|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
100.14.55.26
In Reply to: RE: Polys revisited posted by emailtim on September 19, 2023 at 19:41:06
I'm using absorption at first reflection point on sidewalls and have known for some time it was less than ideal.
Any favorite source of yours for purchasing the polys? Those look nicely made. My room is only 14 feet wide. Based on your evaluation what would the optimal midpoint/radius? Do you have the link to a calculator that explains the optimal target dispersion?
Follow Ups:
The polys in the 1st picture above are in a studio, not mine, that I used to show their usual horizontal dispersion orientation. I made my own DIY polys based on information from the web and in-room testing.I used REW to identify my problem areas first and determine what frequency ranges to address. Based on the delay lengths and some debugging, I discovered what reflections were causing the issues. Thus, this was a targeted fix for a specific problem and not a random "catch-all" fix.
REW (vertical spectrogram plot) showed I had strong reflections ranging from [1-7]kHz at 4.5ms and 9ms which is in the range of smearing the direct signal versus being perceived as an echo.
Once they were identified as side and cross-side reflections, I realized it was also what was causing the sound stage to shift when tuning one's head because the leading ear would switch between the side and cross-side reflection. The polys fixed that leading ear sound stage shift. I also used REW to validate the results.
4ms OC-703 reflection shown here in red, blue afterward with poly.
My commercial broadband traps were 2'x4' so I made my DIY polys similar in size.
To get a 180 degree dispersion, you need a 1/4 slice of a pie, or 90 degree arc from a circle. Other dispersion patterns can be made by changing the shape of the arc.
Below is a link to an ARC/CORD length calculator. If you enter in the height of your material, say 48" (ARC) and then enter 90 degrees, it will give you a CORD of 43.2" inches which will be the vertical height of the poly or semicircle. Since 43" is longer than the [1-7]kHz frequency wavelengths (13" and 2" respectively), this size poly is suitable for the problem at hand.
Frequency Wave Length chart:
https://www.jdbsound.com/art/frequency%20wave%20length%20chart%202013.pdf
I have line arrays centered on the vertical height of my room. They do not exhibit much floor/ceiling bounce so that is not an issue. I centered the vertical dispersion polys to the height of my room so the 43" CORD covers both sitting and standing positions.
There are multiple DIY examples of making polys on the web. I tested with solid (non-perforated) surface and also cloth covered. Both worked according to REW before and after measurements.
Another thing about polys is that they do not distort/scatter phase which can be beneficial for certain applications. QRDs scatter phase and tend to work best in the back of the room.
.
2022/03/30 Historical Records CENSORED
Edits: 09/21/23 09/21/23 09/21/23 09/21/23
Your approach to this is really fascinating, but will take me awhile to digest.
I have REW and have used it, but have never seen a Spectogram plot of frequency v. delay (I'm using an old version). I would have thought those 2 little blips at 4.5 & 9 ms were benign! How did you identify them as side and cross side reflections? I see the 2nd graph (dbFS & distance) and surmise that corresponds to side and cross side distances plus your observations with "head turning".
Just so I'm clear, the chord length corresponds to the vertical height of the poly? I initially thought that the chord/arc was the horizontal width of the poly based on your original studio room picture. I'm guessing you did this for standing and seated positions, but that they would work oriented the other way?
I did not realize that polys do not alter phase unlike QRDs. I'm not sure whether that is good or bad if I'm trying to diffuse side wall reflections.
Anyway, you've given me (too?) much to think about and convinced me to upgrade my ancient version of REW. Thanks again for this!!!!
The CORD is flat surface between the end-points of the ARC facing the wall. The ARC is the curved reflective surface facing in the room.Just to be clear, I mounted them horizontally (in a single centered line) to disperse vertically like depicted on the right side of the picture.
Normally, they are mounted vertically (in multiples) to disperse horizontally (left side of picture) which causes unwanted focused lens effects instead of even dispersion patterns.
My goal was to attenuate AND delay the 1st reflection by dispersing it across a 180 degree reflection pattern in addition to lengthen its travel path (delay its arrival simulating a wider room) by turning portions of the dispersion into 2nd, 3rd, 4th, ... order reflections.
REW will show the dispersion as redistributed lower amplitude peaks indicating the energy is still in the room but decentralized pushed lower into the decay field.
The upper right reflections depict dispersed SIDE-WALL 1st reflections into Nth order reflections.
The lower right reflections depict dispersed CROSS-SIDE-WALL 1st reflections into Nth order reflections.
Similar dispersed patterns are mirrored for the left wall.
Keep in mind this is going on in 3D, not just 2D as pictured so the front to back directions are not pictured.
.
2022/03/30 Historical Records CENSORED
Edits: 09/22/23 09/22/23 09/22/23 09/22/23 09/22/23 09/22/23 09/22/23 09/23/23
The Vertical Spectrogram is basically a topological chart (mountain ridges and valleys) that you can adjust for depth contour detail.It is very helpful because it shows the frequencies constituting the reflections where the bottom chart does not. Both are useful, but for different details.
The bottom chart shows you the exact dB level of each peak, but not what frequencies are causing it. This one is only showing the 4ms peak, not the 9ms peak which is off the right side of the chart. I started testing with 1 poly, so could only test 1 reflection at a time, not all 4 reflections.
The Vertical Spectrogram also details time alignment of your multiple drivers and subs (black dashed line). People who align their subs by other means would be surprised to view this plot after thinking they dialed in their subs. It is well worth getting familiar with.
I would have assumed the same as well, but [1-7]kHz includes the most sensitive range in human hearing so it is a critical region. Ideally, your room reflections should be down 20dB within 20ms (4 and 9ms are well within that range). The 2nd plot easily shows this range (amplitude and time) as well as which peak time is the worst offender (1st to treat).
The closer the reflections are to the direct sound, the more they smear the clarity. In my scenario, it also destabilized the soundstage by mixing the channels.
I blocked various sides of the microphone (with a tripod and heavy folded quilt), remeasured and compared plots to identify where the reflections were coming from using the 4 and 9 ms as hints at which reflections to try identifying first.
After the reflection sources were identified as cross and cross-side, I realized as I twisted my head, the leading ear would cross the tangential of one channel's reflection and then the other channel (another variant of "head in vice", but caused by the room, not the speakers).
On tangent would dictate the soundstage. Switching between the left and right channel with the leading ear explained the sound stage shift. The tangential effect could further be exaggerated by cupping the leading ear (helpful tool) making it more directional.
According to some studio designers and anechoic measuring engineers, polys have a natural sound and specify QRDs for the rear because of the phase scattering. QRDs can also make time aligning speakers harder because the measurements show altered phase from the QRDs.
John/REW has a beta thread with weekly releases of the newest REW development. I recommend grabbing the latest development version which has a lot of nice new features/bug fixes and found it to be quite stable.
.
2022/03/30 Historical Records CENSORED
Edits: 09/22/23 09/22/23 09/22/23
Your posts always contain great information. I especially appreciate any REW info and how to interpret the results. Now all I have to do is get around to downloading the new version of REW and working with it....
My next project is partially unmodding a few TX2575 resistor changes I made to my ARC preamp awhile ago. Turned out to be a little too much of a good thing on some recordings as I listen with a new CD/SACD player.
Thank you again.
Here is a link to the latest releases.
Currently 5_20_14_72
.
2022/03/30 Historical Records CENSORED
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: