The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Loudspeaker Choice for Impulse Testing Reverb & Delay Plugins
Old 23rd December 2010
  #1
Gear Nut
 

Loudspeaker Choice for Impulse Testing

I am just entering the world of impulse testing in an attempt to quantify the effectiveness of treatments I'm putting into my new room. I've done searches here, but cannot find a clear answer.

I've read that most test mics, even inexpensive ones, are flat enough to use for creating relatively accurate FR plots and waterfalls. But what about the loudspeakers creating the impulse source? Speakers have a large degree of frequency variation, and most don't seem to have enough extended LF output to create meaningful information in the low frequencies where rooms are usually most problematic. How do you guys account for this when measuring?

On a related note, when first doing measurements and getting an overall picture of the room, what is the best placement for both test mic and impulse source? Should the mic be at the proposed (38%) listening position with the speakers in a boundary area, or is there a better method?
Old 23rd December 2010
  #2
Gear Guru
 
Ethan Winer's Avatar
 

Lightbulb

Quote:
Originally Posted by Azure View Post
Speakers have a large degree of frequency variation, and most don't seem to have enough extended LF output to create meaningful information in the low frequencies where rooms are usually most problematic. How do you guys account for this when measuring?
If you think about it, the only speaker that matters for testing your room is the same speaker you use in that room. If some day you get different speakers that go lower in frequency, you can re-test and reassess your treatment.

Quote:
what is the best placement for both test mic and impulse source? Should the mic be at the proposed (38%) listening position
Yes, the microphone should be where your ears are when listening, and at the same height. It's not bad to measure elsewhere to see what happens at, say, the "producer's" couch in the back of the room. But what matters most is what you hear.

--Ethan

________________
The Acoustic Treatment Experts
Old 23rd December 2010
  #3
Gear Nut
 

Quote:
Originally Posted by Ethan Winer View Post
If you think about it, the only speaker that matters for testing your room is the same speaker you use in that room. If some day you get different speakers that go lower in frequency, you can re-test and reassess your treatment.
This makes absolute sense, but one of the issues I always have when dealing with a new room and my monitor setup is how to place a subwoofer and then adjust it's level and crossover point. My fear is that in this case, when testing, I will introduce so many variables using my current subbed setup, that I will get lost in the adjustment "abyss" and never come out heh. I guess was wondering if there was a more generic way of approaching it?
Old 23rd December 2010
  #4
Lives for gear
 
johndykstra's Avatar
 

Quote:
Originally Posted by Azure View Post
On a related note, when first doing measurements and getting an overall picture of the room, what is the best placement for both test mic and impulse source? Should the mic be at the proposed (38%) listening position with the speakers in a boundary area, or is there a better method?
Impulse response is not the test tone to use. You want a sweep. Particularly for low frequency measuring.

Actually, your first test should be done with the speaker firing into a corner, and your test mic in the opposite corner. Think floor left and ceiling right. This will be to capture a waterfall reading of the room, with as little direct speaker influence as possible. You are looking for 300hz and below for the range... windowed out as far as the longest "ringing". THis test will truly tell you what the room's response is.

At that point do as Ethan said, as a traditional listening triangle, and adjust speaker and mic to find the best waterfall response as possible. You are still only concerned with 300hz and down. You will see involvement of frequencies you learned about already in the first test, but also some new "dips" or nulls in response related to the speakers relationship with the front/side/floor/ceiling (SBIR) of the room. Moving the speakers location will shift the frequency of these nulls, and after a few tests of moving the speakers on all axis, you will quickly see this relationship. Aligning these new nulls over peaks caused by the room's boundaries can help you use SBIR to minimize peaks you know to be modal from your first test.

After this test, begin to treat the room for it's low frequencies. Depending on how low, or how intense will dictate your course of action. Relatively "weak" or higher freq. can be treated with porous corner traps... the standard things in 90% of the threads here. For strong pronounced very low frequencies, you may want to consider tuned traps. Jens, SAC, Lupo, GE, Magikman... are all good sources to look to in designing these.

Frequency response in the higher register is influenced by specular reflections combination with direct signal. THese are hunted out using the ETC response... what until recently was thought (by me at least) to be obtained using an impulse response, though it's being suggested that some measuring platforms are better suited for a sweep here as well. Any reflection that shows in the first 20msec. after direct signal should be down 20db. These early reflections can be absorbed with relatively thin absorptive panels, or redirected using flat reflective panels. THis will depend on the room response you are looking to achieve. Search "lede/rfz" or "non-environmental" to get a feel for the two "major" control room models and their pros/cons. Based on the direction of a room model choice along with a few other deciding factors (live room size) will determine exactly how you want to "sculpt" the response in or around the first 20msec, as well as what should happen after that 20msec (or so).
Old 23rd December 2010
  #5
Lives for gear
 
johndykstra's Avatar
 

That was, in a nutshell, a regurgitation of what SAC posted on this thread:

Envelope Time Curve - ETC - Impulse

but a little less wordy, and in a larger font. For an in depth understanding of the process, reading SAC's version will certainly provide a LOT more useful info.

Think of my post above as a "noob primer"
Old 23rd December 2010
  #6
Lives for gear
 
johndykstra's Avatar
 

Quote:
Originally Posted by Azure View Post
This makes absolute sense, but one of the issues I always have when dealing with a new room and my monitor setup is how to place a subwoofer and then adjust it's level and crossover point. My fear is that in this case, when testing, I will introduce so many variables using my current subbed setup, that I will get lost in the adjustment "abyss" and never come out heh. I guess was wondering if there was a more generic way of approaching it?
another great read pertaining to this subwoofer subject:

https://www.gearslutz.com/board/bass-...ng-thread.html
Old 23rd December 2010
  #7
Gear Nut
 

Okay, thanks for the clarification, and I appreciate the both you taking the time on my behalf! I have a pair of Mackie HR824's in storage and will dig them out and use them for simplicity's sake (I see that Ethan has used these speakers in his testing, so I'm assuming they will work okay). FM in hand, off I go...
Old 24th December 2010
  #8
Lives for gear
 
SörenHjalmarsson's Avatar
 

Quote:
Originally Posted by Azure View Post
This makes absolute sense, but one of the issues I always have when dealing with a new room and my monitor setup is how to place a subwoofer and then adjust it's level and crossover point. My fear is that in this case, when testing, I will introduce so many variables using my current subbed setup, that I will get lost in the adjustment "abyss" and never come out heh. I guess was wondering if there was a more generic way of approaching it?
Different dogs have different ways of barking, but the the most complete and systematic approach i've seen is that of Friedemann Tischmeyer in his Internal Mastering book/DVD. I have written it down and attached it, as i find it to be a handy 'step by step guide'. I've also attached two more PDF's.

Also check out this thread: Treating a small living room (5.55m\3.03m2.51m) (only post #9 to about post #18) Mr Brandt has a pretty interesting approach aswell, although he doesn't get into specifics.



Cheers,
Attached Files

Last edited by SörenHjalmarsson; 24th December 2010 at 09:11 PM.. Reason: Correcting/Updating PDF
Old 25th December 2010
  #9
Gear Nut
 

Quote:
Originally Posted by SörenHjalmarsson View Post
I have written it down and attached it, as i find it to be a handy 'step by step guide'. I've also attached two more PDF's.
Thanks for the material Sören - reading through it now.
Old 25th December 2010
  #10
SAC
Registered User
 

The ETC (Envelope Time Curve) response is derived from the impulse response and the doublet response.

Both the impulse response and the doublet response reflect complimentary aspects of the system's behavior with respect to time. The impulse response representing the real domain and the potential energy, and the doublet response representing the imaginary domain and the kinetic energy of the system. Related by the Hilbert transform, both are necessary to represent the total system.

The impulse response has no necessary/exclusive relationship to an 'impulse' or Dirac stimulus. I might suggest that part of the confusion lies in the assumption made in the question itself. One can measure the room to determine the characteristic impulse response, but that is not the same thing as "impulse testing", if by the term you are assuming some relation to how the test is made.

Don't let the word "impulse" confuse you - as the term is being used for multiple 'things'.

For general modal evaluation/stimulus of the room, a non-directional source is generally preferred. The common source for this is a 'dodec' - a dodecahedron - a 12 sided 'sphere' with drivers firing in 12 equally spaced directions - a comb filtering nightmare that sounds like @#$&, but which 'drives' the room just fine (Think a Bose direct-reflecting speaker on steroids - or what the Bose direct-reflecting system wanted to be when it grew up). This is a typical configuration used in larger spaces for RT60 measurements as well.

Speakers can be placed in corners as well, provided they are in the far field relative to the the measurement mic - optimally placed in the diagonally opposite corner - which, assuming this is concurrent with the longest dimension, will also provide information on the lowest frequency mode.

For this purpose you are evaluating the room's support of acoustic behavior essentially 'independent' of a source - to the greatest degree possible. You are measuring in the far field without the influence of near field direct signals.

Note, when you move to measuring the actual response of a topology relative to the listening position using the actual system (each response featuring one speaker driven at a time in order to avoid the destructive interference introduced by the superposition of spaced sources), your measurement setup and response may often be characterized by a combination of both the direct signal and of the room supported behavior. Thus, it becomes more difficult to ascertain what part of the response is due to the direct signal (something you do NOT want to be treating) and that portion reinforced by the room. (Its almost as if Heisenberg has taken a short side trip over to the macro world of classical acoustics....sorry... too much time spent in the quantum realm.... ;-)

I suspect this issue is a large contributor to many of the issues with which many are concerned.

There are quite a few measurements that are possible. There is no one size fits all solution to the question. The answer is conditioned by the nature of the information desired.
Old 27th December 2010
  #11
Gear Nut
 

Quote:
Originally Posted by SAC View Post
Don't let the word "impulse" confuse you - as it is used for multiple 'things'...

...Speakers can be placed in corners as well, provided they are in the far field relative to the the measurement mic - optimally placed in the diagonally opposite corner - which, assuming this is concurrent with the longest dimension, will also provide information on the lowest frequency mode.

...For this purpose you are evaluating the room's support of acoustic behavior essentially 'independent' of a source - to the greatest degree possible. You are measuring in the far field without the influence of near field direct signals.
Yes, thank you SAC (and johndykstra as well), this was the information I was looking for regarding determining the "generic" response of a room, independent of frequency response anomalies of a particular speaker or a speaker's placement.

I'm still trying to get my head around how and why an ETC measurement is more advantageous and different than an FR graph displayed over a given time interval (i.e. waterfall plot). And if I understand correctly, an ETC is merely an amplitude measurement for a given frequency (whatever frequency is generated in a particular test) which registers as amplitude "spikes" as it decays over time? And that by computing when these "spikes" occur, we can determine distance and begin to find reflection points that are problematic within a room? Pardon my ignorance, but is there a definitive source that explains this in terms that a layperson such as myself can understand?
Old 28th December 2010
  #12
SAC
Registered User
 

Quote:
Originally Posted by Azure View Post
I'm still trying to get my head around how and why an ETC measurement is more advantageous and different than an FR graph displayed over a given time interval (i.e. waterfall plot). And if I understand correctly, an ETC is merely an amplitude measurement for a given frequency (
No.

The ETC is not a measurement relative to a particular frequency It accounts for all energies at all frequencies.

The ETC is a Time domain measurement.

Forget frequency! Trying to view frequency with a time measurement is like trying to tell the temperature with a standard analog watch!

Literally, its time to move beyond this myopic frequency-centric point of view.

Generally speaking, The frequency domain is not causal. The time domain is. Within the behavioral realm which we aw dealing, the frequency response is derivative - it is determined by the interaction of energies distributed in time.

The frequency domain response that you are thinking in terms of is a result of many signals combining destructively/constructively with respect to time. The frequency response is simply a summary of all of such interactions. But you have NO information as to what the various contributory signals might have been, nor any information about their characteristics. And this information is necessary in order to make any fundamental changes to the frequency response of such a non-minimum phase signal.

The group of measurements grouped as 'impulse responses' (impulse, doublet, ETC) provide an accounting for all of the energy in the system with respect to time.

There is much more to this entire issue that is beyond the scope of this response. If you are desirous to pursue this, PM me and I will gladly address it interactively in as much detail as you desire.

But ultimately, the ETC provides an accounting of the energy response within a room at a particular point with respect to time. Specular energy arrivals are able to be identified with great acuity with respect to the gain and the arrival time.

This energy is also able to be identified in terms of their spatial characteristics within a space. Thus, knowing the gain, time arrival, and spatial dispersion characteristics, we are able to modify individual contributory/causal factors with respect to the other contributory factors.

And since we understand the interactive nature of the superposition (the fancy proper name for the interaction/combining of signals), we are able to determine the optimal manner in which to modify the various signals and their relationships with one another in order to effect the final summary result.

Hence, if you correct issues in the time domain, the resultant frequency response for the system 'magically' is resolved.

But again, within the framework of what we are dealing, we don't care about the frequency. It is what it is. the source generates energy at various frequencies. But it propagates as energy that arrives within time at various gain levels. we address this energy with respect to time. By minimizing the destructive iteration of the various energy 'packets' within time, we reduce the destructive interference, leaving the original energy with as little destructive influences as possible. thus, the only concern regarding frequency is one that is concerned with the creating of the signal prior to its propagation from tine speaker or source- not its reproduction once the energy is generated within a space..
Old 28th December 2010
  #13
Gear Nut
 

Quote:
Originally Posted by SAC View Post
Literally, its time to move beyond this myopic frequency-centric point of view.

...Generally speaking, The frequency domain is not causal. The time domain is. Within the behavioral realm which we aw dealing, the frequency response is derivative - it is determined by the interaction of energies distributed in time.

...The frequency domain response that you are thinking in terms of is a result of many signals combining destructively/constructively with respect to time. The frequency response is simply a summary of all of such interactions. But you have NO information as to what the various contributory signals might have been, nor any information about their characteristics. And this information is necessary in order to make any fundamental changes to the frequency response of such a non-minimum phase signal.

Hence, if you correct issues in the time domain, the resultant frequency response for the system 'magically' is resolved.

...But again, within the framework of what we are dealing, we don't care about the frequency. It is what it is. the source generates energy at various frequencies. But it propagates as energy that arrives within time at various gain levels. we address this energy with respect to time. By minimizing the destructive iteration of the various energy 'packets' within time, we reduce the destructive interference, leaving the original energy with as little destructive influences as possible. thus, the only concern regarding frequency is one that is concerned with the creating of the signal prior to its propagation from tine speaker or source- not its reproduction once the energy is generated within a space..
Wow SAC, brilliantly stated, and I think it's beginning to make sense. If I understand you correctly, much like the "waves in water" exercises we used to do in high school physics class to visualize how waveforms interact constructively and destructively to generate gain changes in the form of peaks and nulls when reflecting off boundaries, it's the same thing here. And it's purely a matter of how these "interferences" or interactions collect to build and diminish in certain areas of a room, amplifying or attenuating at specific frequencies, which determines the "response" of a room. That it's not a function of a generated frequency as the cause, but rather the causal effects of boundary reflection and the resulting interference with other colliding waves in the time domain having an effect on a given frequency.

So now, if I'm correct in visualizing this concept, I'll need to learn how to interpret and visualize what an ETC readout is actually telling me before I can even begin to apply it in some kind of useful manner.

SAC, thanks so much for taking the time, and for the offer of a PM'd lesson. I'm too intimidated to accept, my fear being that I'll be left in the dust from the word go; the information simply too difficult for me to grasp in this type of forum.
Old 28th December 2010
  #14
Gear Guru
Answer

Hi Azure. For recording Impulse Responses, the requirements of the speaker are pretty much the same as for studio monitoring.
i.e. Full range, flat, and high powered. The Mackie is fine but a little lacking in low bass. Length modes can be 30Hz or even lower in a bigger room.
Put it on the floor, even in the corner as john said. This will deliver louder and lower bass.

To get decent decay measurements in a typically noisy (fans computers) room you need to drive the speaker loudly. Be careful, wear headphones or earplugs, watch for clip lights on the powered speakers during the sweep, particularly the HF part.

Try to get an optimum Signal to Noise Ratio. Use speaker volume and mic pre gain to get that Freq graph up near 0dB, like in the old days of 16Bit recording.

All of the graphs are derived from this Impulse Response Recording.
Each has it own purpose and fortes.
e.g. Frequency Response is obvious, becomes a lot more useful if you take IR's at various spots in the listen zone and average them.
Our ears are in two spots.
Waterfalls are great for Mode spotting.
ETC is great for taking time for some Reflection.

And so on...

Anyone that says one graph is 'better' or 'best' is a communist.

DD
Old 28th December 2010
  #15
SAC
Registered User
 

Quote:
Originally Posted by DanDan View Post
Anyone that says one graph is 'better' or 'best' is a communist.
They may not be a communist, but they might be an acoustician. heh

What has been said is that in a space dominated by distributed energy arriving over time, one response displays causal information, while the other response is derivative. One shows in great detail what is happening, and the other simply shows the aftermath wherein something happened.

They each are 'better' for different tasks. As modal behavior is relatively simple in terms of its component nature, and as a result the FR/waterfall is fine (and preferred).

The most common problem on the forum (and in the larger world as well) with regards to the usage of the various responses is that too many still rely almost exclusively on the FR, a response with very limited practical use above the modal region, and where most folks have literally operated blindly in a very critical response region. Thus a 'rebalancing' regarding the usage of tools is beneficial.

Once modes are addressed, one can easily excel with respect to the specular region with literally never seeing the frequency response again.
Old 28th December 2010
  #16
Gear Guru
Commies

I hear Putin is designing some Traps.....
DD
Old 29th December 2010
  #17
Gear Nut
 

Thanks SAC, Dan Dan, Johndykstra, et al; I think I'm good so far...

Okay, so in approaching treating my new, hybrid tracking/monitoring room would the following be a proper methodology in trying to determine a listening position? For reference, the room is decently sized (about 700 sq. ft., 9.5' ceilings), but not ideal because it's oddly shaped and somewhat trapezoidal. Even though it would not be a classic LEDE configuration because of the irregularity of the room, the intent would be to treat RFZs relative to the listening position while keeping a degree of "life" and not making it NE or completely dead because I want to track in it as well. I understand the imaging problems that I will encounter while monitoring due to lack of symmetry, but I am hoping that by using suspended panels, gobos, and tube traps that I have around the listening position, that I can mitigate these effects somewhat. (And sorry if in my noobism I butcher or use the wrong terminology) :

1) Test the raw room "generically" without treatment by doing multiple LF sweeps with a monitor firing into a corner to see what inherent modal problems exist and to determine where in the room might be the least problematic for speaker placement. (I understand that the 38% rule doesn't apply here because of the shape, so it's a bit of a crap shoot.)

2) After finding a best case position, treat the room with superchunk in corners and as much other LF absorbent treatment (i.e. 4+" spaced wall panels with air gaps) as possible. Sweep again to determine the resulting LF improvement, adjusting the proposed listening position if necessary.

3) This next step is probably out of my financial capabilities, but treat any remaining problematic LFs with appropriately designed tuned traps specific to those frequencies.

3) Once the "best case" position is found, place the monitor(s) in the proposed listening position and begin doing IR / ETC testing for the higher frequencies. Determine initially the obvious first reflection points by using the string or mirror methods, treating these zones with absorbent material. These would include side panels and a ceiling cloud.

4) Next step would be to determine the location of secondary reflection ETC "spikes" and back wall reflections. But rather than treating these points with absorption, I would use diffusion such as QRDs, Skylines, or Polys (or even angled reflective panels for scattering) to break up the waves emanating from these points.

5) Repeat ad nauseam. Crack a beer and enjoy.

Does this make sense, and am I on the right track?
Old 29th December 2010
  #18
Lives for gear
 
johndykstra's Avatar
 

Azure,

I'll leave it to SAC for the official thumbs up, but yes, in general you are heading down the right path. Absorptive reflection panels around the listening position may make it difficult to achieve a strong ISD termination (end of reflections of listening zone in regards to time... again about 20msec), and perhaps more importantly, a dense diffuse "afterbloom".

However, if finances are a concern, it's likely you will start with absorption around the mix spot, and only after some $ rolls in, reface those absorbers with geometrically angled specular panels, and only then worry about diffusion. This will allow you to preserve room energy to later arrive diffused to the mix zone.
Old 29th December 2010
  #19
Gear Guru
Afterbloom

Quote:
Does this make sense, and am I on the right track?
Yes, particularly the hotchpotch of hanging and standing stuff around the mix area.

Seems like quite a big room. I doubt if there is any possibility of 'dead'.
In fact you may have genuine reverb in there.

DD
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
peabody / So much gear, so little time
7
Static Glare / Electronic Music Instruments and Electronic Music Production
3
OCDaveWilcox / Electronic Music Instruments and Electronic Music Production
17

Forum Jump
Forum Jump