or Connect
AVS › AVS Forum › Audio › Audio theory, Setup and Chat › Simplified REW Setup and Use (USB Mic & HDMI Connection) Including Measurement Techniques and How To Interpret Graphs
New Posts  All Forums:Forum Nav:

Simplified REW Setup and Use (USB Mic & HDMI Connection) Including Measurement Techniques and How... - Page 26

post #751 of 9543
Quote:
Originally Posted by jevansoh View Post



With that being said, if you look at the graph on a vertical scale in 5db increments it won't look as good.

If you look at it with Audyssey off and always totally unsmoothed, it won't look as good.


With enough EQ and on the proper scale almost any FR can be made to "appear" totally flat, but the real test is to see the waterfall and look at the decay times and any potential ringing.

I'd much rather have a FR that is +/- 5db for instance, yet overall even decay rates that are similar to the higher frequencies than a room that appears to be totally flat but has awful ringing. It's all in how you look at things and whether or not you want pretty graphs or a great sounding room.


BTW, you can absolutely get rid of the effects of that null, at least in one position, pretty easily by simply raising a subwoofer off the ground.


--J

 

In the interests of full disclosure, here are my graphs usning (what I recall) as the posting guidelines.

 

Low frequency response, 15-300Hz, no smoothing, vertical scale in increments of 5dB, left+right+subs, with Audyssey

 

 

Full range 15Hz-20kHz, 1/12 smoothing, left+right+subs, with Audyssey

 

 

Waterfall 15-300Hz, vertical scale in 5dB increments, 600ms time range, left+right+subs, with Audyssey

 

 

ETC with vertical scale in 5dB increments, out to 40ms, left speaker only, Audyssey off

 

 

I am considering adding a dedicated section in the guide covering how to configure REW measurements, and how to post the results.  I'm not sure if I have it all correct now, but we can review and correct the guidelines once they are captured.

Gear mentioned in this thread:

post #752 of 9543
Quote:

Originally Posted by jevansoh View Post

 
We see the first reflection at about .5ms which is almost certainly from the speaker cabinet itself. Depending on how the mic is mounted/placed it could even be from that, but I suspect diffraction effects or the speaker grill, depending on how it's designed (take more measurements without the grills), etc. This does need to be addressed, however.

 

Here's a picture of the MKSound S150II's LCR with my AT screen removed (sub not shown but sits in right corner):

 

 

This is an earlier photo.  I've sinced moved the L and R a little further to the outside (to make the soundstage as wide as possible) so I'll check to make sure they're not reflecting off the frame of the screen as it is 3.7" wide aluminum rails.  As you can see, I've tried to keep all the grilles well over the front edge of the stands even on the L and R where the toe in leaves a pretty good overhang on the outside edge.  I thought I read somewhere that MK suggest leaving the grilles on the speakers due to their design but perhaps Keith can chime in here as he has the exact same speakers.  Nevertheless, I can try removing the grilles when I take my next set of measurements.

 

Quote:
Originally Posted by jevansoh View Post
Second, we see a very high gain specular reflection way too early at around 3ms. The way to figure where this is coming from is to multiply 1.13 times 3ms in this case. (Always multiply the timing of the reflection by 1.13) and ADD that to the "actual" distance between the mic and the acoustic center of the speaker. If the speaker uses a D'Appolito array then you can use the tweeter. That's how long the path of this reflection is traveling in total. You can then mark on a tape measure or cut a string to this length (adding a few inches at each end to be able to tie the string, but making sure the marks on the string are at the exact added distance to the actual distance between mic and speaker in the end) and pull the string taut in ALL directions until you find all the surfaces the string touches. Then, one by one, add an absorber or large piece of insulation to that area and remeasure until the reflection is at least -20db in gain.

Do this for all reflections out to about 20ms.

In a lot of conversations we hear that either all reflections are bad or in some alternative points of view we hear they are helpful and needed.

My philosophy (based on acoustical models that have been around for decades and folks much smarter and with much more experience than me all but unanimously agreeing) is that "high gain" and "early" reflections aren't good and in fact will cause smearing and not give you some of the main psycho acoustic benefits we are striving for in our small acoustical spaces, such as making them sound like larger rooms!

However, a lot of people that follow this philosophy stop there. The fact is, all reflections are NOT bad! In fact, they are necessary! Depending on the size of your room and your level of commitment, anywhere from about 12-20ms (larger rooms up to 25ms) should have a termination of the effectively anechoic space we are creating by reducing the early reflections to -20db in gain relative to the direct sound. There should be reflections after somewhere between 12-20ms but they should be diffuse and dense and not just one or two here and there.

This is difficult to achieve and if you go around filling the room with tons of absorption it could be almost impossible.

There can be too much of a good thing. So, test each spot (mirrors and the "mirror trick" are USELESS for this very reason) relating to the spike in the ETC and ONLY place absorption where it is absolutely necessary to tame that spike.

There are other methods that don't require any use of absorption in the specular region such as reflection/deflection/diffusion but those are a bit more advanced topics we'll touch on later.

As far as the frequency response, it is true you have a null/room mode at 40hz. Due to the wavelength (very long) at this frequency, moving "just a little" to the left/right isn't going to do anything and will just make other things worse. Symmetry is very important.

If you had two subs though, we could effectively cancel this null through placement.

Invest in another sub. smile.gif

If you cannot do that, then through placement of the sub and tweaking of the phase (distance controls, measure in one foot increments starting 5-6' below what Audyssey found and going to 5-6' above) you will still be able (most likely) to smooth things out a bit more.

Make sure to do all this with Audyssey OFF.

 

As I mentioned previously, I will be investing in an upgrade (ideally dual Submersives) to my subwoofer next year or maybe even later this year.  In the meantime, I would like to address as many other issues as possible and address the 40Hz null as much as possible.  I've read a little bit about the "string test" but perhaps it might be worth going into more detail on the exact steps I need to take as I'm sure others following along will undoubtedly benefit from a detailed explanation of this when they start measuring themselves?

 

Quote:
As for the waterfalls, it looks like unfortunately Audyssey made things worse in the very low frequencies. Luckily, in the range from 20-300hz (some noise is showing and you might want to raise the bottom limit to 42 or 43db just to make it easier to see the "truth") you don't have a lot of problems that will be extremely difficult to fix. The biggest one is of course between 30-40hz and regular velocity based (pink fluffy) absorption isn't going to work well at that frequency, so placement, phase, and EQ are the best bet for that one unless you don't mind spending some money and getting some pressure based tuned traps for 30-40hz.

Audyssey, by taming the peak at 30hz helped cut down the ringing a bit, but it's still too much and you need bass traps for the rest. 40hz was made worse by it boosting the null in trying to fix it. This is where another sub and/or placement could help a ton.

The ETC's for the L and R channels (you did measure only one at a time, right? Test sweep tones coming only from the L channel for the measurement marked L, right?) are very similar. This is a good thing. The example I gave for the first reflection (diffraction most likely) and the second one at 3ms applies to all the others that are peaking above -20db as well. Just multiply 1.13 by the ms of the reflection and add it to the physical distance between mic/speaker then pull a string taut in all directions, placing insulation/absorption at each spot one at a time, measuring in between, until you tame the peak. You have to do this for all seating positions that you care about and you have to do this separately for the L and R channels though, ok?

So... You have some work to do. I can't wait to see some results! wink.gif

I'm sure I've missed a few things and hope I haven't confused you but we now have a dialog going so feel free to ask more specific questions and I'll be more than happy to help you.

Hope this helps,

--J

 

My original measurements were only for the L+sub and R+sub separately.  I forgot to measure L+R+sub.  However, as I mentioned in a previous post, these REW measurements were taken with an OM mic in the vertical position and the calibration file loaded.  How much will this impact the suggestions given?  I will be remeasuring everything as soon as my UMM-6 mic arrives.

 

Thanks again to everyone for the great feedback.  It feels good to advance this thread from basic setup to actually taking some measurements - as well as to start exposing some of those pitfalls as well! wink.gif

 

 Joe

post #753 of 9543
Quote:
Originally Posted by AustinJerry View Post

Here are some waterfalls:
Please keep in mind the thread I post on the first page here. That as a matter of principals of signal processing, you can either have resolution in time, or frequency domain. You cannot have both. This directly impacts fidelity of waterfalls. Unless you have optimized for the view you are interested in, you could get the wrong data.

Quote:
The first three reflection points (common to both graphs) are:




Just taking a guess, the second reflection (18.7 ft) could be caused by the back wall (blue line).  The third reflection (24.6 ft) could be caused by the signal from the left speaker bouncing off the right wall (red line).  The first reflection (9.8 ft) could be the ceiling.  The reflections are pretty serious, down only -5dB to -7dB.  Recall, the general objective is to have all reflections below the -20dB level.  To determine for sure what is causing the reflections, you need to conduct the "string test" (familiar with that?).
I am afraid you can't perform such computations. The peak impulses shown in time domain are spectrum sensitive. Change that and the values can be quite wrong. Since your speaker does not have the same frequency response in all directions, the spectrum of the signal as it arrives at different reflection points will be different. Therefore you can't compare the peak values to each other. Listening tests confirm the same. Here is a test from Dr. Toole and Olive that started with a full spectrum sound. They took a reflection and filtered it down to 500 Hz. They then raised its levels until it matched the unfiltered one (both at the threshold of detection). Here are the measurements:

i-sNvGjhR-L.png

Figures (b) show the frequency response. As we naturally see, the filtered reflection level had to be boosted by 3 dB to compensate for the truncated spectrum.

Now look at figures (c). We see that the ETC shows completely erroneous results. It claims that the reflection in the filtered version is 20 dB down. Again per above, the perceptual level was equal and measured actual spectrum was 3 dB higher!

If you are still not convinced, here is additional evidence in the form of peer reviewed paper in the Journal of Accustic Society of America (ASA), "The Active Listening Room A Novel Approach to Early Reflection Manipulation in Critical Listening Rooms", Naqvi and Rumsey (AES Fellow):

"Energy–time curve (ETC) measurements were taken at each stage to establish the relationship of such measurements with subjective assessments. They reported that the frequency spectrum of reflected sounds is normally reduced in high-frequency content, but reducing the high frequency content of reflected sounds had a relatively minor effect on the absolute threshold of several common sounds. However, the ETC measures of these sounds changed considerably, leading to the possibility of underestimating the audibility of a reflection when using the peak ETC level as an indicator. The frequency response of the reflection seemed to be a more reliable indicator. The unsuitability of ETC as a measure is discussed in detail in later sections."

They go on to say:

"The practical consequence of this observation is that using the peak level of ETC spikes as a measure of the audible importance of room reflections can lead to a serious underestimation of the importance of reflected sounds and the normal off-axis response of loudspeakers, which contain less high-frequency energy than direct sound."

The situation will get a lot worse as you proceed to treat your room. Any acoustic products deployed, or even normal furnishings such as carpets, curtains, etc. change the spectrum of the reflected sound. As soon as they do that, you no longer can rely on the measured values and certainly can't compare them to other spikes with different spectrum.

Quote:
If I am correct, this would suggest a requirement for full-band acoustic treatments on the side walls at the first reflection points (always a must), ...
Always a must? On what basis?
post #754 of 9543

Here is a picture of how to configure the "string test":

 

 

Instructions:

 

- Generate the ETC graph

- For the early reflection you want to identify, place the cursor at the peak of the reflection and click.  You will see the the level on the vertical scale, and the distance on the horizontal scale.  For example, in my ETC, I have a reflection of -19dB at a distance of 5m (16.25 ft).

- Multiply 16.25 ft by 1.13 to get 18.4 ft (not sure why we multiply by 1.13, but perhaps Jason will explain).

- Now cut a length of string so that 18.4 ft is marked on it (add 6-12 inches on each end to secure it in place).

- Tape one end of the string so that the mark on the end of the string is on the front grill of your speaker at tweeter height.

- Secure the other end of the string at ear level where the mic was placed during the measurement.  Taping it to the mic stand works well.

- Now, holding the string in the center, move around the room until the string is touching a hard surface.  Recalling your old trigonometry lessons, the angle of incidence should equal the angle of reflection.  When you find such a place, you have likely found the source of the reflection.

- Treat that point on the hard surface with a temporary treatment.  A square of R-30 insulation (i.e. "pink fluffy"), available at your local hardware store, is a good temporary treatment.

- Re-run the ETC measurement to see if you have indeed found the source of the reflection, and whether the treatment reduced the reflection to the target value of -20dB.

 

I have had varying success using this method.  The string test is somewhat difficult to accomplish for ceiling reflections--be creative!

post #755 of 9543

@ Amir

 

So am I correct in concluding that there is no hope and we should all stop generating ETC graphs?

 

And regarding your reference to the link on Whatsbestforum, yes, I have attempted to comprehend the theory that you presented in the thread, but it is over my head.  Perhaps we are moving in the direction of comprehension, but we aren't there yet. 


Edited by AustinJerry - 1/29/13 at 9:38am
post #756 of 9543
Quote:
Originally Posted by jkasanic View Post

 

Here's a picture of the MKSound S150II's LCR with my AT screen removed (sub not shown but sits in right corner):

 

 

 

Joe,

 

May I ask your reasoning for placing the speakers on what look like tables?  Are you not concerned with the effect the flat surface will have in reflecting sound?  Did you consider more conventional speaker stands?

post #757 of 9543
Quote:
Originally Posted by jevansoh View Post

Quote:
Originally Posted by jkasanic View Post

Quote:
Originally Posted by dstew100 View Post

Quote:
Originally Posted by audioguy View Post

If only.smile.gif Not in my room. Short of stacking subs from floor to ceiling, (I've thought about it biggrin.gif )I will never not have a dip (null) at about 90hz from the ceiling bounce. And I have 4 identical subs placed to optimize the room response. We may want to think about using a sample that is this ruler flat as I would suggest that this graph might set unrealistic expectations. It might work in a few rooms, but not all, regardless of the number of subs.


I've read that raising a subwoofer up off the floor can address the floor ceiling mode. Apparently rooms are 3-dimensional under current understanding which is why Geddes approach is 3 subs.

I read this somewhere too but I don't recall any method for determining just how high it should be?  Is there a general rule based on the distance from floor to ceiling?  I suppose measuring is the only surefire way to know.

It can be calculated and is different for every room. It has to do with the height of the room, the mode that creates, and when you calculate that, you can figure out the placement.

There is an Excel spreadsheet on the Harman web-site that allows you to input your room dimensions and it will show you, graphically, where your modes are and at what area, both graphically and printed out in specific hz vs ft and inches where the modes are strongest, weakest, and mid-way.

It's pretty neat and a great tool to be able to visualize best placement for speakers/subs and listeners in the room.

It's a starting point, of course. You will have to measure and it won't be perfect as it assumes true boundries of thick concrete walls for instance, where you may actually have openings, or only one or in some cases two layers of drywall and mass is critical when trying to calculate modes, as is symmetry/rectangle room, but it is always pretty close and I use it a lot.

I'll post the link when I have time to dig it up later unless someone else beats me to it.

If I knew the height of the ceiling and the height of the listener's ears, I could tell you approximately where to place the sub vertically to negate the effects of the null in that position, but would need more detail.

Find that spreadsheet on the Harman web-site and you can see for yourself. I got lucky and in my case only had to raise a sub (it only took 1 and I have 4) 15" off the ground and I already had a 12" riser for the second row in a dedicated theater, but it can be done in any room.

--J

http://www.harman.com/EN-US/OurCompany/Innovation/Documents/Calculators/Room%20Mode%20Calculator.xls
post #758 of 9543
Quote:
Originally Posted by AustinJerry View Post

So am I correct in concluding that there is no hope and we should all stop generating ETC graphs?

his extent of understanding of the ETC is from what he can copy-paste, and not from his actual experience in acoustics or treating/modification of rooms. the ETC is but a tool, and he merely looks for all of the ways the tool can be used or interpreted incorrectly, versus how it can be used beneficially. where i come from, we refer to this type of behavior as OPERATOR ERROR.

you'll notice he displays zero real world experience on 3D polar / IACC ETC's. his commentary is limited by his lack of experience and to what he can "google" and copy-paste from papers taken out of context. he's stated plenty of times in the past that he "sees no reason for time-domain analysis" in our such rooms.



in this application, if a user has made the design decision to attenuate early arriving, high-gain, sparse, indirect specular reflections (eg, first-order reflections from sidewall, ceiling, etc) - then the ETC as shown is more than satisfactory. this is because the treatment procured for such an en-devour (eg, porous absorption or a large reflector panel to redirect the reflection away from the listening position, or diffuser) is called to be BROADBAND in nature. if the treatments are broadband, then the spectral content contained in the indirect reflection is not required - as the broadband treatment is effective throughout the entire specular region. this is why sufficiently THICK porous absorbers are required - to fully attenuate the indirect reflection versus thin absorbers which merely color/eq/filter the reflection as thin absorbers are not sufficient to fully attenuate the lower band.

who cares what the boundary does to the reflection's spectral content, or what spectral content is present due to the speaker's polar response if we're using BROADBAND treatment. but then maybe now you can ask him why he deploys baby diffusers in his company's showroom, versus true broadband diffusers (broadband treatment) that toole calls for,


note also the actual acoustical impedance of the boundary changes the spectral content of the reflection - so you can't just assume the spectral content by the source' polars alone. but this user simply lacks the acoustical experience to understand why this matters and how it can be measured.

oh, and the ETC can also be used to identify signal alignment issues, or speaker cabinet diffraction or other destructive early colored energies - but you notice he will never imply any of this is relevant as his primary goal here is to completely write-off the one tool that was responsible for acoustics taking a quantum-leap some 40+ years ago.

if you want to see him struggle, simply produce a frequency-response graph and ask him to talk about intelligibility, localization, and imaging wink.gif
post #759 of 9543
and this is probably one of the greatest illustrative room mode calculators i've stumbled across: http://www.hunecke.de/en/calculators/room-eigenmodes.html

..as it shows the pressure distributions in 3space quite handsomely.

note what was said above about these calculators assuming fully resistive boundaries (eg, concrete walls). the calcs will fall apart with complex impedance bounadries, and you may even see other behaviors such as room modes from adjacent spaces (coupling) as well. this is why measurements of the acoustical space trump the calculators - but they do play a role.
post #760 of 9543
Quote:
Originally Posted by AustinJerry View Post

Here is a picture of how to configure the "string test":

make note that to accomplish this, you need to be utilizing hardware loopback on your measurement kit!

eg, if you have the direct signal on the ETC corresponding to T=0, then your indirect reflection distances will be invalid.

T=0 should correspond to the time when the signal is generated at the source/speaker (taking into account hardware propagation delay). and thus, the direct sound should be seen on the ETC at the distance (time) it took to traverse from source --> listening position.
post #761 of 9543

The REW Guide has been updated to Version 1.08 (a link to the new document is in my signature).

 

Changes:

 

- Revised Windows Audio Playback Devices configuration to reflect both HDMI and non-HDMI connections.

- Added Section 3b for non-HDMI configurations using the Java interface.

- Moved Revision History and Credits sections to the end of the document.

- Inserted description of the “String Technique” in the Generating Impulse Graphs section.

post #762 of 9543
Quote:
Originally Posted by localhost127 View Post


make note that to accomplish this, you need to be utilizing hardware loopback on your measurement kit!

eg, if you have the direct signal on the ETC corresponding to T=0, then your indirect reflection distances will be invalid.

T=0 should correspond to the time when the signal is generated at the source/speaker (taking into account hardware propagation delay). and thus, the direct sound should be seen on the ETC at the distance (time) it took to traverse from source --> listening position.

 

Thank you for reminding me of this.  You were kind enough to make the same recommendation in another thread several months back.  I'll spare you the excuse that my retention isn't what it should be.  The mind is the second thing to go, and I can't remember which is the first....  wink.gif

post #763 of 9543
Quote:
Originally Posted by localhost127 View Post

and this is probably one of the greatest illustrative room mode calculators i've stumbled across: http://www.hunecke.de/en/calculators/room-eigenmodes.html

..as it shows the pressure distributions in 3space quite handsomely.

note what was said above about these calculators assuming fully resistive boundaries (eg, concrete walls). the calcs will fall apart with complex impedance bounadries, and you may even see other behaviors such as room modes from adjacent spaces (coupling) as well. this is why measurements of the acoustical space trump the calculators - but they do play a role.

 

Interesting link.  I'm not exactly sure how one would use it as an aid on which to base remedial actions.

post #764 of 9543
Quote:
Originally Posted by localhost127 View Post

make note that to accomplish this, you need to be utilizing hardware loopback on your measurement kit!

eg, if you have the direct signal on the ETC corresponding to T=0, then your indirect reflection distances will be invalid.

T=0 should correspond to the time when the signal is generated at the source/speaker (taking into account hardware propagation delay). and thus, the direct sound should be seen on the ETC at the distance (time) it took to traverse from source --> listening position.

If you don't have the loopback in place can you just use the distance (time) for the direct sound from source-->to reach LP and back calculate what the reflected distance would be if T=0 was signal and not signal reaching LP. Sounds too simple, sorry if I missed something.
post #765 of 9543
Quote:
Originally Posted by jevansoh View Post


It can be calculated and is different for every room. It has to do with the height of the room, the mode that creates, and when you calculate that, you can figure out the placement.

There is an Excel spreadsheet on the Harman web-site that allows you to input your room dimensions and it will show you, graphically, where your modes are and at what area, both graphically and printed out in specific hz vs ft and inches where the modes are strongest, weakest, and mid-way.

It's pretty neat and a great tool to be able to visualize best placement for speakers/subs and listeners in the room.

It's a starting point, of course. You will have to measure and it won't be perfect as it assumes true boundries of thick concrete walls for instance, where you may actually have openings, or only one or in some cases two layers of drywall and mass is critical when trying to calculate modes, as is symmetry/rectangle room, but it is always pretty close and I use it a lot.

I'll post the link when I have time to dig it up later unless someone else beats me to it.

If I knew the height of the ceiling and the height of the listener's ears, I could tell you approximately where to place the sub vertically to negate the effects of the null in that position, but would need more detail.

Find that spreadsheet on the Harman web-site and you can see for yourself. I got lucky and in my case only had to raise a sub (it only took 1 and I have 4) 15" off the ground and I already had a 12" riser for the second row in a dedicated theater, but it can be done in any room.

--J

 

OK, here are the results for my room:

 

 

What is missing (at least for my understanding) is how to actually use this as a guide for speaker placement.

 

Take my listening room as an example.  I have my room oriented so that my front speakers and my TV display are on the width of the room (19' 6"), the middle graph above.  So, what does the graph tell me about the recommended horizontal position of my left and right speaker?  How about two subs placed on the front wall (BTW, the two subs are placed at the 1/4 and 3/4 distances).  And should the MLP be exactly in the center for symmetry?

 

Does the top graph suggest how far from the front wall my speakers should be placed?

 

And does the bottom graph suggest whether I should raise a sub off the floow, and by how much?

 

This is how my front speakers and subs are configured now:

 

post #766 of 9543
Quote:
Originally Posted by lbrown105 View Post


If you don't have the loopback in place can you just use the distance (time) for the direct sound from source-->to reach LP and back calculate what the reflected distance would be if T=0 was signal and not signal reaching LP. Sounds too simple, sorry if I missed something.

 

Is this perhaps why Jason said to multiply the distance by 1.13?

post #767 of 9543
Quote:
Originally Posted by jevansoh View Post

Quote:
Originally Posted by AustinJerry View Post

Quote:
Originally Posted by jkasanic View Post

Jerry, I saw your post on specific responses to your questions so even though some of this is repetitive from my previous post, I thought the least I could do is collect all of your questions and answers into one post!  Thanks again for your help!

Now that you mention it, I recall seeing your posts in the Audyssey thread.  Sorry, I monitor too many threads, but I should have remembered.  The advice you received from Roger Dressler is good advice, so I have nothing to add at this time.

I plugged in your dimensions in the Room Mode Calculator:






The primary resonant frequency associated with the width of the room is ~40Hz.  If your MLP is halfway between the left and right walls, this would place you directly in the null associated with 40Hz, which the measurements seem to support.  One way to address this would be to move the MLP to the right or left a bit and re-measure to see if there is an improvement.

The secondary resonant frequency associated with the width is 80Hz, and the secondary for the length is 62Hz, and the primary for the height is 65Hz.  I don't see anything associated with room modes that might be causing the issue at ~75Hz.  You should experiment with the sub/mains phase adjustments.

I think we should wait for Jason's opinion regarding whether bass treatments might be a solution for you.  For example, I have 96 sq ft of bass traps (eight 4x2 velocity traps, half from RealTraps, the other half from GIK).  Even though the treatments are not in a dedicated HT room, I am single and don't have WAF issues to deal with.

I think I covered most of these questions in my previous post, but let me know if I missed anything.

Jerry is on the right track here. What is your ceiling made of? If for instance it's in a basement and is a drop ceiling, you need to measure to the sub floor above as acoustically the ceiling doesn't even exist. That could be an explanation for the discrepancy between the 62hz and 75hz. That's just one example, of course.

For all but your first axial mode you can pretty easily treat the other ones with velocity based absorption.

In looking at the size and layout of your room, you are a perfect candidate for cheap and easy Superchunk style bass traps.

If possible (and it's easiest) I'd simply go buy R30 insulation (15" x 300" x 9.5" thick rolls at Lowe's for $15.87 last time I bought it a month or two ago) and make a soffit style trap in each corner. You could use two or four pieces, hanging from ceiling to floor. I use 4 pieces for a roughly 18x30" trap. This way you don't have to cut a bunch of triangles and support them so they don't compress (you don't want compression - you need it to stay as fluffy as possible) and you'll get good absorption with this type/style all the way down to 40hz.

Depending on your decay times and what it takes to treat your reflections (need you to try and clean up that ETC first) you may want to get some Contac brand self adhesive paper and adhere it directly to the soffit insulation (which will be hanging in a simple frame made up of 1x2's or 2x2's) so you don't over absorb the highs and the trap becomes a true bass trap instead of a huge broadband absorber.

If, though, you don't end up needing a lot of insulation to clean up the ETC and your room is over 300-400ms at an overall/average specular decay rate then you can do without the Contac paper as you'll want to bring those decay rates down a bit.

I've yet to see a room that cannot benefit from absorption in the modal region though. So no matter what, I'd build some inexpensive and easy floor to ceiling bass traps. You have 12 corners in your room, remember. Do the 4 main corners first, measure, and let's go from there.

You can buy them too, if you have more money than time and/or DIY skills, but... Be careful. If you're going to buy products make sure you get products that perform well and have been independently measured.

I recommend GIK Acoustics (at this time - more to come on that later) as they are honest, I've worked with them many times before, they have great performing products, a fast turnaround time, great looks, and quite frankly the cheapest prices, in some cases by far compared to any other company currently on the market today) and specifically the Soffit Trap (still a whole lot cheaper to build yourself though) the Tri-Trap, the 244 bass trap (which I believe should be used as broadband reflection panels to clean up the ETC more so than actual bass traps other than maybe some filler for problems with higher frequencies) and then if you still have problems you need to take care of in the lowest frequencies, they have Pressure based absorption called Scopus Tuned Traps that perform well (you'll need more than a few though) that will take care of your lowest modes and these type are NOT easy to build. It can be done but you need some good skills and a lot of time and patience.

I talk more about this in my previous post, but remember, for your 40hz width mode, if you had two subs, one on each side wall, they would combine to effectively produce 40hz right in the center of the room, which is your current MLP (and the best place for it as moving it just a bit isn't going to make a difference at that frequency) and your null would be gone!

I strongly recommend investing in a second subwoofer or upgrading if possible and buying two new capable subs.

Without knowing your budget (both money and time) and what exactly you hope to accomplish it's a bit difficult to offer any more advice at this time.

Hope this helps,

--J

 

The ceiling is drywall at approx. 8' 8" from the carpeted concrete floor.  I can easily check the distance to the subfloor above (if that's still necessary?) as I have an access panel for some plumbing that is above the drywall (and no, it wasn't there when I moved in!).  The wall behind the LCR speakers is drywall over concrete.  The entire right side of the room is interior drywall with two door size openings (one for the stairwell into the basement which is the back right corner and one leading down a short hallway to the unfinished portion of the basement which is directly to the right of the false wall in the upper right corner.  These are visible in the lower perspective view shown a couple posts above.

 

As for construction of the bass traps, if you have more detail on the DIY construction of soffit traps would be great?  I've only ever seen the "triangle stacks" for corner bass traps but I certainly am willing to invest in making my own based on the description above.  I briefly looked at the ones from GIK.  It would appear as though the frame is just an L shape with the edge meeting the corner and the top and front two sides are wrapped in AT material?  I could easily do two in the corners behind the false wall (as well as the floor/ceiling corners behind the wall) but the exposed corner in the back right of the room could be problematic (aka nothing or maybe have to buy one from GIK for WAF).  There is no back right corner as the back wall forms the opening of the stairwell.  This means I only have 3 main floor to ceiling corners with the exception of the short corner that is created in the step of the back wall adjacent to the sliding glass doors.  Here's an updated photo with some comments on the openings:

 

 

When/if we can determine what other types of panels I need on the "exposed" walls/ceiling, I'm willing to consider some professional looking panels from GIK (has anyone ever figured out a legal way to avoid copyright infringement as these panels used for first reflections would be ideal for some movie poster artwork!).

 

While budgetary constraints are always a concern, I'm of the opinion that investing in gear first (i.e. second sub before additional bass traps) is a better proposition as you know you can always use the gear no matter what room you are in.  Buying expensive (albeit less than new subs) bass traps to fit a custom space is obviously less than ideal especially if you are considering moving within the next 3-5 years (which I may).  That leaves investing in significantly less expensive DIY bass traps to help to fill the void between now and the sub upgrade with the possibility of using these same traps if I stay in this location or reconfiguring them in a new space.

post #768 of 9543
Quote:
Originally Posted by AustinJerry View Post

Is this perhaps why Jason said to multiply the distance by 1.13?

that is for (approximate) conversion from milliseconds (ms) into distance traveled (in ft) given a constant speed of sound in your room's medium.

that does NOT account for the relative accuracy of the data presented in the graph via that of the hardware loopback connection.

however, it doesn't mean you cannot ballpark the area and continue on with the string test. just understand the limitations of what you're working with.
post #769 of 9543
Quote:
Originally Posted by jevansoh View Post

Hi Nyal,

First I wish to thank you for responding in a civilized manor and without getting defensive.

I've read a lot of text books in my time, I've spent a lot of time on forums, and I've been involved in different aspects of business for awhile now, but no matter how much one reads and learns, experience counts the most in my opinion.

Based on what you stated in the text above, it seems as though I may have less than 10% of your real world experience at this point in my life and career, however, I do believe strongly in continuing education, gaining more experience both for knowledge sake and in that it means my bottom line continues to grow (I have work).

I'm sure over time my ideals may change a bit as I continue to learn and grow and I hope I never become a crotchety old man stuck in his ways.

But for now, I only have the education and experience I've gained so far to go on and have to respectfully continue to disagree with just a few of your ideals regarding the targets for 2-channel audio expressed previously.

I do respect where you're coming from and hope you understand that while I have some different viewpoints on certain aspects of what you believe, I don't know that either of us could say the other is "wrong" as all it boils down to is what make the customer happy.

For this thread I want to introduce folks to standard ways of doing things that aren't challenged by most and even if certain goals aren't totally met by each person, I know that nothing I'm teaching is going to make their rooms sound worse. There will always be room for improvement, but at least I'm doing no harm.

I'm sure through your experience you have found numerous reasons to come up with your own way of doing things and go down your own path, so to speak. At this point though, with my level of experience and knowledge, I simply feel more comfortable in teaching what I've learned and know works and for the most part is widely undisputed.

So I guess what I'm saying is, welcome. There is room for many different theories on what is "right" but the most important thing in my opinion is for one to decide what that path is going to be and follow it. If goals are set and worked towards to achieve them by passionate audiophiles, they will simply learn a lot along the way and enjoy their systems that much more and that enjoyment and knowledge gained will last a lifetime.

I'm glad to learn there is a typo in your document as quite frankly I mainly took issue with two of your points, one being a +/- 10db window is a "goal" and is acceptable, however, now realizing that is supposed to read +/- 5db I wholeheartedly concur. We are on the same page.

Where we remain opposed however is in regards to reflections. I never once stated all reflections are bad. In fact, I believe they are absolutely necessary.

I do believe that high gain early reflections are absolutely detrimental. I believe they cause smearing and muddiness as I've heard it myself. I also believe that nobody likes to listen in rooms approaching anechoic conditions therefore reflections after a certain termination point are necessary and they create the psychoacoustic effect of making the room seem larger than it is. Therefore, reflections become necessary, after a certain point and to a point, just not random uncontrolled reflections.

When creating a room that has a small period (12-25ms depending on the size of the room, budget, goals regarding focus on 2-channel or HT, etc) of low gain early specular reflections around -20db terminated by a high gain (around -12db or more) specular reflection and followed by a diffuse and infinitely decaying rate of further reflection, the direct sound will be heard, clear, no smearing or muddiness, and the room will sound spacious and live.

There are of course many other factors that if are not addressed will cause this effect to collapse, however, and I believe we are in agreement on most of them.

I don't think the goals for 2-channel and HT are all that different. Yes, there are differences, but let's face it, most rooms are multi-purpose including dedicated rooms as most folks (the folks reading this thread at least) don't have a dedicated 2-channel room and a dedicated HT).

I disagree that a room with an average 200ms decay time will not sound very dry, however. Maybe "dead" was a bit harsh, but 200ms sounds quite dry and you can't just fix it by adding a ton of diffusers and diffusers, by nature, also absorb and will reduce that further, plus too many diffusers, if not placed surgically and strategically in the room will cause other problems such as interfering with the early reflections you worked so hard to tame, lobing issues depending on the design, etc. If the diffusers are broadband then the room has to be large enough, too. So just placing a ton of diffusers all around a room with a 200ms decay time is NOT an answer IMHO.

As for the rest, we're pretty much on the same page.

I like to target +/- 5db with no EQ - full band. +/- 3db with EQ - full band.

Depending on the size of the room I like to target between 300-500ms and quite honestly will accept 250ms in smaller harder to control rooms and seldom reach much over 400ms in all but the largest of rooms. The target is always 10% but I'll be the first to admit I've yet to achieve it. 20% is more reasonable and I cannot disagree with your 25% recommendation fully, however, I do believe that should be the absolute limit.

With those targets, it's not necessary to state that the bass can decay slower than the specular region other than to say it is common for the lower frequency range to be the hardest to tame and it will take the most skill, time, and budget to get in line without affecting the specular region negatively. Why create two different targets then also state you need to keep the full band decay time within 25% relative?

So again, we may belong to some of the same clubs, pay membership dues to some of the same places, and have similar ideals with but a few differences, but one thing is sure and that is our experience differs by a great amount with you simply having more of it.

I hope we can speak again and would like to get to know you better.

I hope you can at least see that there is room for my philosophy and yours, and that the important thing for folks trying to learn is that they understand the differences, pick a model (yours, mine, or other) and start working towards it to accomplish true audio nirvana!

I hope our paths will cross again sometime, Nyal.

--J

Hey, no offense taken smile.gif

Writing the white paper with Jeff Hedback was a great learning experience for both of us. It helped us clarify and distill our experiences and thoughts based on real world experience into something tangible. We'd be fools to think we'd get it 100% perfect and could please everyone on the first iteration (if ever, come to think about it!). The white paper is a good 'stake in the ground' and both Jeff and I thought that it would be useful both to ourselves and also to the community at large. From the feedback we have received that seems to have held true, with some discussion about the use of RC instead of NC for noise measurement and also the usual disagreement on the subject of reflections (see below).

With respect to the 2nd main point you disagreed with (reflections) I wrote a follow up to our white paper on Hi-Fi Zine that more clearly illustrates our position. I agree that in general high gain specular reflections are bad, smearing images and reducing clarity. The only exception to this sentence would be for the lateral reflections, where I am in the Toole camp. Setting a target of -20dB for reflections as measured on an ETC is overly simplistic and does not reflect (in my opinion) our latest understanding of psychoacoustics.

I can speak from experience that a room with 200ms decay (as measured by T20/T30) through the midrange does not sound dry if the room is small and diffusion is extensively used. A theater room I did recently measured at 200ms and was not too dead. The side walls were left reflective, the whole ceiling and rear half of the side walls were covered in diffusers and we used slat diffusers on the back wall in front of bass traps. On the flip side I've worked in a home theater room with a 200ms number where it sounded overly dead. There was a lot of thick stretch fabric in that particular room. So it depends...and that's why we just gave a range, as a single figure doesn't tell you much.

On the subject of different decay targets for <250Hz and >250Hz That can be improved in iteration 2 and I definitely see your point. Would you suggest abandoning the T20/30 measurements altogether and simply moving to a spectrogram? The nice thing about T20, T30 is that you can octave or one third octave smooth them, which I feel is more useful for >250Hz whereas the spectrogram is pretty high resolution in the frequency domain which is useful in the bass but not so much elsewhere. I think we can agree that decay time should be relatively consistent across the audio spectrum with perhaps a rise in the lowest frequencies.
post #770 of 9543
Quote:
Originally Posted by jkasanic View Post

While budgetary constraints are always a concern, I'm of the opinion that investing in gear first (i.e. second sub before additional bass traps) is a better proposition as you know you can always use the gear no matter what room you are in. Buying expensive (albeit less than new subs) bass traps to fit a custom space is obviously less than ideal especially if you are considering moving within the next 3-5 years (which I may). That leaves investing in significantly less expensive DIY bass traps to help to fill the void between now and the sub upgrade with the possibility of using these same traps if I stay in this location or reconfiguring them in a new space.

It's even more complex if you have two subs already, as well as powered DefTech speakers (Mythos ST towers and CS-8080 center) that can provide powered 'mid-bass'. That's why I was a little leery of only measuring room response one speaker at a time+sub, except for waterfalls. There's a fill-in role that the DefTechs have that may help with what otherwise would be dips in the 80 to 200 Hz region. Maybe I'll "cheat" and at least for waterfalls, do L/R+subs, C+subs, and L/C/R+subs and compare pre/post Audyssey Pro cals..

I have my system all configured with HDMI output - thanks to AustinJerry for the how-to guide. I think I've have been lost without it! The trickiest part was getting the laptop to recognize the Denon 4311 (connected on the Dock HDMI port) as HDCP-capable. I had to download new Geoforce drivers. switch from Intel to Nvidia Graphics control, and work out a sequence to get my Panny VT50 display and 4311 to play nice with the resolution without creating flicker or disabling the laptop's video display. All trouble-shot now; I may produce that 'sounds of silence' waterfall for Jason's amusement from OmniMic before I get my real mic later this week (keeping fingers crossed).

FWIW, I'm in the same boat about permanency of the HT room---due to the "family expansion project", I've only got limited funds I'm going to devote to achieving "audio nirvana" in the current HT/living room outside of any new gear (I'm set except for adding calibration software and equipment, which are portable, if I value time more than money LOL for learning calibration skills). For one thing, in the fall that room might have some objects that create more reflections smile.gif. The odds are strongly deterministic that we'll be moving ourselves in the next 1 to 3 years....hence I'm looking at this thread as a tutorial for skills I can use more seriously in the future if I really jump down the rabbit hole with Jason, AustinJerry and Keith. In the new place we'll have the luxury of designing from the ground up...I can possibly sell treatments as part of the cost of prepping the living room with luck.
Edited by sdrucker - 1/29/13 at 11:38am
post #771 of 9543
Quote:
Originally Posted by AustinJerry View Post

 

Even though the treatments are not in a dedicated HT room, I am single and don't have WAF issues to deal with.

Somehow, I thought that was the case.

wink.gif

post #772 of 9543
Quote:
Originally Posted by Nyal Mellor View Post

The only exception to this sentence would be for the lateral reflections, where I am in the Toole camp.

this is the camp that says it is "a matter of taste", is it not?

and pardon me, but lol at lumping such design decisions into "camps".
post #773 of 9543
Quote:
Nyal Mellor]
LEDE was to me a product of a particular age of studio building and the use of particular speakers with poor off axis performance in that era.

does this imply that the LEDE total specular response and use of psycho-acoustics in its development are no longer valid with today's speaker technologies, given the design requirements for such a critically accurate reproduction space?

no one is implying LEDE is the appropriate response for all pleasuring listening spaces - that is up to the user to decide their own requirements, as some like the larger yet less-precise soundstage and smearing... but for accurate reproduction, is it somehow invalidated with speakers with good off-axis response?

per that of Peter D'Antonio, LEDE is the most comprehensively researched case of design yet! and it's funny the specific types of speakers D'Antonio called for! but that must have been missed in the cloud of history.

it is comical that NE response just doesn't seem to see this same type of criticism.
post #774 of 9543
Quote:
Originally Posted by AustinJerry View Post

Quote:
Originally Posted by jkasanic View Post

 

Here's a picture of the MKSound S150II's LCR with my AT screen removed (sub not shown but sits in right corner):

 

 

 

Joe,

 

May I ask your reasoning for placing the speakers on what look like tables?  Are you not concerned with the effect the flat surface will have in reflecting sound?  Did you consider more conventional speaker stands?

 

Jerry, I tried (ask Keith as he has the PM's to prove it) to find OEM stands for the S150II's when I was considering a setup similar to his (i.e. speakers sitting low to the floor under projection screen and pointing up towards MLP) but they are very hard to find as M&K discontinued their stands even before they went under and MK Sound never offered any for the SII's.  I just assumed that as long as I kept the front of the speakers overhanging the edge of each stand that this would mirror the advice given to many users in the Audyssey forum wrt their center channels sitting on AV stands?  Honestly, once I decided to go with a false wall and an AT screen, I never really considered using stands as I figured this was a much cheaper and equally effective solution behind the false wall?!  Did I overlook something or can you elaborate on the reflections I should be concerned with?

 

I suppose another alternative now that I don't really need pivoting stands (and since the height of the speakers can easily be adjusted by changing the height of the stands I have behind my screen), I could use any generic speaker stand that fits with my speaker cabinet?  I guess I would just like to understand the real benefit of doing this before I head down this path?

post #775 of 9543
Quote:
Originally Posted by localhost127 View Post

does this imply that the LEDE total specular response and use of psycho-acoustics in its development are no longer valid with today's speaker technologies, given the design requirements for such a critically accurate reproduction space?

no one is implying LEDE is the appropriate response for all pleasuring listening spaces - that is up to the user to decide their own requirements, as some like the larger yet less-precise soundstage and smearing... but for accurate reproduction, is it somehow invalidated with speakers with good off-axis response?

per that of Peter D'Antonio, LEDE is the most comprehensively researched case of design yet! and it's funny the specific types of speakers D'Antonio called for! but that must have been missed in the cloud of history.

it is comical that NE response just doesn't seem to see this same type of criticism.

Local I am not going to respond here because what normally happens is the thread gets hijacked. f you want to start a new thread on this be my guest smile.gif
post #776 of 9543
Quote:
Originally Posted by Nyal Mellor View Post

Local I am not going to respond here because what normally happens is the thread gets hijacked. f you want to start a new thread on this be my guest smile.gif

you're aware that you yourself made the initial comment?
post #777 of 9543
Quote:
Originally Posted by jkasanic View Post

 

Jerry, I tried (ask Keith as he has the PM's to prove it) to find OEM stands for the S150II's when I was considering a setup similar to his (i.e. speakers sitting low to the floor under projection screen and pointing up towards MLP) but they are very hard to find as M&K discontinued their stands even before they went under and MK Sound never offered any for the SII's.  I just assumed that as long as I kept the front of the speakers overhanging the edge of each stand that this would mirror the advice given to many users in the Audyssey forum wrt their center channels sitting on AV stands?  Honestly, once I decided to go with a false wall and an AT screen, I never really considered using stands as I figured this was a much cheaper and equally effective solution behind the false wall?!  Did I overlook something or can you elaborate on the reflections I should be concerned with?

 

I suppose another alternative now that I don't really need pivoting stands (and since the height of the speakers can easily be adjusted by changing the height of the stands I have behind my screen), I could use any generic speaker stand that fits with my speaker cabinet?  I guess I would just like to understand the real benefit of doing this before I head down this path?

 

I went back and looked at the picture that shows the front speakers, and now as I look more closely, I see that the speakers are at the edge of the tables.  When I first looked at the picture, I thought they were sitting back from the edge.

 

I suppose any number of stands would work.  I use these stands for my Wides.  The stands are filled with sand and are quite heavy and sturdy.  Another consideration might be a wall-mount, since it is behind the screen (but may be too close to the wall).  Like everything else we are doing, experimenting with different options and measuring the results is a good plan.  If you don't see any differences, I don't see why the current setup wouldn't be fine.

post #778 of 9543
Quote:
Originally Posted by Nyal Mellor View Post


Local I am not going to respond here because what normally happens is the thread gets hijacked. f you want to start a new thread on this be my guest smile.gif

 

+1.  I think this type of discussion is beyond the scope of the relative novices in this thread who are trying to develop basic measuring skills.

post #779 of 9543

 

Quote:
Originally Posted by AustinJerry View Post

Quote:
Originally Posted by jkasanic View Post

 

Jerry, I tried (ask Keith as he has the PM's to prove it) to find OEM stands for the S150II's when I was considering a setup similar to his (i.e. speakers sitting low to the floor under projection screen and pointing up towards MLP) but they are very hard to find as M&K discontinued their stands even before they went under and MK Sound never offered any for the SII's.  I just assumed that as long as I kept the front of the speakers overhanging the edge of each stand that this would mirror the advice given to many users in the Audyssey forum wrt their center channels sitting on AV stands?  Honestly, once I decided to go with a false wall and an AT screen, I never really considered using stands as I figured this was a much cheaper and equally effective solution behind the false wall?!  Did I overlook something or can you elaborate on the reflections I should be concerned with?

 

I suppose another alternative now that I don't really need pivoting stands (and since the height of the speakers can easily be adjusted by changing the height of the stands I have behind my screen), I could use any generic speaker stand that fits with my speaker cabinet?  I guess I would just like to understand the real benefit of doing this before I head down this path?

 

I went back and looked at the picture that shows the front speakers, and now as I look more closely, I see that the speakers are at the edge of the tables.  When I first looked at the picture, I thought they were sitting back from the edge.

 

I suppose any number of stands would work.  I use these stands for my Wides.  The stands are filled with sand and are quite heavy and sturdy.  Another consideration might be a wall-mount, since it is behind the screen (but may be too close to the wall).  Like everything else we are doing, experimenting with different options and measuring the results is a good plan.  If you don't see any differences, I don't see why the current setup wouldn't be fine.

 

Thanks for the quick response.  You had me worried there for a minute.  The 0.5ms reflections are a concern but until I get my UMM-6 mic and load the proper calibration file, I'll reserve changes to simply adjusting the L/R speakers inward and modifying the toe-in to avoid the screen frame or removing the grilles (although I can't remember exactly what these foam blocks do but I recall something about leaving the grilles installed as a result?!):

 

 

 Where's Keith when you need him?  Probably went to bed early! biggrin.gif

post #780 of 9543
Quote:
Originally Posted by AustinJerry View Post

@ Amir

So am I correct in concluding that there is no hope and we should all stop generating ETC graphs?
That would certainly be my recommendation. You guys have done an *amazing* job with that FAQ so far. It is the most high quality work I have seen on the topic yet. It would be a shame to a) derail the thread with the deep dive on the ETC and b) put thing in the FAQ that simply are not supported by research.

Let's get through the frequency response measurements below transition frequencies where psychoacoustic considerations are minimal to non-existent and hence, REW output is trustworthy. Once we get through that, i suggest we tackle T20/T30 as that can be a useful measure of how "live" the room is. If we still have time and energy left on our hands then smile.gif, we can revisit higher frequencies.

On waterfalls, that would be another area I would leave out. Quick and hopefully easier to understand version of that story is that frequency and time resolution are at opposite ends of the spectrum. Increase frequency resolution and you get lower time resolution and vice versa. Now think of you how you use the waterfall. You look at the time axis from back to front looking for "ringing." Once there, you "look up" at the top of the graph trying to find the offending frequency. Well, you can only do one or the other tasks I just mentioned with accuracy. If you have a lot of resolution in time where you clearly see ringing, then the top of the curves, the frequency response, will be too smoothed so you can't determine the precise frequency. If you re-run the graph so that the frequency response is precise, then the time domain will get soft and not show the ringing clearly! In my thread on WBF forum, I showed an in-between set of parameters that has decent resolution in both but even with that, it was easy to mis-guess the frequency of interest.

The best thing to do is to simply use the frequency response measurement and be confident that if you knock of the peaks you have done some good. Hope this is more clear smile.gif.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Audio theory, Setup and Chat

Gear mentioned in this thread:

AVS › AVS Forum › Audio › Audio theory, Setup and Chat › Simplified REW Setup and Use (USB Mic & HDMI Connection) Including Measurement Techniques and How To Interpret Graphs