caminperth Posted June 16, 2013 Posted June 16, 2013 Hi guys - I've recently acquired from Tyuen some framed panels using glass insulation. They have now been recovered in appropriately WAF acceptable fabric which I'm slowly getting used to... I've put two across the rear corners to act as bass traps and it has definitely cleaned things up a bit. The original plan was to put the rest of the panels along the walls to bring down the "ringing" in the room - clapping makes for some weird high pitched echoes - however now we have them ready I'm not so sure. Room dimensions are 3.5m wide x 5m long x 3m high. Listening position is unfortunately on the back wall of the room given the couch design. There is a long, thin window at the top of the front/TV wall and one behind the listening position. Gear is B&W 683, Emotiva XPA-2, Matrix Mini-i DAC & AppleTV. Panels to place are: 2 x 1.8m long x 0.9m rectangle 4 x 0.6m square Any thoughts on where I should place the panels to best effect?
Steve M Posted June 16, 2013 Posted June 16, 2013 Wow Cam! Funky fabric ...wife acceptable panels, is your wife a hippy?? ;-)) Steve. 1
Davey Willo Posted June 16, 2013 Posted June 16, 2013 Good god man turn off the light!!! My eyes my eyes!!!! Sorry mate, only joking, personally I would make sure that you have at least one pair at the first reflection points, I apologise if I'm telling you how to suck eggs here but just in case you don't know how to find the first reflection points, get your missus (yes the colour blind one) to move a mirror along the side wall until you can see your speaker in it from your listening position. As your room is so long and you sit so far back this might not be as simple as it could be but the above method is how you would find the FRP which is a quite important one as far as room treatments go. best regards Dave. p.s. Get the wife's meds changed ASAP
Guest Peter the Greek Posted June 17, 2013 Posted June 17, 2013 Wow....certainly adds some "colour" to what is otherwise a very conservative room Hard to suggest something without measurements....try the back wall as well
svenr Posted June 17, 2013 Posted June 17, 2013 Play recordings of single clicks or hand claps through the left and right speaker seperately and listen while seated. If the flutter echos remain, you should work on them. The most likely flutter echos exists between the upper protions of your front and rear wall. Then use the image source technique to find out where your first and second order reflection points are located. I recommend treating the ceiling reflections, but leaviing the first lateral reflections in tact to result in good apparent width and adequate listener envelopment.
caminperth Posted June 17, 2013 Author Posted June 17, 2013 Thanks for the comments guys. The print was a little brighter than anticipated once applied so is in the process of being changed out for something a little more neutral... One thing I did notice was that even with the panels sitting as shown in the photos, the sound was a little less "live" - if that sounds right? Dave - thanks for the info, I'll try the FRP's tonight and see how it sounds. Peter - I agree, just haven't purchased the gear yet. Any recommendations on a mic/software package that would suit a mac? Sven - I was thinking that the four square panels should be placed in the top corners of the room, do you think this would work? I also have a long skinny panel that I could put either above the window on the source wall or above the window on the wall behind the listening position - any thoughts on which would be best? Thanks again for the responses..
svenr Posted June 17, 2013 Posted June 17, 2013 The two windows make things a bit tricky. Try to cover as much reflective surface as possible. Hang them so to get about 150+mm of an air gap between the back and the wall - this will somewhat increase low frequency absorption. Think about some fancy shelves or other rara on the walls above the lateral reflection points to increase diffusion.
caminperth Posted June 20, 2013 Author Posted June 20, 2013 Thanks Sven - I'll do some playing around with panel placements over the weekend. Also, just ordered one of these - http://www.minidsp.com/products/acoustic-measurement/umik-1 - so will take some room measurements when it arrives... 1
Davey Willo Posted June 20, 2013 Posted June 20, 2013 Hey Cam I'm assuming that you are going to use REW for your measurements, here's a cool little video giving the basics of getting started. http://gikacoustics.com/room-eq-wizard-tutorial/ Apologies for the voice, within a few minutes you will want to strangle the man with the incredible weird accent, well I did but you may handle it better...
ArthurDent Posted June 21, 2013 Posted June 21, 2013 Also assuming you're going to be using REW you need the v5.01 BETA version in order for the UMIK to be automatically detected etc., available here... http://www.hometheatershack.com/forums/downloads-area/47460-v5-01-beta-downloads-asio-support.html
davewantsmoore Posted June 21, 2013 Posted June 21, 2013 Play recordings of single clicks or hand claps through the left and right speaker seperately and listen while seated. If the flutter echos remain, you should work on them. The most likely flutter echos exists between the upper protions of your front and rear wall. Then use the image source technique to find out where your first and second order reflection points are located. I recommend treating the ceiling reflections, but leaviing the first lateral reflections in tact to result in good apparent width and adequate listener envelopment. This. Treating the first reflection points can be right ... but is not a universal truth.... and the ceiling can give the biggest reward, but it difficult so often overlooked. Experiment lots. Oooo... I see you will. There will be lots of help if you can show data.
davewantsmoore Posted June 21, 2013 Posted June 21, 2013 bass traps If you are looking to fix real bass frequencies, it will be interesting to do extreme things like put all of them in one or two corner (bass traps need to be really really thick) ..... Not because that would be a reasonable way to arrange them (WAF, etc) .... but just to see what happens, how much difference it makes, etc.
caminperth Posted June 21, 2013 Author Posted June 21, 2013 Hey Cam I'm assuming that you are going to use REW for your measurements, here's a cool little video giving the basics of getting started. http://gikacoustics.com/room-eq-wizard-tutorial/ Apologies for the voice, within a few minutes you will want to strangle the man with the incredible weird accent, well I did but you may handle it better... Also assuming you're going to be using REW you need the v5.01 BETA version in order for the UMIK to be automatically detected etc., available here... http://www.hometheatershack.com/forums/downloads-area/47460-v5-01-beta-downloads-asio-support.html Thanks for the links guys - some weekend reading.... This. Treating the first reflection points can be right ... but is not a universal truth.... and the ceiling can give the biggest reward, but it difficult so often overlooked. Experiment lots. Oooo... I see you will. There will be lots of help if you can show data. If you are looking to fix real bass frequencies, it will be interesting to do extreme things like put all of them in one or two corner (bass traps need to be really really thick) ..... Not because that would be a reasonable way to arrange them (WAF, etc) .... but just to see what happens, how much difference it makes, etc. Dave - treating the ceiling may be on the cards, though will require some tactful discussions with the better half. As for data - I'm an engineer so this won't be a problem..
caminperth Posted June 24, 2013 Author Posted June 24, 2013 (edited) Well...... Received the mic in good time and condition today. The unit seems quite robust and of decent quality. After a crash course in setting up REW and downloading the cal file for the mic I started on some measurements - thats were it has now left me quite confused... I tried quite a few combinations and settled on a 256k, 0-300Hz sweep to check out the bass and the effects of the bass traps on the response. I tried one measurement each and there seemed to be a positive difference with the addition of one bass trap and a minor change with adding a second in front of the other. Like any reputable engineer, I decided to make a second set of measurements to check the repeatability/reliability of the results - and could not get a consistent result for any of the three conditions. It seems that one measurement taken is completely different to the next. So I tried a different tack - do 5 measurements and average them. This hasn't really worked as the outliers are so significant that they affect the overall results. I've posted results of measurements of the three conditions below for your amusement... Condition 1 - No bass traps Condition 2 - One bass trap Condition 3 - Two bass traps Using the averages of the results means that the traps do nothing. Using some of the results individually means they do a heap of good... Time to do some reading up on REW and to learn how to use it properly.. Edited June 24, 2013 by caminperth
svenr Posted June 25, 2013 Posted June 25, 2013 You need to perform a series of checks. In many cases the incoming mic signal is fed to the output again, creating a feedback loop with net gain less than unity so this failure is not obvious. I recommend starting with a 2channel measurement of the soundcard itself (no mic, outputs straight back to inputs) to check whether you get something close to a Dirac finction as the impulse response and then check what the frequency response of the soundcard is.
svenr Posted June 25, 2013 Posted June 25, 2013 (edited) So what - in the Windows mixer or other audio software running on the system you can still route the USB mic signal to an output creating the described problem. Have a look at all these beautiful comb filtering glitches. Swept sine excitation is particularly sensitive to noise present during the measurement. The workaround is to use very long excitation time (say 10sec for a single sweep) and /or quite a number of automatic averages to increase the signal-to-noise ratio (doubling of averages results in +3dB SNR). I imagine that with the relatively low sweep playback level used in a living room during the evening noise could be a real issue. MLS is much more robust to noise. And by noise I mean not only noise from the electronics but also environmental noise present in your house. Do you happen to live close to a major road, a power station, some big-ass air con heat exchangers or similar? Edited June 25, 2013 by svenr
davewantsmoore Posted June 25, 2013 Posted June 25, 2013 Yep, if you haven't already... backup a bit and try a loopback measurement, or a near-filed measurement of a woofer ... and lets see if that looks ok.
caminperth Posted July 3, 2013 Author Posted July 3, 2013 (edited) Thanks for the replies guys - work is getting in the way and I haven't taken any measurements or done any research on how to use this since last week.. You need to perform a series of checks. In many cases the incoming mic signal is fed to the output again, creating a feedback loop with net gain less than unity so this failure is not obvious. I recommend starting with a 2channel measurement of the soundcard itself (no mic, outputs straight back to inputs) to check whether you get something close to a Dirac finction as the impulse response and then check what the frequency response of the soundcard is. So what - in the Windows mixer or other audio software running on the system you can still route the USB mic signal to an output creating the described problem. Have a look at all these beautiful comb filtering glitches. Swept sine excitation is particularly sensitive to noise present during the measurement. The workaround is to use very long excitation time (say 10sec for a single sweep) and /or quite a number of automatic averages to increase the signal-to-noise ratio (doubling of averages results in +3dB SNR). I imagine that with the relatively low sweep playback level used in a living room during the evening noise could be a real issue. MLS is much more robust to noise. And by noise I mean not only noise from the electronics but also environmental noise present in your house. Do you happen to live close to a major road, a power station, some big-ass air con heat exchangers or similar? Ken is right - I'm using a USB mic with Apple OSX software on a 15" macbook pro. I've had a dig around and couldn't easily locate setting to provide loopback - though I understand the concept of checking actual output vs mic measurement. Sven - what does MLS stand for? I'm not near any major sources of noise, aside from the normal suburbia - A/C's etc.. Measurements were taken at a fairly high level - any louder and I'd be hitting against the WAF and would need some earplugs. Yep, if you haven't already... backup a bit and try a loopback measurement, or a near-filed measurement of a woofer ... and lets see if that looks ok. Drew - I'll give a near-field measurement a go, though not sure what this will show? Is it just another way to check if the mic is relatively accurate? Edited July 3, 2013 by caminperth
caminperth Posted July 3, 2013 Author Posted July 3, 2013 Tried to edit the above post... How loud should it be when measuring? I suppose for low frequencies quite loud and just cut the sweep at around 200Hz?
davewantsmoore Posted July 7, 2013 Posted July 7, 2013 I'll give a near-field measurement a go, though not sure what this will show? Is it just another way to check if the mic is relatively accurate? Yes. Put the mic less than 1cm from the speaker cone. You should get no holes in the response like the measurement you got earlier.... depending on your result, we can proceed from there.
davewantsmoore Posted July 7, 2013 Posted July 7, 2013 How loud should it be when measuring? I suppose for low frequencies quite loud and just cut the sweep at around 200Hz? It doesn't matter that much... depends what you want to know. 90dB is about right. Quieter is fine, but you may run into SNR issues... but even then, it won't matter a lot if you're just looking for the general response of the driver.
svenr Posted July 7, 2013 Posted July 7, 2013 (edited) Weird - I posted a reply days ago but it vanished - check this thread and the links given there: http://www.stereo.net.au/forums/index.php?/topic/53647-room-treatment-article-series-paul-spencer/ There is a good link to a page explaining the calibration. Edited July 7, 2013 by svenr
Recommended Posts