AVS Forum banner

The Audyssey Pro Installer Kit Thread (FAQ in post #1)

385K views 6K replies 258 participants last post by  Mike_WI 
#1 · (Edited)
For the Audyssey Pro FAQ click:
 http://www.avsforum.com/forum/90-re...kit-thread-faq-post-1-a-102.html#post22429521
 
The only thread I could find on this topic on AVS is defunct so I thought I'd try starting a new one. We shouldn't have to go to HTShack for this discussion, should we? There seems to be some renewed interest in Pro calibrations lately. So let's see if it flies.

I'm new to the game, and have only run Pro a few times on my XT32 equipped A100. I only ran 8 positions as the instructions recommend 8-10 for average-sized rooms as a good start. Can add more later. So far I'm very impressed with XT32 as it really kicked things up a notch from XT in my AVR4310 without Pro. With Pro so far I haven't tweaked any curves or played with midrange comp. I overrode the suggestions and opted for 80 Hz xovers. So far, my impression is that Pro adds further SQ improvements, smoothness, integration and polish.
 
#77 ·

Quote:
Originally Posted by SoundofMind /forum/post/20695426


A "Luke tip": One can optimize the speakers, subs, acoustic treatments, etc. using the Pro kit by comparing the "Before" graphs from successive runs.

No EQ is in effect during Audyssey setup, so why would the "before" graphs differ other than in sub level?


I could see some differences if speaker distance settings were in effect, due to the effect on phase, but aren't they zeroed too?
 
#78 ·

Quote:
Originally Posted by SoundofMind /forum/post/20695426


A "Luke tip": One can optimize the speakers, subs, acoustic treatments, etc. using the Pro kit by comparing the "Before" graphs from successive runs.
Quote:
Originally Posted by noah katz /forum/post/20697097


No EQ is in effect during Audyssey setup, so why would the "before" graphs differ other than in sub level?

Because changing speaker position, sub location and adding/subtracting acoustic treatments would make a difference.
 
#79 ·

Quote:
Originally Posted by SoundofMind /forum/post/20695364


It also looks to me per your subs "Before" response that your setup required much less correction than mine. I wonder if dialing in time alignment with the Behringer might have improved overall response in the room- do you have a pre-Behringer cert to compare to?

Here is the Audyssey Pro certificate that goes with this graph (which I posted previously)...




As you can see my sub trim level was well below the proverbial -10.
 
#80 ·

Quote:
Originally Posted by SoundofMind /forum/post/20695642


Of course the Audyssey cert AFTER graphs are not real measurements. I'm trying to get more adept with the OmniMic system. But my OmniMic graphs do not sync well with Audyssey before OR after graphs so I'm messing up somewhere.

The Audyssey graphs show a spatial average of all measurements from a single speaker. Is this also how OmniMic works?
 
#81 ·

Quote:
Originally Posted by counsil /forum/post/20697281


Here is the Audyssey Pro certificate that goes with this graph (which I posted previously)...As you can see my sub trim level was well below the proverbial -10.

Yes indeed, so there goes my little hypothesis. The Audyssey Subs After curve looks really great though, about as impressive as your more current graph with the Behringer in place.


So this takes us back to whether the Audyssey After curves are accurate. Sorry if I missed this, but has there been an audible improvement with the Behringer in place that correlates with the better XTZ curves?
 
#82 ·

Quote:
Originally Posted by markus767 /forum/post/20697524


The Audyssey graphs show a spatial average of all measurements from a single speaker. Is this also how OmniMic works?

Hi Markus. I am a total newb with OmniMic and it has more features than I know how to use. If I understand your question, yes OmniMic can do averages from a single speaker. Also see my response to audioguy's question just a few posts up. I have been most focused on the averaged curves for the 2 subs together from both Pro and Omni.
 
#83 ·

Quote:
Originally Posted by SoundofMind /forum/post/20697767


Hi Markus. I am a total newb with OmniMic and it has more features than I know how to use. If I understand your question, yes OmniMic can do averages from a single speaker. Also see my response to audioguy's question just a few posts up. I have been most focused on the averaged curves for the 2 subs together from both Pro and Omni.

You need to make sure the measurements are properly aligned in time so the average is vaild.

Can you show the graphs (sorry if I missed it)?
 
#84 ·
No I did not post the OmniMic graphs, not sure I saved them properly but i sketched them and took notes so that is what I reported here. I do not understand the time-align stuff. I used the technique I described above, sampling a 3-5 spots in a relatively small area and using the summation feature in the OmniMic software for the combined curve. For a full Pro calibration I use a slightly larger area (almost 9 feet sq) and 8 positions. OOps gotta go to work...
 
#87 ·

Quote:
Originally Posted by markus767 /forum/post/20697959


You need to, otherwise your summed graphs could be plain wrong.
OK, Markus, but as I said, I'm a newb to OmniMic. Clarification: when I say, "I do not understand the time-align stuff," I'm not politely disagreeing. I mean I have no clue what you mean, why it's true or how to do it.
 
#88 ·
Quote:
Originally Posted by SoundofMind
OK, Markus, but as I said, I'm a newb to OmniMic. Clarification: when I say, "I do not understand the time-align stuff," I'm not politely disagreeing. I mean I have no clue what you mean, why it's true or how to do it.
Time alignment means all the wavefronts arrive at the measurement point in phase at the same time. It's the Audyssey "distance" measurement for the MLP. If you're measuring very far from the MLP your numbers could be skewed by phase issues. I don't know what (if anything) Audyssey is doing to compensate for it. I think markus767 might have asked but wasn't answered for trade secret reasons but I could misremember.
 
#89 ·
^Hmmm...Here's what I understand on the topic, which may seem fairly basic, so I apologize to those with a more sophisticated understanding. Feel free to correct me if necessary and further educate me where possible.


Distances (delay) of multiple speakers can only be correct at one spot, with Audyssey or without. That spot, referred to as MLP, is mic pos #1 for Audyssey.


On the topic of using systems for measurements, it has been repeatedly stated by Chris (audyssey) that to more accurately measure what Audyssey is doing, one needs to take multiple measurements at the same mic positions used for Audyssey. There was no mention of time alignment.


I have posted that for OmniMic I used a small grid, less than 2'X2' and for Audyssey I use a slightly larger one, maybe 3'X3'.
 
#90 ·
SoundofMind, here's an example. 3 measurements. Here are the impuls responses:




Now I've shifted the 3rd measurement by 2ms in time:




This is what the average frequency response from all 3 measurements in both cases looks like:




The two graphs look pretty different. One needs to know what the measuring software is doing in order to obtain valid results.
 
#91 ·

Quote:
Originally Posted by SoundofMind /forum/post/20704360


Distances (delay) of multiple speakers can only be correct at one spot, with Audyssey or without. That spot, referred to as MLP, is mic pos #1 for Audyssey.

True. It's impractical (but if you had a Kinect ...) to compensate for movement without something like a Realiser. A deeper question is how much it matters to the end result. The EQ is a single static transform on the signal and will be more correct in some spots that in others.

Quote:
On the topic of using systems for measurements, it has been repeatedly stated by Chris (audyssey) that to more accurately measure what Audyssey is doing, one needs to take multiple measurements at the same mic positions used for Audyssey. There was no mention of time alignment.

As I said I don't know if or how Audyssey compensates for phase changes.

Quote:
I have posted that for OmniMic I used a small grid, less than 2'X2' and for Audyssey I use a slightly larger one, maybe 3'X3'.

The example provided by markus767 is ~ .7 m (2.3 ft).
 
#92 ·
^Interesting. OK, so what's the bottom line? What system, which measurements are to be trusted? I feel like I walked into the wrong classroom and haven't taken the prereqs.



What I'll do is continue to experiment with the Audyssey Pro settings (delete midrange compensation, reset xovers to 80, select a different target curve), listen for changes and post results.


Maybe I'll end up sending OmniMic back w/i 45 days as it is not worth $300 if I can't understand and trust the graphs. Besides, the HT system sounds great now and I really need to catch up on the day job stuff...
 
#93 ·

Quote:
Originally Posted by SoundofMind /forum/post/20705333


OK, so what's the bottom line? What system, which measurements are to be trusted?

If you want to compare Audyssey results to other systems you need measure from a single spot. As soon as you start doing averaging to emulate multiple Audyssey measurement positions you've introduced things you cannot know or control because they're Audyssey secrets.
 
#95 ·

Quote:
Originally Posted by bodosom /forum/post/20706016


If you want to compare Audyssey results to other systems you need measure from a single spot. As soon as you start doing averaging to emulate multiple Audyssey measurement positions you've introduced things you cannot know or control because they're Audyssey secrets.

Agreed. The approach I take is to make before/after Audyssey calibration with an independent system and compare these, not to the Audyssey measurements.
 
#96 ·

Quote:
Originally Posted by Kal Rubinson /forum/post/20706420


Agreed. The approach I take is to make before/after Audyssey calibration with an independent system and compare these, not to the Audyssey measurements.

That is the right thing to do. Unfortunately nobody publishes these kind of measurements online.
 
#97 ·

Quote:
Originally Posted by Kal Rubinson /forum/post/20706420


Agreed. The approach I take is to make before/after Audyssey calibration with an independent system and compare these, not to the Audyssey measurements.


One point to keep in mind is that each Room EQ scheme has its own unique target final EQ response curve..

Once the system's calibration software is run, then the AVR's processor takes those coefficients, plugs them in, runs the equations and then...

Out comes a revised, room transfer function..


The challenge is that each flavor of Room EQ software, be it Yamaha, Pioneer, Harman/Kardon, NAD, Meridian and those brands running Audyssey or Trinnov S/W can/will sound differently..


Maybe its time for a respected publication or testing house runs a comparison face-off between the major EQ software players. The room and loudspeakers should be neutral, but one could eliminate the variable of amplifier contribution by measuring each output @ pre-outs..


Just my $0.02...
 
#98 ·

Quote:
Originally Posted by bodosom /forum/post/0


A standard run sets my subs to different distances and trims. A pro run sets them to be the same. Is that typical?


Here's one sub versus two (in the front corners).

I see the same behavior. It looks like this is the way Audyssey designed the two pieces of software. I certainly hope Pro is establishing the correct delays for my two subs, which are not equidistant from the MLP.
 
#99 ·

Quote:
Originally Posted by M Code /forum/post/20706682


One point to keep in mind is that each Room EQ scheme has its own unique target final EQ response curve..

Once the system's calibration software is run, then the AVR's processor takes those coefficients, plugs them in, runs the equations and then...

Out comes a revised, room transfer function..


The challenge is that each flavor of Room EQ software, be it Yamaha, Pioneer, Harman/Kardon, NAD, Meridian and those brands running Audyssey or Trinnov S/W can/will sound differently..

Maybe its time for a respected publication or testing house runs a comparison face-off between the major EQ software players. The room and loudspeakers should be neutral, but one could eliminate the variable of amplifier contribution by measuring each output @ pre-outs..


Just my $0.02...

The realist in me say that it'll never happen and even the optimist agrees.
 
#100 ·

Quote:
Originally Posted by bodosom /forum/post/20706350


A standard run sets my subs to different distances and trims. A pro run sets them to be the same. Is that typical?

Not in my experience. Audyssey Pro always set the distances differently. Maybe Audyssey Pro was hitting it's limit on how much delay it could handle? Maybe that's part of the reason why Audyssey Pro isn't calibrating my subs very well?
 
#101 ·

Quote:
Originally Posted by bodosom /forum/post/20706350


A standard run sets my subs to different distances and trims. A pro run sets them to be the same.
Quote:
Originally Posted by counsil /forum/post/20706932


Not in my experience.

I did another run and got different values for the distances but the trims are still the same. They still don't match standard XT32 but they're close. It's also not useful that the certificate graph only gives a single value for Subwoofer 1+2. And as long as I'm complaining it would be nice if all the trim numbers in the GUI matched the 'CV?' command.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top