Simplified REW Setup and Use (USB Mic & HDMI Connection) Including Measurement Techniques and How To Interpret Graphs - Page 106 - AVS Forum
Forum Jump: 
 267Likes
Reply
 
Thread Tools
post #3151 of 12760 Old 05-30-2013, 01:09 AM - Thread Starter
Senior Member
 
jevansoh's Avatar
 
Join Date: Oct 2010
Location: Southern, Ohio
Posts: 471
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 53
Quote:
Originally Posted by AustinJerry View Post

OK, time for a bit of instruction for us neophytes: what does the term "terminator" mean?

To terminate the ISD Gap which is the biggest step in creating a LEDE model.

Once you tame the high gain early reflections to about 20ms and about -20db you then need to terminate that ISD gap (the gap is the time after the direct signal out to the 20ms or so goal you've set where the high gain reflections are tamed to below -20db from the direct signal) by returning the energy with a HIGH GAIN reflection at as close to (or more) -12db followed by an exponentially decaying sound field of lateral / diffuse reflections.

So it should look like this... First spike on the ETC is the loudest. This is the direct sound. This is to be followed by no reflection higher than -20db up to about 20ms or so. At 20ms or so you will see a reflection as close to around -12db as possible followed by DENSE reflections that decay smoothly over time into the noise floor.

That's the basis of LEDE.

A lot of people get rid of the early reflections before 20ms but then never "terminate" their ISD gap by reintroducing the energy back into the rooms via a HAAS Kicker or laterally arriving dense reflections and a lot of times I'll see sparse reflections in folks' ETC measurements past 20ms (just a high gain reflection [above -20db] here and there) which only serves to confuse things instead of the proper method listed above which will allow ALL the ambiance and information in the original recording to come through and make your small acoustical space sound much larger than it is.

If you do NOT terminate the ISD Gap by reintroducing laterally arriving reflections past 20ms or so yet the reflections continue to decay and there are no sparse high gain reflections past 20ms then it starts to become more of an NE (Non-Environment) design which is basically effectively anechoic.

It's much easier to do a NE design as there is a finite amount of energy in a room. It is precious. If you absorb it all to get the reflections down to -20db before 20ms then where does the energy come from to terminate the ISD gap (introduce reflections at a high gain at 20ms or so to then decay logarithmically to the noise floor)??

BTW, it "IS" important to use the ETC properly. To do that, you MUST band filter (using the IR window and selecting 1/3rd octave filtering starting at 500hz up to at least 8khz or so) the ETC measurements, in addition to looking at the full unfiltered ETC for each individual channel to get an accurate picture and verify you are not just EQ'ing your response by not using broadband absorption (IE: Only taking care of reflections from 1000hz to 20khz by using too thin of an absorber).

If you use this method you will be using the ETC to the fullest and it will never let you down. smile.gif There is no better tool to use, no more comprehensive or accurate tool in the cookbook either.

You just have to know what it's for and how to use it.

Or..... You could click the RT60 tab, look at it for 5 seconds, do nothing with the data (which doesn't mean anything anyway) and be done. wink.gif

Hope this helps.

BTW - A lot of good posts over the last 24hrs or so... I will be replying to many posts and answering specific questions as time allows tomorrow. I get off work early and should be able to devote a good amount of time to this thread tomorrow as I did yesterday. Sorry for the delayed responses. tongue.gif

--J
BarracudaDelGato likes this.

Thanks to EVERYONE that Helps Make These Threads so Awesome!

jevansoh is offline  
Sponsored Links
Advertisement
 
post #3152 of 12760 Old 05-30-2013, 01:54 AM
AVS Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 5,545
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 973 Post(s)
Liked: 379
Quote:
Originally Posted by amirm View Post

I am sorry but I am not clear about your question. Are you asking us to dismiss a metric because it has a range? If my car engine manufacturer says the idle speed for it is 700 to 800 what should do? Ignore it and let the idle speed be 2000 or 500?

That's not the point. RT doesn't tell you anything more than a look at an ETC would. An ETC shows more information than RT. That's why I don't look at RT at all. Explore the "Filtered IR" tab in REW. Tells you a ton more about the sound field in a room than RT.
By the way, if you need to see RT then it's a single click away in REW.
Quote:
Originally Posted by amirm View Post

We need a measure of late reverberation since it determines whether we are impeding speech intelligibility.

Then simply look at an ETC or a spectrogram or... it's all just a transform of the impulse response. Of course one has to understand what an impulse response represents and what the Fourier transform is.
Quote:
Originally Posted by amirm View Post

Or we have put so much absorption there as to make the room uncomfortable, i.e. "dead." You saw how quickly using RT60 we determined two of the rooms which were posted were in the target range and hence fine.

If you believe in single scalar values. I don't.
Quote:
Originally Posted by amirm View Post

For the single mouse click it took to get that data and make the determination, the ROI is superb. On the other hand, folks still struggle what to do with ETC. If we are going to throw one overboard, I vote for ETC. biggrin.gifbiggrin.gifbiggrin.gif

So ETC is useless because people "struggle with ETC"? Nobody every said that analysis of an LTI system would be easy.
Quote:
Originally Posted by amirm View Post

I provided their names already: http://www.avsforum.com/t/1449924/simplified-rew-setup-and-use-usb-mic-hdmi-connection-including-measurement-techniques-and-how-to-interpret-graphs/3090#post_23368089. It included Dr. Toole, Grimoni, Erskine, Nyal and Jeff. Do you need more? Here is Keith Yates:
" Ideally, reverberation will ramp down to -60dB in approximately 500ms (RT60=0.5s)."

Room acoustics and speech perception, Arthur Boothroyd, Ph.D.Distinguished Professor Emeritus, City University of New York, Scholar in Residence, San Diego State University, Visiting Scientist, House Ear Institute:
"The time taken for the level to fall by 60 dB is known as the reverberation time (RT60). This quantity provides a rough measure of the reverberant properties of a room. Reverberation times in large, reflective spaces such as gymnasia can be as high as 2 or 3 seconds. In small classrooms with many absorbent surfaces (including the surfaces of the students), reverberation times may as low as 0.3 or 0.4 seconds."

Famous Siegfried Linkwitz and his AES paper, Room Reflections Misunderstood?
"The listening tests took place in the authors living room. It is a large room that is enclosed on three sides and extends towards a kitchen area and a hallway in the back. (Figure 3). No special acoustic treatment has been used and the listening area is acoustically fairly live with RT60 around 500 ms."

Dr. Sean Olive AES paper, A Multiple Regression Model for Predicting Loudspeaker Preference Using Objective Measurements: Part II - Development of the Model
"The acoustical properties of the listening room used in the development of this model are not unlike those of many professional and domestic listening rooms. It meets current requirements of ITU-R BS 1116 and its reverberation time (RT60 = 0.3 s.) falls close to the average value of 0.4 s measured in 603 domestic rooms by Bradley [28]. On this basis, it is more than likely our model can be generalized to many typical rooms."

From Todd Welti's famous AES paper, How Many Subwoofers are Enough:
"The room ,as set up for these tests, was fairly live (500 Hz RT60 = .85)."

OK, my fingers are getting tired so I stop here. smile.gif So the answer to your question is "tons of luminaries who write papers and do this for a living." Hopefully your list is just as long and impressive smile.gif.

What is the point you're trying to make? Nothing I've said negates anything in that quotes. What makes you think I'm not familiar with the literature?
Quote:
Originally Posted by amirm View Post

Oh, I apologize for having said anything disrespectful.

Punching someone in the face followed by an apology doesn't undo the punch. If you continue that behavior this is my last reply to you.
Quote:
Originally Posted by amirm View Post

As i said, the words Dr. Toole has for people who trust ETC is quite harsh. In person, he is even less tolerant. It takes a lot to have a person with his calm demeanor to ratchet up his emotions. But using a tool without knowing the basis of how it works and how it can misfire will certainly get him there. And I guess me too to some extent, causing me to write these missives. smile.gif Thanks for taking the time to read my posts and asking questions. Talk to you later.

What makes you think that I wouldn't know "how it works and how it can misfire"? That's ignorant at best.

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is offline  
post #3153 of 12760 Old 05-30-2013, 02:50 AM
AVS Addicted Member
 
sdurani's Avatar
 
Join Date: Oct 2001
Location: Monterey Park, CA
Posts: 19,949
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 1485 Post(s)
Liked: 1041
Quote:
Originally Posted by Roger Dressler View Post

And yes the center channel control can raise the dialog in relation to all the rest of the track.
I was talking about the receiver's volume control, which raises the entire track. But yes, turning up just the centre speaker can make the dialogue louder, and make everything else in the centre channel louder, and throw the front soundstage out of balance. Better to treat the room for better intelligibility (if you believe treating the room can have that effect).
Quote:
Originally Posted by Roger Dressler View Post

But none of that really settles the question as to whether early reflections improves intelligibility in a home theater.
I don't see why they would stop having that effect in a home theatre. They're either having a positive, negative or neutral effect on intelligibility. Do you believe it is negative or neutral?

Sanjay
sdurani is offline  
post #3154 of 12760 Old 05-30-2013, 03:07 AM
AVS Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 5,545
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 973 Post(s)
Liked: 379
A little bit of control room design history:
http://www.rpginc.com/docs%5CTechnology%5CPresentations%5CStudio%20Design%20From%20Mono2Surround.pdf

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is offline  
post #3155 of 12760 Old 05-30-2013, 05:17 AM
AVS Addicted Member
 
kbarnes701's Avatar
 
Join Date: Jan 2008
Location: Main Listening Positon
Posts: 19,579
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 2463 Post(s)
Liked: 2211
Quote:
Originally Posted by AustinJerry View Post

Jim, I agree that what's best for one is not necessarily best for the other.  Furthermore, at least from my perspective, getting things right for music reproduction is considerably more challenging than getting things right for movies.  I can become totally engaged in a movie without paying that much attention to whether the sound reproduction is perfect or not (and I'm not sure I would recognize perfection if I was hearing it).  However, if there is something wrong with the audio when playing music, the entire mood is gone.  IOW, the audio is close to 100% of the experience for me.

 

 

I concur, Jerry, Movies, of course, have the massive distraction of video, not to mention all the other stuff going on. How many times do you watch a movie and see the music credits roll past at the end and say to yourself "I don't remember hearing <insert name of popular and well-known song> in that movie". This happens fairly often with me, and I concentrate hard on the movie when it is playing. With music, of course, there is just audio. So if the audio is poor, it becomes really really apparent and, as you say, spoils the experience. While the audio is 50% of the movie IMO, there is much to enjoy in a movie even if watching it with TV speakers <shudder>. 

 

Because I am a sort of movie 'freak', I often watch a good movie several times, concentrating on a different aspect at each separate viewing. First pass is to get the gist of the plot, the character arcs etc and a general impression. Subsequent passes may focus on cinematography, the score, the editing, production design, set design, costume, art direction and so on. But I would agree that this is perhaps "unusual".  I also love the discs I have that have an isolated score. I have a few of those of Hitchcock movies, scored by the wonderful Bernard Herrmann, and it is amazing to watch the movie, devoid of dialogue and effects and all other sounds, just concentrating on the score with no other distractions. If you know the movie well, eg Vertigo which I must have seen 20 times, the lack of dialogue etc is entirely without importance and just having the score really shows up the incredible contribution it can make to the movie. It truly expresses the emotion that the dialogue etc does not. Or enhances it let's say. But I am rambling... :) 

 

 

Quote:

I also recognize that others, especially Keith, value the audio quality of their movies, and there is nothing wrong with that.  That's why we have different models to work towards.

 

 

Indeed, and concurred again. 

kbarnes701 is offline  
post #3156 of 12760 Old 05-30-2013, 06:43 AM
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 7,983
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 1097 Post(s)
Liked: 914

A true aficionado, Keith.  I respect and admire that!

AustinJerry is online now  
post #3157 of 12760 Old 05-30-2013, 06:48 AM
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 7,983
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 1097 Post(s)
Liked: 914
Quote:
Originally Posted by jevansoh View Post


To terminate the ISD Gap which is the biggest step in creating a LEDE model.

 

 

Thanks, J.  It's now clear what the terminator means.  Going forward, I will be looking forward to discussions on techniques on how to make it happen.  Looking at Jim's listening room, I have a pretty good idea.  wink.gif

 

And posts like this provide the type of information that will be useful to capture in the consolidated "tips" that I will be collecting.

AustinJerry is online now  
post #3158 of 12760 Old 05-30-2013, 06:50 AM
AVS Addicted Member
 
kbarnes701's Avatar
 
Join Date: Jan 2008
Location: Main Listening Positon
Posts: 19,579
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 2463 Post(s)
Liked: 2211
Quote:
Originally Posted by AustinJerry View Post

A true aficionado, Keith.  I respect and admire that!

 

Thanks Jerry. Aficionado is so much nicer than the word that many of my friends use to describe my passion for all things movie wink.gif

kbarnes701 is offline  
post #3159 of 12760 Old 05-30-2013, 06:54 AM
AVS Addicted Member
 
kbarnes701's Avatar
 
Join Date: Jan 2008
Location: Main Listening Positon
Posts: 19,579
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 2463 Post(s)
Liked: 2211
Quote:
Originally Posted by jevansoh View Post

 
BTW, it "IS" important to use the ETC properly. To do that, you MUST band filter (using the IR window and selecting 1/3rd octave filtering starting at 500hz up to at least 8khz or so) the ETC measurements, in addition to looking at the full unfiltered ETC for each individual channel to get an accurate picture and verify you are not just EQ'ing your response by not using broadband absorption (IE: Only taking care of reflections from 1000hz to 20khz by using too thin of an absorber).

If you use this method you will be using the ETC to the fullest and it will never let you down. smile.gif There is no better tool to use, no more comprehensive or accurate tool in the cookbook either.
 

 

J (or anyone) - can you elaborate for me please? When I open the IR Window, I see this, which doesn't gell with your remark bolded above:

 

Where are the controls you mention?  Thanks.

kbarnes701 is offline  
post #3160 of 12760 Old 05-30-2013, 06:58 AM
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 7,983
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 1097 Post(s)
Liked: 914
Quote:
Originally Posted by AustinJerry View Post

 

Local, this is the second time you have advised me regarding the "Use Loopback as Timing Reference" setting for the impulse response measurements, and I appreciate your advice.  For others who are not familiar with where this setting is, open the Analysis tab on the Preferences screen:

 

 

 

I am familiar with how to configure the hardware loopback when using the "legacy" REW gear, but am not sure whether it is even possible when using the USB Mic + HDMI connection to the AVR.  Can you (or anyone else who knows the answer) comment on whether this is possible or not?  If not, does this mean that if we are using the HDMI connection, the impulse measurement is wrong, or meaningless?

 

I'm going to ask this question over on HTS as well.

 

I didn't receive any feedback WRT the accuracy of ETC measurements without the hardware loopback enabled, and how the loopback could be accomplished using HDMI connections.  A similar posting over on HTS received one reply, pointing me to a discussion on how to enable the loopback.  I read it, but didn't understand it completely.

 

http://www.hometheatershack.com/forums/rew-forum/67726-how-configure-hardware-loopback-when-using-asio-hdmi.html#post615095

AustinJerry is online now  
post #3161 of 12760 Old 05-30-2013, 07:02 AM
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 7,983
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 1097 Post(s)
Liked: 914
Quote:
Originally Posted by kbarnes701 View Post

 

J (or anyone) - can you elaborate for me please? When I open the IR Window, I see this, which doesn't gell with your remark bolded above:

 

Where are the controls you mention?  Thanks.

 

Not J, but I believe he is referring to the "Filtered IR" tab, which has a control at the bottom left of the measurement panel allowing you to select various filters, at either 1/1 or 1/3 octave.

AustinJerry is online now  
post #3162 of 12760 Old 05-30-2013, 07:07 AM
AVS Addicted Member
 
kbarnes701's Avatar
 
Join Date: Jan 2008
Location: Main Listening Positon
Posts: 19,579
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 2463 Post(s)
Liked: 2211
Quote:
Originally Posted by AustinJerry View Post

Quote:
Originally Posted by kbarnes701 View Post

 

J (or anyone) - can you elaborate for me please? When I open the IR Window, I see this, which doesn't gell with your remark bolded above:

 

Where are the controls you mention?  Thanks.

 

Not J, but I believe he is referring to the "Filtered IR" tab, which has a control at the bottom left of the measurement panel allowing you to select various filters, at either 1/1 or 1/3 octave.

 

Thanks Jerry. I will go and take a look.

kbarnes701 is offline  
post #3163 of 12760 Old 05-30-2013, 07:11 AM
AVS Special Member
 
localhost127's Avatar
 
Join Date: May 2009
Posts: 2,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 8 Post(s)
Liked: 40
Quote:
Originally Posted by markus767 View Post

So ETC is useless because people "struggle with ETC"?

in an earlier thread, the salesman "struggled" with understanding just what the x-axis represented within the ETC (even while furiously debating against it like an "expert"). he literally did not understand the x-axis represents distance traveled of the indirect energy - and how this "perspective" can be used to place treatments or achieve a desired response - based on the user's own design requirements.


the frequency-response is useless because i struggle with the frequency-response.

how on earth do i derive information regarding accurate localization, imaging, identifying sources of edge diffraction, identifying whether individual components are out of alignment, identifying sources of flutter-echo, identifying where to place diffusers/absorbers/etc to achieve my desired response, identifying how the energy decays in my room, etc - from a 20hz-20khz frequency-response?

can someone please do a tutorial regarding these scenarios so i can stop struggling with the frequency-response?

maybe the salesman should inform Keith Yates that he has been doing it all wrong for all of these years. nyal mellor too. peter d'antonio too. and every other acoustician involved with studio/room design. have you ever seen a user so overly scared and utterly terrified of a different perspective regarding acoustical analysis? funny - it was TDS/time-domain analysis that resulted in the quantum leap in understanding of acoustics many decades ago via that of the work from the acoustical giants of whose shoulders we stand on (not salesmen's shoulders). where was he then to cry to Richard Heyser that he was doing it all wrong?

sonar and radar engineers are laughing their heads off.
localhost127 is offline  
post #3164 of 12760 Old 05-30-2013, 07:23 AM
AVS Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 5,545
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 973 Post(s)
Liked: 379
Quote:
Originally Posted by AustinJerry View Post

I didn't receive any feedback WRT the accuracy of ETC measurements without the hardware loopback enabled

The ETC is a transform of an impulse response. You can obtain a perfectly fine impulse response without loopback hence no loopback is required in order to calculate a valid ETC.

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is offline  
post #3165 of 12760 Old 05-30-2013, 07:38 AM
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 7,983
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 1097 Post(s)
Liked: 914

The vertical axis should extend from 0dB to the noise floor, at least -40dB.  I set it at -60dB.  I set the horizontal axis right limit to 40ms.  The first 20ms is critical, and you should see most (if not all) peaks below -20dB beyond 20ms.  This should be especially true in your well-treated room.  I set the left limit for the horizontal axis at -1ms, which allows the initial impulse at T=0 to be more easily "grabbed" using the CTRL-Right Click technique to display peak distances.

 

AustinJerry is online now  
post #3166 of 12760 Old 05-30-2013, 07:39 AM
AVS Special Member
 
localhost127's Avatar
 
Join Date: May 2009
Posts: 2,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 8 Post(s)
Liked: 40
Quote:
Originally Posted by markus767 View Post

The ETC is a transform of an impulse response. You can obtain a perfectly fine impulse response without loopback hence no loopback is required in order to calculate a valid ETC.

the context was with respect to identifying total flight path of the indirect energy.
with the direct signal at T=0 you do not have accurate total flight paths as propagation delay is not accounted for.
localhost127 is offline  
post #3167 of 12760 Old 05-30-2013, 07:42 AM
AVS Addicted Member
 
kbarnes701's Avatar
 
Join Date: Jan 2008
Location: Main Listening Positon
Posts: 19,579
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 2463 Post(s)
Liked: 2211
Quote:
Originally Posted by AustinJerry View Post

The vertical axis should extend from 0dB to the noise floor, at least -40dB.  I set it at -60dB.  I set the horizontal axis right limit to 40ms.  The first 20ms is critical, and you should see most (if not all) peaks below -20dB beyond 20ms.  This should be especially true in your well-treated room.  I set the left limit for the horizontal axis at -1ms, which allows the initial impulse at T=0 to be more easily "grabbed" using the CTRL-Right Click technique to display peak distances.

 

 

Thanks Jerry. Appreciated. I am at least -25dB down beyond 20ms.

kbarnes701 is offline  
post #3168 of 12760 Old 05-30-2013, 07:43 AM
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 7,983
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 1097 Post(s)
Liked: 914
Quote:
Originally Posted by localhost127 View Post


the context was with respect to identifying total flight path of the indirect energy.
with the direct signal at T=0 you do not have accurate total flight paths as propagation delay is not accounted for.

 

Well, until I completely understand what is going on here, I will be using my "legacy" REW kit, which easily allows me to configure the hardware loopback.  Since I am focusing on mastering the use of ETC, and identifying/treating some remaining <20ms >-20dB reflections, I think this is the safe approach.

AustinJerry is online now  
post #3169 of 12760 Old 05-30-2013, 07:57 AM
AVS Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 5,545
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 973 Post(s)
Liked: 379
Quote:
Originally Posted by localhost127 View Post

the context was with respect to identifying total flight path of the indirect energy.
with the direct signal at T=0 you do not have accurate total flight paths as propagation delay is not accounted for.

Check "Set t=0 at IR Peak" in Preferences > Analysis
Other than that it shouldn't be too hard to subtract two numbers to get the time difference between them wink.gif

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is offline  
post #3170 of 12760 Old 05-30-2013, 08:24 AM
AVS Special Member
 
localhost127's Avatar
 
Join Date: May 2009
Posts: 2,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 8 Post(s)
Liked: 40
Quote:
Originally Posted by markus767 View Post

Other than that it shouldn't be too hard to subtract two numbers to get the time difference between them wink.gif

what does subtracting two numbers have to do with total flight path accuracy with respect to the data presented within the ETC - accounting for propagation delay, true acoustic center etc. im not sure how what is confusing about the initial statement?
localhost127 is offline  
post #3171 of 12760 Old 05-30-2013, 08:26 AM
AVS Special Member
 
jim19611961's Avatar
 
Join Date: Mar 2011
Location: Texas
Posts: 1,431
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 150 Post(s)
Liked: 89
Quote:
Originally Posted by AustinJerry View Post

OK, time for a bit of instruction for us neophytes: what does the term "terminator" mean?

They say a picture is worth a thousand words. Well, here is a pic from a Don Davis paper.



Here is the same pic, but I put a grid into for more precise interpretation.



And here is my attempt to emulate it.


My Room
My Music
Rega - Apollo, Rega - DAC, Goldpoint Passive, (2) Classe CA-100 bridged power amps (350w)
Jenzen Next
jim19611961 is online now  
post #3172 of 12760 Old 05-30-2013, 09:16 AM
AVS Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 5,545
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 973 Post(s)
Liked: 379
Quote:
Originally Posted by localhost127 View Post

what does subtracting two numbers have to do with total flight path accuracy with respect to the data presented within the ETC - accounting for propagation delay, true acoustic center etc. im not sure how what is confusing about the initial statement?

Yes sure, for some measurements a time reference is needed but not for all of them.

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is offline  
post #3173 of 12760 Old 05-30-2013, 11:01 AM
AVS Addicted Member
 
kbarnes701's Avatar
 
Join Date: Jan 2008
Location: Main Listening Positon
Posts: 19,579
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 2463 Post(s)
Liked: 2211
Quote:
Originally Posted by jim19611961 View Post

Quote:
Originally Posted by AustinJerry View Post

OK, time for a bit of instruction for us neophytes: what does the term "terminator" mean?

They say a picture is worth a thousand words. Well, here is a pic from a Don Davis paper.



Here is the same pic, but I put a grid into for more precise interpretation.



And here is my attempt to emulate it.

 

Interesting. I am nowhere near your level of expertise and I am striving towards the NE model. Here is my ETC set to the same time/dB parameters as you have used.  As you can see, by 140ms (before in fact)the graph in my case has disappeared off the radar. Is this because of the model I am trying to emulate, or is it some sort of basic error I am making, or have I misunderstood something?  I am very pleased with the sound BTW. The graph shows my left speaker, but the right and centre are essentially similar.

 

kbarnes701 is offline  
post #3174 of 12760 Old 05-30-2013, 11:08 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,384
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 986 Post(s)
Liked: 547
Quote:
Originally Posted by markus767 View Post

That's not the point. RT doesn't tell you anything more than a look at an ETC would. An ETC shows more information than RT. That's why I don't look at RT at all. Explore the "Filtered IR" tab in REW. Tells you a ton more about the sound field in a room than RT.
Oh, I didn't realize there was a competition going on between the metrics so the one that "shows more information" wins and we throw the other one out of the boat smile.gif. How would you judge the fuel gauge and speedometer in your car? Should we throw out the fuel gauge because the speedometer tells me more about what the car is doing?

So no, we don't judge our measurements that way. We are seeking clues as to what can be improved in our rooms and if a measurement provides data that the other one does, or doesn't readily, we use it. RT60 is one such measure.

As to me exploring filtered IR, I have already done that and then some. I provided a link to that post already. Here it is for the third time: http://www.avsforum.com/t/1453370/do-bass-traps-produce-noticeable-audible-difference/90#post_23024737. There, I explained exactly how it works, and how RT60 is computed using Schroeder Integral. There are graphs in it like this:

i-cgRFQVS-L.png

As I show there, the filtered IR is actually used to compute RT60. So if you think that filtered IR is good, you just accepted RT60 is good! biggrin.gif
Quote:
Then simply look at an ETC or a spectrogram or... it's all just a transform of the impulse response. Of course one has to understand what an impulse response represents and what the Fourier transform is.
A Fourier transform? The fourier transforrm takes a time domain signal and gives you the equiv. one in frequency domain. The X axis in ETC is in units of time. Clearly it is not a fourier transform. Here is both Impulse response and ETC overlaid in the same graph:

i-XcJ7btX-XL.png

The spikes are impulse response and ETC is that curve that follows those peaks and valleys to create their "envelop." The x-axis is still in time. So your notion that one is a fourier transform of the other is clearly incorrect.

The mathematical transform that gives the ETC from IR is called a Hilbert transform. I don't recommend that you try to read or learn it but since you asked if I know what I am talking about, here is a tutorial I wrote on that a year ago: http://www.avsforum.com/t/1413173/does-sound-sounds-better-in-a-room-full-of-furniture-and-stuff-or-without/210#post_22185014.

That said, REW uses a sweep log input signal instead of a true "impulse." By deconvolving that, we can extract the impulse response and from there, ETC. A fast way to implement that is using FFT. But that is an implementation detail and does not describe the relationship between ETC and IR.

So if you really understand ETC and you misspoke above, please come back and explain properly as what just happened, puts one in a very uncomfortable situation of confusing the speedometer in a car with the RPM meter. smile.gif
Quote:
If you believe in single scalar values. I don't.
Then I wonder why you stare at those spikes in ETC instead of stepping back, and understanding the macro picture of what reflections do in a room, which ones are beneficial and which ones are not, and acting on them that way. Instead, you look at single values in ETC which can be completely detached from reality. You are guilty of trusting scalar values which can be very removed from their audible effects.

I and countless other experts use RT60 at mid-frequencies, because for that one characteristic of the room, it gives us a quick answer. We don't use it as sum total of what we use to analyze the room, or remotely so as you tried to imply there. It is a tool in our toolbox and it happens to be one that gives an answer quickly and accurately for the purpose we use it. The continued resistance against its use appears to be an emotional reaction from folks who thought it is not a valid measure. Now that I have shown it to be, to satisfaction of many experts on Gearslutz including those who have written books on acoustics, is puzzling to see so much energy put to fight it. And constantly getting upset over it to boot.
Quote:
So ETC is useless because people "struggle with ETC"? Nobody every said that analysis of an LTI system would be easy.
But it can be "easier" if you rely on modern acoustic science that is based on listening tests and science of acoustics discussed outside argumentative forum posts. That is the beauty of that system. I explained how once you know that science, i.e. how our ears and brain work, you can immediately know which walls you should treat and which ones not. No measurement is needed at all.

I see folks now say "let's use filtered ETC." Do you all know what a window function and lobing is? Because if you don't, you don't know what artifacts you create when you apply a filter that way. All of those obscure parameters in REW is there for a reason. Click on Preferences->Analysis and look at the top parameters for Impulse Response: There is a reason they are there because there is no one right answer. Teaching someone signal processing concepts like this is extremely difficult. You want to see someone struggle? That would be the definition of it.
Quote:
Nothing I've said negates anything in that quotes. What makes you think I'm not familiar with the literature?
You claimed no one should use RT60 and I showed you expert after expert who uses it. You asked again who is 'we' as if it was only me. So again, I provided more citation. Now you say it doesn't negate anything? Of course it negates things. It negates you saying RT60 is of no use. As to familiarity with literature, you need to demonstrate that by not asking me questions that are so readily answered in literature. And cite them to back your claims. You have done neither and you want me to think you have read them? Have you read Dr. Toole's book? Do you have access to ASA papers and can answer a question if I asked you about one of the papers there? How about AES?
Quote:
What makes you think that I wouldn't know "how it works and how it can misfire"? That's ignorant at best.
So you know it can misfire? That's great! Why did you wait 'till now to state your agreement? Please list the ways it can so that folks know to avoid those land mines.

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #3175 of 12760 Old 05-30-2013, 11:23 AM
AVS Special Member
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 7,983
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 1097 Post(s)
Liked: 914
Quote:
Originally Posted by kbarnes701 View Post

 

Interesting. I am nowhere near your level of expertise and I am striving towards the NE model. Here is my ETC set to the same time/dB parameters as you have used.  As you can see, by 140ms (before in fact)the graph in my case has disappeared off the radar. Is this because of the model I am trying to emulate, or is it some sort of basic error I am making, or have I misunderstood something?  I am very pleased with the sound BTW. The graph shows my left speaker, but the right and centre are essentially similar.

 

 

I see some significant reflections above -20dB and below 20ms.  Have you been able to identify what those are?

AustinJerry is online now  
post #3176 of 12760 Old 05-30-2013, 11:38 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,384
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 986 Post(s)
Liked: 547
Quote:
Originally Posted by kbarnes701 View Post

Interesting. I am nowhere near your level of expertise and I am striving towards the NE model. Here is my ETC set to the same time/dB parameters as you have used.  As you can see, by 140ms (before in fact)the graph in my case has disappeared off the radar. Is this because of the model I am trying to emulate, or is it some sort of basic error I am making, or have I misunderstood something?  I am very pleased with the sound BTW. The graph shows my left speaker, but the right and centre are essentially similar.


Looks like you have confirmation that you have accomplished what you set out to do in your room Keith. To wit, eyeballing the RT60 time from your response above, gives a value of about 0.15 seconds. That means your room is highly absorbant which optimizes speech from late reflection point of view. You have been telling us speech is very clear to you; that likely is the reason. I expect the sound to be focused where the speakers are as opposed to a wider "image source" due to heavy absorption in the room (especially at early reflection points). You will then be relying 100% on ambiance or such characteristics in the content which you have been telling us is your strategy. I expect the room to not be as optimized for music which is again what you have already observed. It may be a bit uncomfortable as far as a space to converse in due to high level of absorption. If this is for movie watching, then it is fine that way as one can go as low as 0.2 RT60 there.

Given that these are all the things you desired to do, using this type of macro analysis we know you are already there. Just be careful to not keep adding absorption as you will soon create your own anechoic chamber smile.gif.

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #3177 of 12760 Old 05-30-2013, 11:39 AM
AVS Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 5,545
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 973 Post(s)
Liked: 379
Amir, you don't seem to read my posts or you don't understand what was said or I didn't express myself clear enough so you could understand what was meant. Either way, I'm not interested in a conversation with you any longer as I feel it's a waste of my time.

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is offline  
post #3178 of 12760 Old 05-30-2013, 11:39 AM
AVS Special Member
 
jim19611961's Avatar
 
Join Date: Mar 2011
Location: Texas
Posts: 1,431
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 150 Post(s)
Liked: 89
Quote:
Originally Posted by kbarnes701 View Post

Interesting. I am nowhere near your level of expertise and I am striving towards the NE model. Here is my ETC set to the same time/dB parameters as you have used.  As you can see, by 140ms (before in fact)the graph in my case has disappeared off the radar. Is this because of the model I am trying to emulate, or is it some sort of basic error I am making, or have I misunderstood something?  I am very pleased with the sound BTW. The graph shows my left speaker, but the right and centre are essentially similar.



Firstly, the models you and I are trying to emulate are different. Mine the LEDE / RFZ and yours the NE model. So comparing our ETC's is a bit apples and oranges.

I am no expert on NE rooms, so take anything I say, and verify it with someone who knows more than I about the NE model, but from what I understand, the ETC on a NE room should look relatively anechoic. In your case, those first 10ms need some reflection attenuation. But what makes the model work is a reflective floor and a reflective front wall in order to establish necessary psycho-acoustic room cues. I also believe it incorporates some ceiling diffusion aimed back at the front reflective wall for bringing up the level of those cues.

Again, look into this yourself, but I think all these criteria must be there for a NE room to sound like it should. Its much more than just emulating a anechoic ETC response.

My Room
My Music
Rega - Apollo, Rega - DAC, Goldpoint Passive, (2) Classe CA-100 bridged power amps (350w)
Jenzen Next
jim19611961 is online now  
post #3179 of 12760 Old 05-30-2013, 12:10 PM
AVS Addicted Member
 
kbarnes701's Avatar
 
Join Date: Jan 2008
Location: Main Listening Positon
Posts: 19,579
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 2463 Post(s)
Liked: 2211
Quote:
Originally Posted by AustinJerry View Post

Quote:
Originally Posted by kbarnes701 View Post

 

Interesting. I am nowhere near your level of expertise and I am striving towards the NE model. Here is my ETC set to the same time/dB parameters as you have used.  As you can see, by 140ms (before in fact)the graph in my case has disappeared off the radar. Is this because of the model I am trying to emulate, or is it some sort of basic error I am making, or have I misunderstood something?  I am very pleased with the sound BTW. The graph shows my left speaker, but the right and centre are essentially similar.

 

 

I see some significant reflections above -20dB and below 20ms.  Have you been able to identify what those are?

 

I thought that the aim was to get -20dB *after* 20ms?  I am fairly sure that one of the earlier reflections is from a light fitting in the centre of the ceiling, about which I can do nothing at all. Not a thing unfortunately. I have mooted removing it (it is rarely used) but Mrs K is against it, and I am not going to antagonise her over it given how co-operative she is in 'most all other ways. When I did the string thing it landed right on the light fitment. There may be other things that are issues too. 

kbarnes701 is offline  
post #3180 of 12760 Old 05-30-2013, 12:13 PM
AVS Addicted Member
 
kbarnes701's Avatar
 
Join Date: Jan 2008
Location: Main Listening Positon
Posts: 19,579
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 2463 Post(s)
Liked: 2211
Quote:
Originally Posted by jim19611961 View Post

Quote:
Originally Posted by kbarnes701 View Post

Interesting. I am nowhere near your level of expertise and I am striving towards the NE model. Here is my ETC set to the same time/dB parameters as you have used.  As you can see, by 140ms (before in fact)the graph in my case has disappeared off the radar. Is this because of the model I am trying to emulate, or is it some sort of basic error I am making, or have I misunderstood something?  I am very pleased with the sound BTW. The graph shows my left speaker, but the right and centre are essentially similar.


Firstly, the models you and I are trying to emulate are different. Mine the LEDE / RFZ and yours the NE model. So comparing our ETC's is a bit apples and oranges.

I am no expert on NE rooms, so take anything I say, and verify it with someone who knows more than I about the NE model, but from what I understand, the ETC on a NE room should look relatively anechoic. In your case, those first 10ms need some reflection attenuation. But what makes the model work is a reflective floor and a reflective front wall in order to establish necessary psycho-acoustic room cues. I also believe it incorporates some ceiling diffusion aimed back at the front reflective wall for bringing up the level of those cues.

Again, look into this yourself, but I think all these criteria must be there for a NE room to sound like it should. Its much more than just emulating a anechoic ETC response.

 

Thanks. I am reading around the subject in general but I have much to learn wrt to this aspect of the hobby. What you say about the floor is interesting because I have a fitted carpet on the floor (and it will have to stay there to keep Mrs K happy) but the ceiling is fairly bare. I am getting there slowly I think...

kbarnes701 is offline  
Reply Audio Theory, Setup, and Chat

Tags
Dayton , Dayton Audio , Room Equilizer Wizard Rew

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off