Trying to chase down resonant peaks with a parametric EQ is going to require massive cuts of 12 to 18 dB. I have a hard time believing that these sorts of radical narrow band EQ settings are good for fidelity.
There are some interesting articles on this by theater sound system designer John Allen. In 20 years, he has NEVER EQ'd the bass of one his systems and most theater systems where people have tried end up being worse. Rather, he simply sets the subwoofer levels properly, does some listening tests to make sure the levels actually sound right on reference material, and calls it a day.

You may choose to use a handy compensation chart when calibrating your sub using the Radio Shack Analog Sound Pressure Level meter. Go to this link to view the chart: http://www.svsubwoofers.com/faq_rscomp.htm

Thanks. I've seen the chart. Unfortunately, it doesn't give an overall correction factor when measuring broader band pink noise (i.e. from 20 to 125 Hz). That's what Dolby appears to be doing when they suggest setting the subwoofer levels 4 dB below what the Radio Shack SPL meter would indicate.

 

[Paul Clarke]

Harold,

Yes, I too have found the same thing using the Rat Shack. My sub runs 4-5dB low and gives great sound. Because my receivers save all settings according to input and sound mode, there are some combinations which stray from this general setup. But when not using DD or DTS (which I only use for HT) I prefer this bass level with music anyway with my room acoustics.

 

[brucek]

Harold,
You're quite correct in saying that using the Radio Shack meter to measure band limited pink noise for subwoofer calibration will produce errors. Most people will just average the wildly varying meter reading. This seems to work OK if your not thinking of EQ'ing your sub.
It's generally accepted, (that unlike a large commercial theater), in a home theater where the room dimensions conveniently create axial resonances that are centered in the all important 20Hz to 100Hz spectrum, it's useful to use a parametric equalizer combined with a knowledge of modal distribution. For example, a room with dimensions of 22ft x 12ft x 8ft will have first order effects at 22Hz, 47Hz, 70Hz.........
To properly excite these room modes, sine waves at sixth octaves are a good tool to use and the Radio Shack meter with the proper calibration constants factored in works fine for this measurement. It's fairly accurate when the constants are used.
I would not use pink noise for calibrating frequencies less than 100Hz. It doesn't properly excite these room modes. I also don't feel a 1/3 octave RTA is resolute enough below 100Hz. Sampling at sixth octaves or greater would make more sense.
You're right though, you don't have to move too far from your equalized area and it's a new ball game. But since most home theaters have limited seating, it's acceptable to optimize a single seat or a front row. If someone wants to sit in the corner - tough.
Primary axial modes can produce some fairly nasty resonant peaks in your low end response. This forces you to calibrate your sub to these peak levels causing the very undesireable result of "single note" bass. Sounds like one of those cars with the big systems in them. Yuk. Personally, I wouldn't use a subwoofer in my theater if I couldn't use a parametric equalizer.
With a parametric equalizer, a cheap RS SPL meter and a simple CDr of sixth octave tones, you can flatten the subwoofer response in most rooms for a fairly wide seating area. This allows you to properly set the wholesale level of your sub. It's a transformation of your music and HT sound that is a revelation when you hear it for the first time.
brucek

 

[BruceD]

Harold,
I'm with brucek on this one. I do go one step further and use a PC based acoustic measurement system called ETF5 (MLS test signals for impulse response and FFT analysis for frequency response) with a calibrated microphone. website here.
I can measure at 1/6 octave or even 1/12 octave precision with this system, and for me the parametric EQ (20Hz-80Hz) makes a world of difference in the quality of the bass in my room. This is very easy to hear and quantify after using a parEQ to cut all of the nasty modal room peaks (standing waves).

 

[Harold_C]

You're right though, you don't have to move too far from your equalized area and it's a new ball game. But since most home theaters have limited seating, it's acceptable to optimize a single seat or a front row. If someone wants to sit in the corner - tough.

That's where I thinkt the whole thing falls apart. In the sample room shown in the demo at the above website, the subwoofer response at 60 Hz changes by as much 14 dB just by moving the microphone 2 feet. This means that the best you can hope for is to EQ for ONE very rigidly fixed location in the room. This limits the usefulness to a small subset of homeowners who are willing to stick to such rigid seating charts.

In any "normal" family room environment where people are sprawled about watching movies, the EQ results go out the window. For example, sitting bolt up right on the sofa will produce a wildly difference response curve than stretched on on the sofa with your shoes off and a comfy blanket.

I do believe that if someone is suffiently motivated, FFT impulse analysis can be very instructive about a room. But, I stick to my contention that wandering around measuring 1/6th octave test tones with a Radio Shack meter is as likely to provide false measurements as real ones. At the most, I would probably use the parametric to address perhaps one large resonant peaks. I don't think try to deal with five or six little 3 dB peaks is likely to accomplish much, since moving your head 12 inches will throw that EQ setting out the window.

I'd love to see someone run a series of plots from six different seating locations -- say two on the sofa, a couple of chairs, lying on floor propped up on pillows, and over on a barstool at the counter and then try to figure out WHICH one to EQ for! That's the basis of my contention that you are just as likely to mess up the sound as make it better.

The other interesting thing that is apparent from the demo plots on that website is that room reinforcement effects can be your best friend. The least "flat" curves come from subwoofers crammed into the corner. From an audiophile perspective, this is "bad". However, these room boundary effects appear to provide as much as 30 dB of rising bottom end gain. That has huge implications on required amplifier power. Setting the sub for roughly equal broadband average output to the main speakers will require significantly less gain on the subwoofer amp and therefore much less power output. Since the vast majority of consumer grade subwoofers are underpowered for Dolby Digital soundtracks, this room gain can be the difference between clean output and having to drive the snot out of the subwoofer amp and driver to come close to Dolby reference levels. Whether you achieve these lower gain settings by turning down the subwoofer or by using a series of a half a dozen notch filters may not make much difference if the end result is getting broadband bass response (say from 30 to 60 Hz) where it needs to be.