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
Envelope Time Curve - ETC - Impulse Dynamics Plugins
Old 11th January 2011
  #61
Gear Guru
OK

[joke that lead to off-topic sequence below removed by mod.]

I have no idea what they are talking about either, but it appears to have not gone well and stopped the thread. That seems a pity given the very direct illuminating answers earlier.

As far as I can see the most significant topic here is still the original question and later additions to it. i.e. How do I use ETC?
I don't believe it is necessary to be Enzio Ferrari to drive to the off licence.

In an effort to reboot, may I point out that FMPro3 has a Third Octave Decomposition. This does deliver a different ETC graph for each third octave.
With a little work this could actually deliver a third octave level for each reflection. Thus illustrating if your RFZ panels are absorbing throughout the full range adequately. If will check if REW has the same type of thing and come back to edit. Unless JohnPM wishes to speak on that?

Anyhoo, instead of a thousand words.
Envelope Time Curve - ETC - Impulse-screen-shot-2011-01-10-22.27.31.jpg

DD
Old 12th January 2011
  #62
Gear Maniac
 

Quote:
Originally Posted by DanDan View Post
I will check if REW has the same type of thing and come back to edit. Unless JohnPM wishes to speak on that?
In REW that is on the "Filtered IR" graph, octave and 1/3 octave filters.
Old 12th January 2011
  #63
Gear Guru
Useful

Thank you John, that saves me going into the studio to fire up REW.

So, open question. I will welcome any input to turn the following notion into a practical method.

Point a loudspeaker at a hard massive wall. This could even be outdoors to avoid room problems.
Sweep. Window the IR to lose the speaker. Use other methods to further minimise distraction, directional speaker, speaker position to avoid mode, suggestions welcome here.
We now have a clear isolated full range reflection of that wall.
Introduce a trap.
Sweep again, the reflection should be diminished by the trap.
The spectrum of the trapped reflection vs the bare wall reflection will hopefully deliver a spectrum of the action of the trap.
This would allow comparison of actual traps in a real situation.

Not looking for 'proof' level here just what may turn out to be a very useful practice.

EDIT, PM's with Andreas in the background. hmm, seems like I can't have the cake and eat it.
I want filters without ringing! We seem to have them in the audio world. The nature of the filter here makes my simplistic suggestion impossible. Unless...gate/window out the direct speaker sound. Evaluate/integrate the area under the curve, i.e. the total energy, of each third oct ETC for a defined period or window.
John, Chris, would you consider including such a Leq or 'dose' type of summing?

DD
Old 12th January 2011
  #64
Lives for gear
 
Dange's Avatar
 

There will be a low frequency limit to your measurements DD, but I believe what you propose is possible with TDS or MLS Quasi-Anechoic measurements

Audio Engineering Society UK » ‘Can we make quasi-anechoic measurements in normal rooms?’

However what about other reflections from other walls/surfaces if you were doing it in a room? How do you discriminate?

You can use an ETC curve to see if your absorbers are effective on early reflections without trucating the impulse response/ETC, i.e. how many dB's it knocks off the reflection.

Also I believe the ETC is only an estimation of magnitudes so using it to rate the effectiveness of an absorber might be difficult too. (the ETC is derived from an inverse FFT of the frequency response)

I think all you can gauge is whether an absorber has worked or not
Old 12th January 2011
  #65
Gear Guru
Wishes

Thanks Dange, interesting link.

Quote:
However what about other reflections from other walls/surfaces if you were doing it in a room? How do you discriminate?
Placement of speaker and mic to avoid or provoke modes, shielding panels, I am hoping for others to join in here.

Quote:
You can use an ETC curve to see if your absorbers are effective on early reflections without trucating the impulse response/ETC, i.e. how many dB's it knocks off the reflection.
Quote:
For sure. I am sure you can see that I am trying to extract a spectrum of absorbtion out of this. Even if crude, I would find this most useful.
e.g. remove the direct and see what the trap is doing within small frequency bands. The current graphs have very long timelines at LF, making it impossible to see how a paricular reflection is being hit at a particular frequency zone.
I hope John and perhaps Chris Liscio are watching-
What would happen if we restrict the Sine Sweep to say 70-90 Hz only?
Would an unfiltered ETC graph give an indication of the action around 80 or is this just mad heh


Quote:
Also I believe the ETC is only an estimation of magnitudes so using it to rate the effectiveness of an absorber might be difficult too.
Is it only an estimation or can we take the dB's of an individual (not filtered) spike literally?

DD
Old 17th January 2011
  #66
Lives for gear
 
Hannes_F's Avatar
 

EDIT:

Regarding the original topic I will try to find time for an understandeable but true explanation (which is is not an easy combination) that hopefully can put some light into it (into the part that I understood as there is always more).
Old 12th February 2011
  #67
Lives for gear
 
Hannes_F's Avatar
 

This thread should be a sticky
Old 12th February 2011
  #68
SAC
Registered User
 

Quote:
Originally Posted by Dange View Post
Also I believe the ETC is only an estimation of magnitudes so using it to rate the effectiveness of an absorber might be difficult too. (the ETC is derived from an inverse FFT of the frequency response)

I think all you can gauge is whether an absorber has worked or not
No.
The ETC shows MUCH more than that.

What you describe describes the log squared response of the impulse response. The impulse response displays only a portion of the total system behavior.


The ETC summarizes and includes MUCH more information.

It 'summarizes' the information contained in both the impulse and the doublet response.

Put simply, these are two views of the Analytic (see Dennis Gabor's "Theory of Communication", the mathematician who invented the hologram and was responsible for Nobel Prize holography research...), a rotating phasor oriented about the Z-axis representing, in this case, time.

The Analytic of which these two responses are simply projections onto the X (real) and Y (imaginary) planes respectively provides an accounting for both the magnitude and phase of the Analytic which encompasses the total system energy behavior with respect to time.

Frequency plays not direct part in this representation, although the time domain Analytic and the frequency domain Analytic are mapped to one another by the Fourier Transform.

The impulse response displays the potential ('stored') energy in the system represented in the real domain while the doublet response displays the kinetic ('dynamic') energy represented in the imaginary domain.

The ETC, by use of the Hilbert transform is able to 'extract' and 'combine' the data in both perspectives of the total Analytic in usable form and provide a complete accounting of the total energy signature (both potential and kinetic energy in the system) with respect to time.
Old 16th February 2011
  #69
Lives for gear
 
Dange's Avatar
 

So this means the magnitude is accurate?

My comments about the magnitude accuracy were referenced from a manual for hardware made by Audio Precision, the AP2700.

Quote:
When Energy-Time is selected as a Time Domain Display, the MLS Analyzer
processes the measurement result somewhat differently before display.
The impulse response is transformed into the frequency domain where further
processing is performed (including the application of a window, if selected,
and a Hilbert Transform). This result is then inverse transformed back to the
time domain for display. This result is commonly called the energy-time curve,
or ETC.
The energy-time curve is useful for determining arrival times and relative
energy distribution in time. It may be considered as similar to the envelope of
the impulse response, tracing along the top of the impulse response curve but
not showing any of the negative excursions.
However, the ETC is only an approximation of the actual energy arriving at
the microphone, since this energy can never be known without the simultaneous
measurement of both the velocity and the pressure of the sound; the
term energy-time response is actually a misnomer. A more accurate term sometimes
found in technical papers is the analytic signal magnitude, but as energy-
time has become common usage that term is used here.

ETC measurements (especially with the use of the Hann window; see below)
can be misleading. The Energy-Time choice is included as an MLS Analyzer
selection to accommodate comparison of current tests to legacy
techniques.
It was the "the ETC is only an approximation of the actual energy arriving at the microphone" bit that I might have misunderstood.

Anyway if it is accurate I can see DD's thinking to work out a rating for an absorber, a guide to the dB reduction it gives in the specular frequency region.
Old 16th February 2011
  #70
SAC
Registered User
 

The "approximation" aspect is due to their applying the Heisenberg uncertainty principle.

I wouldn't worry to much about that.

The bigger issue to which they refer is the difference between the more limited and incomplete log squared response of the impulse response that only accounts for the real/potential energy in the system (the impulse results squared - making them all positive), and the envelope time curve (ETC - note the more accurate modified name) that also accounts for the imaginary/kinetic energy of the doublet response of the total system (the Analytic) as well.

While on one level I want to rebel against the overly simplistic use of the term "magnitude" with regards to the accounting for energy within the modulation envelope of the Analytic... since the results are displayed in terms of gain versus time, you would be safe referring to the results as the magnitude (gain) of the energy with respect to time in the system.
Old 18th February 2011
  #71
Gear Maniac
 

Quote:
Originally Posted by Dange View Post
It was the "the ETC is only an approximation of the actual energy arriving at the microphone" bit that I might have misunderstood.
The statement from the AP manual is correct, ETC is only an approximation to the true energy profile of a system - it is acausal for one thing, though the extent to which that occurs is strongly dependent on the method used to compute the Discrete Hilbert Transform of the impulse response from which the analytic signal is generated (the ETC is the magnitude of that analytic signal). Nonetheless the overall shape and levels of an ETC are not far from reality. You can find an analysis of the differences between ETC and actual energy profile in an AES paper by Don Keele from 1992, entry number 25 in the list here: AES Papers -- Official website of D.B.Keele
Old 12th June 2011
  #72
Lives for gear
 
takman's Avatar
 

ohh boy....this thread gets really confusing...

So what is the difference between an impulse response and ETC? can someone explain it somewhat simply?

in REW you have a section for Impulse. then at the bottom, you have Envelope ETC listed with a check box next to it.

When checking and un-checking the ETC box, there doesnt seem to be a difference between the impulse and ETC readings....
Old 12th June 2011
  #73
Gear Maniac
 

The ETC is the envelope of the signal, which means it follows the overall shape of the signal rather than its individual oscillations. This image, from wikipedia, shows the difference between a signal (in blue) and its envelope (in red).
Envelope Time Curve - ETC - Impulse-envelope.jpg
REW shows the impulse response in either a normal linear amplitude display (when the y axis is set to % FS) or a logarithmic amplitude display (when the y axis is set to dB FS). Because the logarithmic display is generated from the squared amplitude of the impulse response it does not have any changes in sign, so the signal and its envelope appear quite similar. If you zoom in you will see there are differences, but they are much less evident than when using the % FS display.

Last edited by JohnPM; 12th June 2011 at 05:20 PM.. Reason: Remove ambiguity in use of "impulse"
Old 12th June 2011
  #74
Lives for gear
 
johndykstra's Avatar
 

Impulse is one stimuli (sound) you can use to generate an ETC response graph.


"Impulse response" is something different, where you take an impulse of a room, and as opposed to reading as an ETC graph, you feed it into a "reverb creation" program.


[edit] I think I may be wrong here {edit}

Last edited by johndykstra; 12th June 2011 at 04:41 PM.. Reason: or no?
Old 13th June 2011
  #75
Lives for gear
 
takman's Avatar
 

john, quick question...

why does the impulse/ETC graph show the arrival time of the direct signal to be 0?

is this why in the controls section theres an option "t=0 offset(ms)? So i can offset the graph in ms to account for the direct signal delay?


for example.....say if i measured the distance from my speaker to the mic....do the math, and find that the direct signal is arriving to the mic within 2ms....then i offset the time on the graph by 2ms?
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
kafka / So many guitars, so little time
9
Static Glare / Electronic Music Instruments and Electronic Music Production
3
chuck78 / Electronic Music Instruments and Electronic Music Production
2
The Magic Hoof / Electronic Music Instruments and Electronic Music Production
7
jlsgear / Music Computers
5

Forum Jump
Forum Jump