We Built it, We've measured it, HELP US TWEAK IT - acoustics of the BLACK CAT

I'm posting mostly to subscribe to the thread, because I want to see how this develops (and I'm taking copious notes ).

However, with regard to the reflection from the ceiling, is there enough room to put some treatments above the LCR behind the screen to attenuate the energy before it gets to the ceiling? You know, treat the source before it becomes a problem.

Also, any idea what's causing the peak in the quiet room response at 60 and what looks like 120? Could it be electrical noise at 60Hz and at the first harmonic rather than actual acoustic noise?

Being that the peaks are at multiples of 60 Hz, it does look a little suspicious. Like maybe the power supply to your laptop isn't filtering out the 60 Hz sine wave from the outlet. Although, looking at the graph again, that second peak may not be quite at 120 Hz.

If you're troubleshooting those peaks, you might try turning off as much electrical stuff in the room as possible, like lights and the chair transformers (which can both produce audible and electrical noise), and maybe even let the laptop run off of batter power long enough to make the measurements.

It may very well be a noise in the room, and not worth the effort to locate. I was just curious

Glad to hear my other random thought was a success!

Like J_P_A, I mostly want to get subscribed to this thread - I'm exceedingly interested. Thanks to you guys for taking the time.

That said, there might be an improvement to be made in the way your ETC graph is put together. The spike at 0 is, I believe, the initial direct sound from the speaker(s?). There is something to do with loop-back (I think) in REW that allows the time to start when the sound begins to be emitted, and the first spike happens later in time - it helps to look at it this way to place everything in the same reference for time and space. Also, I'd suggest producing ETC graphs in this way for each of the seven (?) main channels individually. We may find that the surround and rear channels are less useful in tuning the overall response, but I think there will be useful information to be gained - not to mention it will be helpful to label each graph very precisely in terms of driven channel and listening (mic) position. And if we're being thorough, one should properly know the response of the speaker outside the room (both on and off axis) - that will help, I think to define the nature (frequency content) of specular reflections.

I say all this as though I know what to do with all this extra data when I see it. That's far from the truth, but it's all part of the full picture. Also, can we get room dimensions and speaker distances to the listening position(s)?

Someone please correct me if I'm wrong - I don't want to give these guys extra work to do without reason.

Regarding the 60Hz noise - was the sound card calibration run under the same conditions as the measurements? (in terms of battery operation, power supply, etc)

Fred

Well, I was really hoping that someone more experienced than myself (as I have no - ZERO - practical experience) would offer some ideas.

In the absence of that, let me give you some suggestions for what to try. Feel free to ignore me, or maybe my wrong-ness will at least spark some conversation.

My feeling is that one should start with the bass. The porous absorption that will be used (unless you are open to more advanced tools, that I can't really help with) will continue to suck high frequency energy out of the room. It would seem backwards to me to go about balancing reflected energy with direct energy and then go add a bunch of porous absorption messing it all up - so start with the bass so that you know how much of the reflected sound energy you've lost and can go about carefully adding it back (by either removing linacoustic and poly batting, adding thin reflective panels, or adding diffusors or some sort). Plus, if you've got an option to move subs, it's best to do it now, not later. Is that possible?

I don't recall exact built details, but I see in Morph1c's first post in the build thread he's indicated one sub behind the screen. It would seem that moving it side-to-side behind the screen would be easy. What about moving it out into the room or adding another? I'm looking at the decay times and seeing the ringing at 28Hz, which I would assume corresponds to a response peak (but we don't see a graph of the frequency response - can we get white or pink noise through the sub and see the response?) There's a similar hang-over around 50Hz. If these are troubling when listening to music, they would be the first thing I would look into. If they are a problem, I'd suggest (in order of decreasing effectiveness): adding more subs, moving the sub(s), a little EQ cut, a tuned bass absorber (or two), more bass traps. Of course, if the bass response is acceptable, we can move on to specular energy (above approximately 500Hz).

The psycho-acoustics I have read about say that you should not have high gain reflections at the mic within about 10-20 ms of the direct sound. Second, as you've noted, the room is somewhat dead. Given that, the goal should be redirecting or diffusing specular energy (in the horizontal plane as much as possible) without absorbing it. As you noted, the first high-gain reflection comes very soon after the direct energy (only 2 or 3 ms). If that's coming off the ceiling, then at least you've found it. Allow suggest a tilted panel that would reflect the sound to the back wall, over the heads of the listeners. If you have a spare sheet of plywood, maybe you can rig up a way for it to hang down in front of the screen at an angle? Here's a visual reference:


To go along with the "quiet time" following the direct sound from the speaker, at least in a two-channel environment, listeners often like a "kicker" to clearly define the end of that period. (I couldn't tell you why, or if it will translate well to home theater, but that's the only frame of reference that ever gets discussed and cited around here) So, if you want a kicker, where do you generate the high-gain reflection that will produce that psycho-acoustic effect? A little math - sound travels approximately 1.13 ft in each ms. For the reflected energy to arrive at least 10ms after the direct energy, the path it travels must be at least (10ms divided by 1.13ft/ms equals 8.8ft) 9 feet longer than the direct sound. In order for the reflected sound to be of adequate gain, it will need to come from a large area. The large central area of the rear wall is probably adequate size. If it's 4.5 feet from the rear rows ears, I'd say it's a good candidate. All that is to say: try it again with no linacoustuc on the rear wall and see if that helps.

And naturally, as a general statement, it's tough for anyone to know with any confidence what changes in a graph are going to be audible with real content. At every step, listen to some real sounds for reference. Maybe pick the first three minutes of The Dark Knight and watch it over and over - it's got good bass, surround effects and dialog so that you can compare.

Now that I've vomited that into the thread, I hope it's useful.

Fred

The HVAC Noise


The Quiet Room


Going back to the 60Hz and 120Hz noise, is there any reason that we can't compare these two graphs? I'm sure the mic locations weren't the same, but clearly there is much less noise in the new graph. The 60Hz and 120Hz spikes weren't in the measurement from the HVAC duct noise. Any explanation? If the testing methodologies were the same, I don't have any...

Screen and PJ are hung, isn't this where are work stops?

Would the chairs transformers even make noise when the chairs aren't moving? - seems like no movement implies no current being drawn...

Before obsessing over thinking this room is too dead, I've heard that different criteria might need to be applied for systems with surrounds. Specifically, I heard that it might be desirable to have a slightly dead room. Does anyone know anything about it?

.

 

FM.pdf 103k . file

Dennis posted the answer. It is irrelevant. No sense in worrying about issues that are irrelevant. I would also look into acquiring a better sound card than using the one in the laptop as well as a multiplexer.

Cheap, unreliable equipment. Cheap, unreliable results. Equally to the point, you'll never get accurate results when you're attempting to measure beyond the ability of the equipment. Equally to the point, you should not be measuring room response until you've measured the speakers. BTW, a $2000 microphone becomes a cheap microphone if it is dropped, the speaker stand falls over or you just toss it in your overnight bag.

Given your proximity to the rear surrounds, you might experiment with a 6mil or so plastic on the front of the bass traps to make them reflective at higher frequencies. Same could be true on the back wall at surround height. Not sure that will solve any of your measurements, but would liven up the surround presentation.

Another random rambling, but I wonder if you can make the rough pass with the equipment you have. Get the low hanging fruit, and get comfortable with taking the measurements and how room changes affect the results. Then, could you find somewhere to rent a high end mic for a day? I know a lot of photography shops rent those high end lenses that us normal folks can't afford, seems like an audio/recording shop might rent a high end calibrated mic as well.

There is a difference between "cheap" and "inexpensive".

There's a bunch of info on the 'net about making a measurement mic based on a Panasonic electret mic capsule. Here's an example that compares a Panasonic capsule-based mic to an expensive B&K: http://www.purebits.com/miccomp.html

If you google "Panasonic electret mic capsule," you'll find a bunch of links.

The other components in the measurement system -- the software and the laptop's sound card -- might also be limiting factors. I haven't ever seen any third-party tests of the OmniMic system. The system might be great, but it's hard to judge without third-party testing and/or a large group of users.

It might be interesting to run REW as a comparison point. With REW, you would at least have a good sized user base to provide help.

So far it seems that the suggestions are:

regarding testing equipment

• either

  • make do with what we have for now
  • rent a professional mic for a day when ready

• or

  • use REW (larger community)
  • get another mic and use external sound card [note: I've read reports of others using Omnimic with REW so maybe just a sound card is all that's needed]

regarding room

• put 6mil or so plastic over bass traps and perhaps back wall
• construct some kind of visor to better modulate the center-channel's ceiling reflection
• move the bass around (behind the screen) and/or add another bass in room

other thoughts

• listen to the same part of a track over and over, like the 1st 3-min of Dark Knight (e.g. forget the readout, what are your ears telling you?)

Looking into REW, it looks like it might take a couple weeks to get a good sound card and build a mic - any NoVA folks have a more expedient solution?

Who's local?

http://www.homeacoustics.net/ for who's local. Choose Dealer locator. Input country and state. I would consider only those who hold HAA Level II Guru status. These are the folks who have proven their capabilities and usually own the appropriate test gear.

For Big's journey http://www.homeacoustics.net/DealerS...DealerHome.asp

A class is usually held the week after CEDIA in September. I would not use the Omnimic. There is more to it than just having the right test gear. You also need some way to control the response.

Looks like the closest who meets Shawn's criteria is Fred Dietzel of Lebabon, PA. That website shows no one with Level II Guru status in VA or MD (is there a higher status that supersedes LII Guru, such that Guru is not listed?)

Name Fred Dietzel
Company Name Enhanced Audio & Video
City Lebanon
State/Province PA
E-mail enhancedav@comcast.net
RTA Model DAG 5161
Certifications
HAA Level I Certification
HAA Level II Guru

That's closer than I thought, at a 2.5 hour drive (Lebanon to Leesburg).

Only other level is Level III. Only three people hold that level to my knowledge. Two of which hold the distinction of master calibrators. Dennis holds instructor level as well. Level II calibrators though are usually quite good, and have accomplished the prescribed requirements for that level. Many have taken the course. Few have completed what it required.

I do appreciate the fact many want to measure and calibrate their own rooms. I better appreciate when people want to understand what's happening. By the same token, you can't have 3000 hours on Microsoft Flight Simulator, go to the airport and hop into a real airplane and expect a positive result ... experienced, professional assistance is still required. (Not suggesting MS FS experience is not helpful.) So, some hints.

First, you absolutely need to know the capabilities of your kit ... from the microphone(s) to the display. If your microphone doesn't have a calibration report, you could be chasing a "reported" problem that can't be fixed. In some cases, for example, you can turn your on your gear with no mic connected and read a response ... basically the internal noise of your measurement gear. In the plots above, that equipment is not at all accurate below 20dB if not 30dB. It takes some really good "stuff" to measure below 30dB.

Second, you have to get your arms around what you can fix, and what you cannot fix. You can move, add, subtract, change treatments, move speakers, and move seats. You have to have the fundamental knowledge and experience to know what can be fixed (and why) using the above methods...along with the impacts of doing so. Twisting one dial can muck up the other dial. If you do not have, or do not care to have, some form of EQ device, you have to recognize that limitation ... you cannot fix everything with positional changes or treatments. Understanding goes a long way ... you cannot "fix" a width modal reponse by putting stuff on the front or back wall. If you have a dip at 3kHz, what's the quarter wave length so you can look for potential phase cancellation problems. Are you chasing one of those many "nasties" that can be measured but is not at all audible? Or a "nasty" that doesn't occur in a seating location?

Third, you absolutely must do near field (on and off axis) measurements of all your speakers. You need to know when reviewing room response measurements which of those responses are room generated vs speaker generated. I cannot tell you the number of times someone has spent days trying to fix a FR problem which, in fact, a speaker issue. (Is that dip in response SBIR or a poor crossover in the speaker?)

Some other pointers. Do not move the bloody microphone between measurements! You cannot compare two measurements if the microphone has moved. Do not change three things, run the measurement again. Yup, the result changed; but, which of the three things you did fixed the problem. If you change one thing (speaker toe-in, treatment, seat position, whatever), measure! Then make the second change and measure again. Don't assume the change you made to the left speaker will be the same change required for the right speaker.

Why does a professional calibrator want two, maybe three days on site? Consider all the individual measurements which must be taken .... seat by seat. The mantra is "no bad seats". That takes time (some of which may be wasted if you don't have the kit to fix an issue).

Let's review the "Quiet Room" and "HVAC" plots above (I wasn't there so some speculation here). We have to assume (perhaps a bad assumption) the microphone didn't move between measurements. From 500Hz to 20kHz, no change. The plots show all of that below 15dB. That is bogus...there is no way that kit can provide results at that low a level. Below 500Hz we see some differences; but, again at such low SPL we have to question the accuracy. None-the-less, is it clear more than one thing changed. We can see what would appear to be noise contribution from the HVAC system; but, there are other differences as well. What caused those changes? The compressor wasn't running? The Fridge upstairs wasn't on, was on? Once again, however, these readings are really on the fringe of the equipment's capabilities and likely not at all accurate (and the two plots aren't overlaid which makes it all that more difficult to compare).

Should BIG run out and buy (or rent) more expensive equipment? No (at least not until he takes some classes). Just recognize the limitations of the kit you're using. The classes likely cost close to what it would take to bring in a top pro. Get a pro to measure and calibrate the room...learn from what they do so you can continue the care and feeding of your room in the future. If you're planning on providing calibrations as a service...get better equipment and get some training (it will save time and frustration).

I've sent email to Fred Dietzel. I have no idea how much calibration service might cost, but five hours of driving and a hotel for a couple nights is already up there...

I can't believe there isn't a Level-I more nearby - seems like there is a market for this skill in the area