"Official" Audyssey thread (FAQ in post #1) - Page 1702

It's nothing to do with DEQ, Feri. See my other more detailed reply.

Grasshopper, to reach the highest levels of Audyssey technique requires one to be accomplished in the art of acoustic invisibility.

Quite. I am booked in to THX's labs to be retro-fitted with the new Stealth Mode.

lol I'm working at it!


All of this SPL vs x-over vs distortion talk, is exactly why I brought the question up. It seems to me that as I turn it up louder and louder, my speakers may start to struggle with a 60hz setting, and might not with a 90hz setting.

It seems to me this afternoon's posts add yet more reasons (than those discussed recently in this forum) to push the xover up to 80 Hz. If you think the woofers in your mains are going to exceed Xmax (become nonlinear) at your max listening level or you aren't sure, then push the xover up to 80 Hz. If your mains are particularly small and you want to listen at high levels, push the xover up to 90 or 100 Hz. Most subs are limited by Xmax only at the low end and do fine at 80-100 Hz.

Speakers that do not remain linear at higher frequencies (above the subwoofer region) when pushed to high listening levels remain a problem and we have no options there but to EQ them at reference level and hope for the best at higher levels.

These reasons to push the xover up apply even with XT32 since XT32 can do no more about keeping our drivers all within their Xmax than the earlier Audyssey incarnations. Here I'm referring to some early speculation that with XT32 lower crossovers may be preferable.

I have plenty of capability in my mains so hadn't given the above any thought. But, when responding to queries from those new to Audyssey EQ, it's worth asking about mains and surround capabilities before suggesting that staying with Audyssey's recommendation is a safe bet.

I think folks are really talking about what I think of as compression - - not necessarily compression from high levels over time, which is a thermal effect, but compression from lack, ultimately, of excursion. So a speaker that can cleanly hit 85 dB with one watt of input may put out 105 at 10 KHz, and at 500 Hz, but not be able to physically move far enough to make 105 dB at 40 Hz with 100 watts.

So because of the way physical limitations of the drivers themselves, a speaker's frequency response may be different as you push it hard enough. You can see tons of examples of this in the data-bass tests of subs. After some point, the sub cannot "keep up" at lower frequencies, although at, say 50 Hz or so it may continue to increase in spl linearly with increased input.

Great info, thanks for your opinion. I have been A/B/C-ing all sorts of settings while watching Saving Private Ryan @ -5db (about as loud as I'll realistically ever go) and while I can't hear any distortion at any x-over setting, I believe 90hz sounds better at that level. I know I *feel* better at that level, knowing that I've taken some burden away from my mains.

Thanks again for everyone's input!

Hey all. I have a onkyo 809, polk tc265 in wall for mains, csia6 for center, and rc80 for surrounds in 5.2 setting. My subs are pa-120.

My mlp is on a couch along back wall. I am about 13 feet away from front wall. Can't change this since tied to the WAF.

When running xt, I get all xovers at 40. I normally raise to 80. I tried playing with all variations myself, but can't seem to find that sweet spot. I wanted to ask you all what settings u might suggest with my setup.

Thanks

Nice table behind couch enabling you to get couch 2 feet from rear wall might help and be consistent with WAF

If an EQ could magically make all speakers perform to any desired performance level, we could buy the worst cheapest speakers and achieve perfection.

I think the conflation is between the concept of a "brain" ASKING for a different response from the speaker, and the concept of a speaker being CAPABLE of that response.

Sure, the eq can reduce output at a frequency. It can even ask for more from the speaker at certain frequency ranges. Lots of speakers can be eq'd for a more smooth response (at some easy to produce spl).

But each 3db boost is double the watts to the speaker, eating into its headroom of power handling, either continuous or peak. If we're talking about 105db peaks, and it needed 300 watts before the 3db boost, then now you're asking for 600 watts. If the speaker has 400 watt peak power handling, it will fail to perform by distorting / compressing.

I think one was within a 3-foot diameter (couch), while the other was around 4 feet (two rows).
Do all 32, with mic positions 1.5 inches apart. (I'm joking) For a 4x3 area I would do a roughly 3-foot diameter circle, described by 6-8 mic positions. Any other listeners sitting on either side of you? Are they as critical as you when it comes to audio?

You have this right. If you are able to look up the TH parameters of the driver in question, you should see one called Xmax. That is the limit of the cone excursion within its linear region. The cone can usually move further and the next limit is Xmech. Xmech is the physical limit where the voice coil form hits the magnet end structure with the cone moving in or tears the spyder with the cone moving out.

Between Xmax and Xmech the driver becomes nonlinear, with increasing drive causing proportionally less cone movement. I.e., the spyder begins restricting the cone movement. It's similar to an elastic material or spring exceeding it's linear region. Stretch a rubber band and you will sense that it is getting longer with the same increase in force for each inch of stretch. Then at some point it gets harder to move it another inch. Then a bit more and it breaks.

Of course, the Xmax and Xmech are of most concern at the lowest frequencies that a driver is asked to produce because that is where the greatest cone excursion is required to achieve the desired sound power level.

Thanks. Other listeners not at all critical. My theatre only has two seats, mine and, er, the other one. The other one is sometimes occupied by my SO, but she isn't critical, and in any case the sound in her seat is pretty good since I added the second sub anyway.

Thanks for the advice - I will try that when my Pro kit arrives and the FW is ready. Will report back with my findings of course.

Great post. My speakers are THX spec, so 80Hz is an easy choice for me. Nonetheless, Audyssey recommends 50 or 60Hz usually, but I always raise them back to 80Hz. There seems to be so many compelling reasons to do so that I find it hard to recommend anyone stays with lower XOs. HST, many regulars here do run XOs in the 40-80Hz reason and they are happy with the results. This latest round of discussion wrt to frequency vs SPL has added yet more knowledge to the knowledge base IMO.

Guys, I think the issue of speaker FR vs. SPL needs a bit more consideration. If we are talking about driving our speaker systems to their upper limits there will surely be issues, the least to say, eh? Not many HT systems will reach 105 dB SPL without degradation of specs, I'm sure we can agree here.

But again, why would a speaker behave differently at "normal" listening levels, i.e. why would a speaker show different FR at -30 dBfs input than at -20 dBfs input or -40 dBfs input. I always thought speakers are linear at their "normal" working level and do not suffer changes in FR at different SPLs. If that would be the case, then how can Audyssey set their filters with a -30 dB sweep, meanwhile the FR of a given speaker will change as the SPL changes throwing off the already calculated Audyssey FIR filter response.

Anyone?

Hi. Settings can only go so far; we're likely trying to make the best of a suboptimal layout of speakers and MLP. Photos and a room layout drawing with speaker placement, etc. would help. Some flexibility on the part of SWMBO irt room layout is required. See the attached illustration for the recommended ideal angles.

I'm unsure what you mean by "can't seem to find that sweet spot". Using "sweet spot" in the traditional audiophile sense, as your FR/L speakers are fixed in place, to find it, just measure out an equilateral triangle using those speakers-your MLP/sweet spot will be the third point of the triangle. Use an online calculator like this one . If FR/L are 8' apart, MLP should be about 7' from the front wall. If they're 10' apart, MLP should be 8.5' away.

ps I am assuming your speakers are not the interesting and innovative Polk "-ip" model with built-in amps, networking capability and DSP, as that processing would conflict with Audyssey.

Feri, read back on the posts made so far on this issue and it should all become clear.

Frequency response does change as SPL changes - there is no argument about that at all. If you need further convincing, google is your friend.

I can't personally explain it any clearer than I already have. This paper goes into great detail but you only need to read the first page or two to see that FR does indeed change with power:

http://audioweb.cz/down/dynamic-line...udspeakers.pdf

The issues wrt to Audyssey are a red herring AFAICS. Audyssey cannot change the physical properties of a driver and nobody has ever suggested it could.

I know you put a lot of faith in Chris - ask him if FR varies with SPL - a yes or no answer will do. He will say 'yes'.

as you basically say, speakers are linear until they are not. Like any sub you see tested on data-bass, the combination of woofer excursion, cabinet and power will cause the low frequency response to stop keeping up at some point. Somebody does linearity testing of speakers but I can't remember who. I've looked at several of their tests and in general linearity tracks well until you hit a wall, which of course differs from speaker to speaker.

As you say, speaker's are linear up to a certain point. It may be Zaph Audio that you are thinking of, he's done testing of linearity.

I think this speaker issue is analogous to amplifiers. We all know that a spec of "100 watts" means nothing if we don't know the max distortion or frequency range.

So picture the information we have on speakers like what we had prior to the FTC requiring more qualifiers on amp specs. The manufacturers are usually not telling you how the speaker behaves at the limits of its performance.

Program Peaks
djbluemax1's discussion of XMax and XMech explains how on program peaks, you can reach the magnetic and mechanical limits of the driver. At these high levels, if for example output falls off for bass/midbass drivers, then you can see how there'd be too much treble in the speaker's overall output? This is a nonlinear response at high peak levels.

Program Continuous
For continuous levels of high volume, it's about thermal compression.

At levels that are alarmingly a small fraction (starting at 1/10th) of the speaker's rated continuous power input, you'll start to get thermal compression. Speakers are inefficient (lots of heat, not much sound). As power input goes up, the voice coil heats up. As its temperature rises, its resistance goes up. For instance, add 3dB more power (double), and you only get 2dB more SPL, and you get an increasing amount of heat, and so on. At some point, more power doesn't yield more SPL. Also, if the impedance curve is changed for some drivers and not others, you're going to get a different frequency response.

A speaker's max continuous power rating usually only suggests when the speaker might fail; not when it might start to sound bad.

Below is the graph provided by JBL for their LSR6332 speaker. Here's the original PDF. You can see the effects of 10, 30, and 100 watts of continuous power.

What a great comparison that is!

And distortion through magnetic flux modulation is another form of instantaneous compression that yields a non-linear response.

It's a problem that appears at high drive levels.

There are all sorts of things they do to make the driver more linear and I don't pretend to really understand any of them.

Along these lines, what would be the procedure to optimize a (non-pro) MultiXT32 run for *JUST* the MLP? Would it be what sanjay described, or something more?

I so seldom have visitors that the MLP is all that matters to me. Because of this I ran XT32 and used very close measurement positions. I like how this worked.

Take the first reading at the MLP as usual. Then cluster the other 7 measurements around the MLP in an area about 3.5 feet x 3.5 feet, making sure to follow the usual rules of mic placement as per the setup guide. You could use an even smaller measuring area if you wanted to but I do think that Audyssey needs a bit of room info to go on so clustering too tightly might be counter-productive. Try it and see which you prefer.

attached are some ht photos. No, don't have the "ip" models.

The mlp to wall is 13 feet. Little less with center since not in wall. I have a square Ottoman 3x3 , in between the mlp and front soundstage.

I'm always interested in your setup, Theresa - how close is 'very close' for you?

About a foot.

This happens to be what I do now, simply because all the 'other' common LPs in my room are within 3.5 feet of mine (laterally only...there are no seats in front of or behind my MLP, though I still do take measurements in front and behind).

So I wonder how 'tight' one can go in the clustering and still get improvement at the MLP.