|
Post by Ethan Winer on Aug 1, 2016 20:36:06 GMT
I've never experimented with angled panels, but I know for certain that putting panels parallel to the ceiling works perfectly well.
--Ethan
|
|
|
Post by rock on Aug 1, 2016 21:27:24 GMT
FWIW, in some high school choir and orchestra rooms I have seen, I notice they have installed panels from the ceiling about 1" thick at an angle that look like they would reflect away from the source. They don't appear to be very absorbent but IDK. These are fairly large rooms and flutter echoes are probably what they are trying to control.
Cheers, Rock
|
|
|
Post by Hexspa on Aug 1, 2016 22:13:44 GMT
I think if you're using fluffy you should go minimum eight inches thick.
|
|
|
Post by Hexspa on Aug 4, 2016 2:03:27 GMT
Thanks guys, So grateful for all your opinions, especially during these time! Well, really interesting, but no energy to change them, have finished half fluffy cloud… I will increase the thickness, then do the measurement. If the result not well, I might try some method... I will upload my test later. : ( hard hard hard… W. I only found the building phase to be fun the first time. Not so much the subsequent iterations. The result is worth it tho. -m
|
|
|
Post by Hexspa on Aug 5, 2016 21:32:11 GMT
Well, seems strange you'd get a curve similar to pink noise. Did you use a sine sweep?
It's also weird that those waterfall graphs are inverted. If you can use the regular ones that'd be better.
Maybe post a screenshot of your sweep settings.
|
|
|
Post by Hexspa on Aug 6, 2016 7:57:00 GMT
Ok, thanks. Looks like you are doing a sine sweep. I looked at AMROC and your room is a little small. Given that you have concrete walls you'll definitely suffer maximally due to reflections. Here are your results: amroc.andymel.eu/?l=16.08&w=10.17&h=9.09&ft=true&r60=0.6Now, predictably we are seeing a peak around 35Hz. You should have a peak at 53Hz but your length is defeating it by having 3x a quarter wavelength thus creating a relative null. I'm taking a rough guess that your ceiling is causing your null at about 110Hz. So, your ceiling is half treated and you said you did your rear wall. But there's one thing that I can't explain: why does your response look like pink noise? You can see that, sans your peaks and nulls, your response is almost identical. I have to assume it's your speakers. Your bass is almost 30dB above 1k and your highs dip out about 12dB from there. I, in my limited experience, have never seen anything like it. By looking I'd think your tweeters are blown and your sub is turned up way too high. Looking through your thread you never mentioned which speakers you use. Why don't you do a close measurement of each of your speakers? If their response is skewed no amount of treatment will get you a linear response. You also didn't post a conventional waterfall so I can't help you regarding decay because I can't interpret that graph. Thanks, -m
|
|
|
Post by rock on Aug 6, 2016 13:03:45 GMT
Hi Wei, I think your big problem is that your measurements are making you second guess your treatments. I agree with m that you should verify your test procedures and speaker response with close miking your speaker. With close micing, you effectively eliminate your room response and see only your speaker response which should look similar to your speaker spec sheet graph if you have one. After you get a good response with close miking, you can make more measurements at incrementally farther distances from the speaker into the room. This should show you the gradual increase in room modes as speaker/mic distance increases. For your low frequency graphs, turn the smoothing off as you want to see all the detail. For the most part, you will be looking at LF response when analyzing room modes so limit your graphs to 300-400Hz. For wide band graphs using smoothing is good because all the peaks and nulls in the upper freqs are unavoidable. Your primary focus right now should be to gain confidence in your measurements. Without that, you'll be forever running in circles. Also experiment with speaker and mic position as that can affect the peaks and nulls too. As far as your absorbers, you mentioned earlier "I used plastic bag with no more than 1mm thickness to package the OC703, then used cotton sheet to package them and install." The 1mm plastic should be attached to only the front; spray glue works good. It's not clear from your description but if you did that already, great. If you have pictures and or diagrams of your room, that may help. Sorry for asking if you have posted them on a previous thread. I'll look to see if I can find them. As you requested, here's a link for Helmholtz absorbers www.mh-audio.nl/acalculators.asp#showcalc but I really think you need to confirm your measurements and procedures before you abandon porous absorbers; they are the most practical solution for small rooms and I don't think there is any reason they won't work for you too. Cheers, Rock
|
|
|
Post by rock on Aug 6, 2016 16:54:05 GMT
Wei,
Just to be clear, when I say limit to 300 - 400Hz, I mean on the high end of the graph. The low end should start at about 20 to 30 Hz. In actuality, 30 to 300Hz is probably good enough but if you play with the values, you should see what is most useful.
Cheers, Rock
|
|
|
Post by Hexspa on Aug 6, 2016 22:59:50 GMT
Correct way to display response graphs: the-audio-expert.freeforums.net/post/1028 Correct way to display waterfalls and decay plots: the-audio-expert.freeforums.net/post/1029^ Maybe you can give a little more decay range to the waterfall (30-35dB from peak) and I have never seen Ethan use a decay plot but I find them useful complements to the waterfall. I'm pretty sure Ethan will advise against resonant absorbers because, and I paraphrase, "small rooms need as much broadband absorption as possible. Tuned traps not only take up valuable space but only treat ONE frequency and not multiples thereof. Besides, they're expensive and hard to build. Therefore stick to broadband absorption - especially for small rooms." Besides, resonant absorbers are targeted and your response is currently nowhere near needing fine-tuning like that, imo. -m
|
|
|
Post by Hexspa on Aug 7, 2016 20:34:17 GMT
Yeah turn that off.
The waterfall in FuzzMeasure is under the Plugins menu though you can only use it in the premium version.
I'd really recommend getting away from Fuzz Measure. It's pointlessly expensive. REW is great and for Mac now (I don't think it always was).
Wei, you have to turn off smoothing. You have it set to 1/12th octave. You're only hurting yourself by leaving it on - especially below your schroeder frequency.
My brain is getting overwhelmed but I think your ceiling and side walls are both contributing to that null above 100Hz. If you've treated your rear wall but not your sides or ceiling it would make sense because otherwise you should have modal support at that frequency from your rear wall.
Other than that looks good. We're typically advised to practically ignore anything below 50Hz so your peak at 36Hz (room length I think) can be ignored for now.
Pretty much if you use 4" with up to 12" gap/12" fluffy all around you should be fine.
Don't ignore waterfall graphs though - some say they're more important than actual response graphs.
Glad you worked out your issue.
-m
|
|
|
Post by Ethan Winer on Aug 11, 2016 21:32:37 GMT
Forget tuned bass traps. The way to improve low frequencies is with more bass traps in more places, and or making them thicker.
--Ethan
|
|
|
Post by rock on Aug 12, 2016 13:36:44 GMT
Hi Wei,
Just out of my curiosity, can you post wide view photos of your room from a few angles so we can get the "big picture"?
Thanks, Rock
|
|
|
Post by Ethan Winer on Aug 12, 2016 14:34:25 GMT
Yes, a big picture would help.
Once you get to a foot thick or more you can use fluffy insulation which costs less than 703 and is easier to find locally.
|
|
|
Post by Hexspa on Aug 12, 2016 23:26:51 GMT
Yes, a big picture would help. Once you get to a foot thick or more you can use fluffy insulation which costs less than 703 and is easier to find locally. So I uploaded, and edited, don't judge me… it's not beautiful, because not finish yet…. This is the rear wall in the middle. And the two corner is the big corner trap, and I didn't fill them yet, because I don't know how to fill them is right, to install the second corner basstrap inside with some air gap, or, just filling…? And the right part of the rear wall, it's a moving 'door', because of my room structure's limitation, I only allow to open a door there… pain. Ethan, if I used 703 to fill the inside instead of fluffy one, is that ok? Because I have so many left.. I might to change all the rear wall, to get rid of two big corners, and re-build as a big rear wall, to do that cloud array one, I don't know.. is that necessary? According to the frequency response, there is a huge deep around 105hz and a peak at 37hz around, is that ok? Scared... Wei, I already told you about your null and ringing but now you've asked twice about it.
|
|
|
Post by Hexspa on Aug 12, 2016 23:27:44 GMT
My brain is getting overwhelmed but I think your ceiling and side walls are both contributing to that null above 100Hz. If you've treated your rear wall but not your sides or ceiling it would make sense because otherwise you should have modal support at that frequency from your rear wall. We're typically advised to practically ignore anything below 50Hz so your peak at 36Hz (room length I think) can be ignored for now.
|
|