Interpreting impulse responses - Home Theater Forum and Systems - HomeTheaterShack.com

Old 04-01-13, 02:52 AM Thread Starter
Senior Shackster

Matt

Join Date: Jun 2006
Location: London, UK
Posts: 244
My System
Interpreting impulse responses

I have been attempting to understand these charts recently in an attempt to assess whether the channels in my system are adequately time aligned (and out of general curiosity). As such I have a few, quite basic, Q's about the data. I believe the charts I have so far contain some measurement errors so have not posted them.

1) the y axis is amplitude therefore 2 impulses that have equal but opposite amplitude are exactly out of phase, correct?
2) is it best to use dbfs or %fs? Why? Or rather when would you use one or the other?
3) can you align sub with mains by overlaying 2 distinct measurements or is this shown via an excess group delay?
4) how should one treat a tripole surround (ie designed to produce reflections)? Same as a monopole or set to arrive earlier?

Thanks
Matt
3ll3d00d is offline

Old 04-02-13, 08:56 AM Thread Starter
Senior Shackster

Matt

Join Date: Jun 2006
Location: London, UK
Posts: 244
My System
Re: Interpreting impulse responses

anyone with any views on the above? or any links to a decent laymans treatment of the subject?
3ll3d00d is offline
Old 04-02-13, 12:59 PM
Elite Shackster

Join Date: Mar 2009
Location: Michigan, USA
Posts: 1,368
My System
Re: Interpreting impulse responses

Quote:
3ll3d00d wrote: View Post
1) the y axis is amplitude therefore 2 impulses that have equal but opposite amplitude are exactly out of phase, correct?
??
If 2 IRs of similar speakers (such as FL and FR) appear to flipped vertically from one another such that one initially moves positive and the other negative then the 2 are "wired" with opposite polarity. This is different from "Phase". This problem should always be corrected first. It is not necessarily a problem if the speakers are dissimilar.

Quote:
2) is it best to use dbfs or %fs? Why? Or rather when would you use one or the other?
We must use %fs to identify polarity as the %dBfs presents all data as negative dB. We could use either for Mains timing, but I prefer to look at %fs. For timing purposes the REW loopback feature must be activated.

Quote:
3) can you align sub with mains by overlaying 2 distinct measurements or is this shown via an excess group delay?
It's easiest to align the SW to mains by using the REW RTA function. Other methods can be use, but are much more complicated and best left aside unless you are just after a major learning experience.

Quote:
4) how should one treat a tripole surround (ie designed to produce reflections)? Same as a monopole or set to arrive earlier?
This would be relatively non critical so just align the initial IR rise/fall to the other mains.
jtalden is offline
Old 04-02-13, 05:21 PM Thread Starter
Senior Shackster

Matt

Join Date: Jun 2006
Location: London, UK
Posts: 244
My System
Re: Interpreting impulse responses

Thanks for the response.

Quote:
jtalden wrote: View Post
If 2 IRs of similar speakers (such as FL and FR) appear to flipped vertically from one another such that one initially moves positive and the other negative then the 2 are "wired" with opposite polarity. This is different from "Phase". This problem should always be corrected first. It is not necessarily a problem if the speakers are dissimilar.
OK, I have my terminology confused. I have another Q related to this, what does + and - mean on a %fs scale? for instance in this thread, the "it should something like this" chart shows an initial, smaller -'ve value before all channels peak at +100% whereas in this thread your sub responses shows the opposite (initial +'ve peak followed by a larger -'ve peak) pattern. Is this meaningful or just completely irrelevant as long as they're all going in a consistent direction?

Quote:
jtalden wrote: View Post
It's easiest to align the SW to mains by using the REW RTA function. Other methods can be use, but are much more complicated and best left aside unless you are just after a major learning experience.
I think I am after a minor learning experience is this method just by running 1 front + sub and adjusting phase to achieve maximum spl in the crossover regio or something else?

Quote:
jtalden wrote: View Post
This would be relatively non critical so just align the initial IR rise/fall to the other mains.
OK thanks. I have some doubt about the measurements I currently have but they appear to show the surrounds arriving at ~5ms whereas the mains arrive at ~10.5ms. These values correspond v closely to actual distances which are entered in the processor (albeit the specific values are those suggested by audyssey).

Thanks
Matt
3ll3d00d is offline
Old 04-02-13, 06:49 PM
Elite Shackster

Join Date: Mar 2009
Location: Michigan, USA
Posts: 1,368
My System
Re: Interpreting impulse responses

Quote:
3ll3d00d wrote: View Post
...
I have another Q related to this, what does + and - mean on a %fs scale? ... Is this meaningful or just completely irrelevant as long as they're all going in a consistent direction?
Check out This Thread The 2 IR traces are opposite polarities.

It is not relevant as to which direction is set so long as all matching speakers are the same.

Quote:
I think I am after a minor learning experience is this method just by running 1 front + sub and adjusting phase to achieve maximum spl in the crossover regio or something else?
Yes, the RTA method adjusts the distances of the SW to maximize the SPL in the XO region. It optimizes the XO handoff by aligning the phase there. This is a manual method though and it is not relevant if Audyssey is used. Audyssey sets the distances and then optimizes the EQ for those settings. If the distances are manually changed afterward by a significant amount, the EQ in the XO range will be negatively impacted. A small distance adjustment is trivial, so I wouldn't expect any adjustment to be necessary. You can check it out to confirm however.

Quote:
OK thanks. I have some doubt about the measurements I currently have but they appear to show the surrounds arriving at ~5ms whereas the mains arrive at ~10.5ms. These values correspond v closely to actual distances which are entered in the processor (albeit the specific values are those suggested by audyssey).
It does sound like something is wrong. Audyssey normally does a reliable job of setting speaker distances.
jtalden is offline
Old 04-03-13, 09:05 AM Thread Starter
Senior Shackster

Matt

Join Date: Jun 2006
Location: London, UK
Posts: 244
My System
Re: Interpreting impulse responses

Quote:
jtalden wrote: View Post
Yes, the RTA method adjusts the distances of the SW to maximize the SPL in the XO region. It optimizes the XO handoff by aligning the phase there. This is a manual method though and it is not relevant if Audyssey is used. Audyssey sets the distances and then optimizes the EQ for those settings. If the distances are manually changed afterward by a significant amount, the EQ in the XO range will be negatively impacted. A small distance adjustment is trivial, so I wouldn't expect any adjustment to be necessary. You can check it out to confirm however.
OK thanks. I should get time to take more measurements tomorrow to doublecheck. Just to be clear, are you saying that if I use audyssey then I should not (must not) manually edit the distances as audyssey should have done the right thing (and if I hasn't then I should repeat the audyssey setup)?

Cheers
Matt
3ll3d00d is offline
Old 04-03-13, 09:44 AM
Elite Shackster

Join Date: Mar 2009
Location: Michigan, USA
Posts: 1,368
My System
Re: Interpreting impulse responses

Go ahead and check it out and try making adjustments. Usually Audyssey is pretty good in my experience, but that is limited to my personal system and maybe 3 or 4 others here on the REW forum. Check out the first post on the AVS "Official" Audyssey forum for links to the Setup Guide. Some others users have found significant benefit in making distance adjustment. They have posted a detailed procedure and provided some examples in the Guide.

Any automated setup system makes assumptions about the situation and if the conditions are significantly different the results may be less that satisfactory.

jtalden is offline
Old 04-05-13, 07:10 AM Thread Starter
Senior Shackster

Matt

Join Date: Jun 2006
Location: London, UK
Posts: 244
My System
Re: Interpreting impulse responses

I reran audyssey last night & took some additional measurements; each individual channel alone and with the SW with all EQ (audyssey and BFD) off. The individual channel measurements were with the sub switched off, the SW measurement was from the FL with the power amp switched off.

A couple of caveats to the measurements

- I was having a lot of trouble with input levels, the SPL readings were correct (and by ear, I'd say they were correct) but it took me a long time to get REW not to complain about them being low. I'm not sure what was going on here, possibly some artefact of having played around with ASIO (these were taken with the non asio driver). The upshot is the main channels readings came out way high on the SPL (about +30dB) but they definitely were not that loud, I am unsure whether this renders the readings useless
- I think the SW reading is wrong, I think the processor may have been in direct mode & my processor only outputs to the sub in direct if it's in LFE+Main mode (I have small speakers so use it as LFE only)

If the readings, excluding the sub, are valid despite the level errors. My interpretation is that;

- FL & FR are spot on

- CC is v close

- SL & SR are quite a way off

The ETC plot suggests rather a lot of relatively loud reflections going on too.

Is that a reasonable interpretation of the above?

Thanks
Matt
3ll3d00d is offline
Old 04-05-13, 10:51 AM
Elite Shackster

Join Date: Mar 2009
Location: Michigan, USA
Posts: 1,368
My System
Re: Interpreting impulse responses

Your comment are all correct.

The distance settings of the SL/SR are off by 1.32 m (4.3 ft) . Does the distance setting in the AVR match a physical tape measure distance? Did Audyssey apply these settings? The front speaker distances look fine.

The Measurements of the front speakers alone are exactly the same as the front speakers + SW (see SPL chart below). Possibly the SW was not working for some reason. Check out all the AVR settings and try to correct this. The SW alone appears to be just baseline measurement noise. Did you hear the signal when the measurement was made?

The REW SPL calibration is incorrect as you stated, but that did not impact the relative levels of the main speakers. You should correct this as well. The low level that REW reported indicates the incoming mic level was low. Possible you just used the wrong range on an SLM being used for a mic? If not, then a recalibration of the REW SLM is in order.

jtalden is offline

 Bookmarks

 Tags impulse , interpreting , responses

Message:
Options

## Register Now

Random Question
Random Question #2

User Name:
OR

## Log-in

 User Name Password Remember Me?
Human Verification

In order to verify that you are a human and not a spam bot, please enter the answer into the following box below based on the instructions contained in the graphic.