|
Post by mikewellington on Feb 19, 2019 18:32:56 GMT
Hi. Background:I have been in and out of a couple of small home studios over the past five years and when I have time I try to improve my understanding of acoustics as well as improve my mixing environment. I moved into a new place about a year and a half ago. I made my own bass traps and I bought the rest of my acoustic treatment from a small independent eBay business. I followed the room set up guides, correct listening position and so on. However as this location is only temporary I wasn't willing to go the whole way and create the acoustic excellence Ethan so carefully describes. I will definitely do this when I get my own place. I needed something portable, storable and relatively quick to put together as time is very limited. However I also need it to work. SO I went OTT and thought I could refine afterwards. This leads me to where I am now. Room:My room dimensions are: 5.5 meters long, 3 meters wide and 2.4 meters high. My bass traps & other panels:
I went for an adaptation of Ethan's more simple 'fill each corner' technique because as I said, I needed something that I could store and reuse. So I have 14 large bass traps: 1200x600x255mm (2 stacked on top of each other, floor to ceiling, in each corner in front of my listening position (behind the speakers) and then I literally created a wall at the back of the room leaving an air gap behind of 200mm, again stacks of two on top of each other (floor to ceiling) The only issue is I have to get in and out of this wall so there is an opening with a heavy curtain to the left hand side which leads to the door. There was no way around this. My design for the bass traps was a 25x35mm wooden shell filled with 3x900mm Knauf Earthwool Flexible Slab (density of 33kg/m3) Unfortunately this is all I could get my hands on as I live in the wilderness of Scotland and I couldn't get anyone to deliver Rockwool (we can't get Owens Corning over here). They are covered in a very thin transparent fabric. For the acoustic absorption coefficients I can only get access to 400mm and 1000mm info (not my 900mm) both of which are measured against a "solid backing" - I was advised that this could be plasterboard/typical indoor wall. I presume this is why Ethan emphasises that traps are securely placed against walls if not sealed to create an air lock.... Anyway, these are the coefficients for Knauf RS33 (Flexible Slab) Thickness 125Hz 250Hz 500Hz 1000Hz 2000Hz 4000Hz NRC 400mm 0.19 0.46 0.79 0.92 1.00 1.00 0.80 1000mm 0.57 1.00 1.00 1.00 1.00 1.00 1.20 So one would presume my 900mm would be close to the 1000mm My traps use three = 2700mm and the majority are in a snug position but not exactly sealed. Moving on: I also have sixteen large panels: 1240x640x50mm 9 on the ceiling 7 on the walls Wrapped in an acoustically transparent fabric covering a sound insulating slab (Rockwool - I don't have the exact model at hand so I can't tell you the coefficients), which absorbs low, mid and high-end frequencies. Open back design with a fabric rather than a wooden panel to maximize sound absorption , My experience with this setup:Generally this is probably the best set up I have had however there is something wrong with my lower frequencies and how I hear them. I have a feeling I have removed too much bass now (I wonder why...) So after making some measurements I decided to put in a graphic eq (Yamaha 2031B) just to see if it could make any difference to my listening position. This takes me to where I am now and my decision to start this thread as I know Ethan is the man to ask and I hope it will help others. Acoustic measurements:
I am measuring each speaker (Yamaha NS10M & Sub SW10) from my listening position using the USB UMIK-1 microphone, my studio audio interface and REW 5.20 Beta 5. Running my speaker output at 33dB setting REW to a scale of 50-110dB from 40Hz-20kHz with 1/12 octave smoothing. My latest results are as follows: LEFT & RIGHT WITH GRAPHIC OFF: LEFT & RIGHT WITH GRAPHIC ON LEFT WITH GRAPHIC ON & OFF: RIGHT WITH GRAPHIC ON & OFF: DEQ2496 RESULTS:LEFT & RIGHT LEFT, RIGHT & LEFT+RIGHT MONO (Green) Here are a couple of pictures to give you a better idea: BOTTOM: Conclusion:Generally I would be extremely grateful if you could help me establish the best course of action to fix my lower frequency issue and perhaps create a flatter response from 1kHz to 2kHz (I am considering getting a parametric for this). Many thanks. M.
|
|
|
Post by Michael Lawrence on Feb 19, 2019 19:46:44 GMT
Hi Mike, and welcome. (Good name!) The response dips and peaks around 70 Hz and 130 Hz are almost certainly modal in nature. You've done a good job of taming it with the GEQ, and your "sub bump" looks very much like the sub bump one might see on a large-scale sound system so I don't see an inherent problem there, unless razor flat is your goal. As a system tech, I will leave the acoustic treatment aspects to the other gurus around here, but you can definitely tune up your response pretty well with a parametric EQ. The industry best practice for system EQ has moved away from GEQ simply because it doesn't work very well, so you're right on the money in thinking a PEQ is a better tool for this job. If budget is a concern, I recommend the Behringer DEQ2496. It's a cost-effective unit that has way more EQ than you'll need. (It does a bunch of other stuff too, but the EQ will do the job for you.) I used to own one and have no complaints about the unit. I would place a filter at 2 kHz or so and bring it down maybe 4 or 5 dB with a low Q (maybe between 1.4 or 1, just eyeballing it) and then use a hi-shelf tuned down to 1 kHz and bring the whole upper range down by maybe 1 or 2 dB. Other alternatives include MiniDSP boxes which are great if you're a DSP nerd. For larger applications (more than 2 channels) I still use the Ashley Protea series. I showed a little bit of an optimization job I did with the Ashly here. Regardless of who makes the DSP, the end goal is the same.
|
|
|
Post by mikewellington on Feb 19, 2019 22:25:37 GMT
Hi Mike, and welcome. (Good name!) The response dips and peaks around 70 Hz and 130 Hz are almost certainly modal in nature. You've done a good job of taming it with the GEQ, and your "sub bump" looks very much like the sub bump one might see on a large-scale sound system so I don't see an inherent problem there, unless razor flat is your goal. As a system tech, I will leave the acoustic treatment aspects to the other gurus around here, but you can definitely tune up your response pretty well with a parametric EQ. The industry best practice for system EQ has moved away from GEQ simply because it doesn't work very well, so you're right on the money in thinking a PEQ is a better tool for this job. If budget is a concern, I recommend the Behringer DEQ2496. It's a cost-effective unit that has way more EQ than you'll need. (It does a bunch of other stuff too, but the EQ will do the job for you.) I used to own one and have no complaints about the unit. I would place a filter at 2 kHz or so and bring it down maybe 4 or 5 dB with a low Q (maybe between 1.4 or 1, just eyeballing it) and then use a hi-shelf tuned down to 1 kHz and bring the whole upper range down by maybe 1 or 2 dB. Other alternatives include MiniDSP boxes which are great if you're a DSP nerd. For larger applications (more than 2 channels) I still use the Ashley Protea series. I showed a little bit of an optimization job I did with the Ashly here. Regardless of who makes the DSP, the end goal is the same. Hi Michael. Thanks very much for your detailed answer. This is all very helpful and I look forward to others hopefully helping me with the modal issues. Re the parametric one of the staff members over at AV Nirvana advised this: Yamaha YDP2006 Parametric - you can have 6 bands per channel. Sounds good but looking rather hard to get hold of over here.... I had seen the MiniDSP but will take a closer look. I will report back to update. Thanks again - much appreciated.
|
|
|
Post by rock on Feb 20, 2019 14:13:17 GMT
Sounds and looks like you've done a very good job with treatment and your graphs show peaks and dips about 10dB so that's not bad at all. As Michael Lawrence points out, the problem is most likely modal. I have not done the math but using a mode calculator, you can find what room dimension is the likely culprit. If it's the width, you might try adding bass traps to your side wall/ceiling and/or side wall/floor corners (remember, a rectangular room has 12 corners, not just 4). Another place to experiment may be moving speakers and listening position. See this thread the-audio-expert.freeforums.net/thread/546/help-room-peak-130hz-bass
|
|
|
Post by mikewellington on Feb 20, 2019 18:12:46 GMT
Sounds and looks like you've done a very good job with treatment and your graphs show peaks and dips about 10dB so that's not bad at all. As Michael Lawrence points out, the problem is most likely modal. I have not done the math but using a mode calculator, you can find what room dimension is the likely culprit. If it's the width, you might try adding bass traps to your side wall/ceiling and/or side wall/floor corners (remember, a rectangular room has 12 corners, not just 4). Another place to experiment may be moving speakers and listening position. See this thread the-audio-expert.freeforums.net/thread/546/help-room-peak-130hz-bassThanks Rock, I will try a modal calculator. I'm glad I have finally reached this stage as I very much want to get to grips with this. Leave it with me and I will report back. Thanks very much for your help. Best, M
|
|
|
Post by Hexspa on Feb 21, 2019 18:03:27 GMT
Greetings, Michael. Indeed, it is a great name we share among us. Please allow me to give my analysis and advice.
The room's height is very close to half the wavelength of 70Hz which will cause peaks around 140Hz. This is extremely common in North America where ceilings are 8', 243.84cm, high. If the bass energy is seeing your walls as even 12.5cm larger than your stated dimensions, then it's clear that your length is causing the null at approximately 75Hz as that correlates to 5/4ths the wavelength which is a nulling dimension. In addition, if your measurement was taken 337.50cm from any large boundary, you will also experience nulling at 75Hz as that is 3/4 the wavelength - another null. The same goes for 124.47cm which is the 1/4 wavelength. This is known as Listener-Boundary Interference response, LBIR. Besides your measurement location, your speaker's distance from boundaries also causes interference known as Speaker-Boundary Interference Response, SBIR. I made a video about these interference responses which you can watch here. In short, while part of the issue is modal, the other is probably placement. Experiment with measuring different points. Indeed, it's best practice to measure in a reasonable radius - 30-122cm - from your ideal listening position. This is especially true before applying any corrective EQ. I hope this was accurate and helpful. Thanks
-----notes follow----- 69.30Hz = 497.87cm 73.42Hz = 469.92cm
/2 (peaks) (248.94cm) ~ 135Hz (234.96cm) ~ 146.83Hz
/4 (nulls) 1/4: 124.47cm 3/4: 373.40cm 5/4: 622.34
75Hz ~ 450.00cm
/4 (nulls) 1/4: 112.50cm 3/4: 337.50cm 5/4: 562.5cm
dimensions:
5.5m L (550cm) /4 (nulls) 1.375 (137.50cm) 1/4 ~ 255Hz 4.125 (412.50cm) 3/4 ~ 85Hz 6.875 (687.50cm) 5/4 ~ 50Hz 3m W (300cm) 0.75 (75.00cm) 1/4 ~ 460Hz 2.25 (225.00cm) 3/4 ~ 150Hz 3.75 (375.00cm) 5/4 ~ 90Hz 2.4m H (240cm) 0.60 (60.00cm) 1/4 ~ 550Hz 1.80 (180.00cm) 3/4 ~ 190Hz 3.00 (300.00cm) 5/4 ~ 115Hz
|
|
|
Post by mikewellington on Feb 23, 2019 22:57:11 GMT
Greetings, Michael. Indeed, it is a great name we share among us. Please allow me to give my analysis and advice.
The room's height is very close to half the wavelength of 70Hz which will cause peaks around 140Hz. This is extremely common in North America where ceilings are 8', 243.84cm, high. If the bass energy is seeing your walls as even 12.5cm larger than your stated dimensions, then it's clear that your length is causing the null at approximately 75Hz as that correlates to 5/4ths the wavelength which is a nulling dimension. In addition, if your measurement was taken 337.50cm from any large boundary, you will also experience nulling at 75Hz as that is 3/4 the wavelength - another null. The same goes for 124.47cm which is the 1/4 wavelength. This is known as Listener-Boundary Interference response, LBIR. Besides your measurement location, your speaker's distance from boundaries also causes interference known as Speaker-Boundary Interference Response, SBIR. I made a video about these interference responses which you can watch here. In short, while part of the issue is modal, the other is probably placement. Experiment with measuring different points. Indeed, it's best practice to measure in a reasonable radius - 30-122cm - from your ideal listening position. This is especially true before applying any corrective EQ. I hope this was accurate and helpful. Thanks
-----notes follow----- 69.30Hz = 497.87cm 73.42Hz = 469.92cm
/2 (peaks) (248.94cm) ~ 135Hz (234.96cm) ~ 146.83Hz
/4 (nulls) 1/4: 124.47cm 3/4: 373.40cm 5/4: 622.34
75Hz ~ 450.00cm
/4 (nulls) 1/4: 112.50cm 3/4: 337.50cm 5/4: 562.5cm
dimensions:
5.5m L (550cm) /4 (nulls) 1.375 (137.50cm) 1/4 ~ 255Hz 4.125 (412.50cm) 3/4 ~ 85Hz 6.875 (687.50cm) 5/4 ~ 50Hz 3m W (300cm) 0.75 (75.00cm) 1/4 ~ 460Hz 2.25 (225.00cm) 3/4 ~ 150Hz 3.75 (375.00cm) 5/4 ~ 90Hz 2.4m H (240cm) 0.60 (60.00cm) 1/4 ~ 550Hz 1.80 (180.00cm) 3/4 ~ 190Hz 3.00 (300.00cm) 5/4 ~ 115Hz
Thanks Hexspa, I really appreciate you taking the time to send over this info. I realised my NS10s were about 3" too low so I raised them. I don't really have time to do much more speaker placement tests unfortunately. The only thing I could try is moving the bass traps at the bottom of the room and perhaps repositioning my sub to a different location. Interested to know if you think this is worthwhile. Apart from this I took Michael's (Lawrence) advice and got myself a second hand DEQ2496 to try and fine tune my listening position. I will add the results below the others is my original post. I tried the RTA and auto eq but settled on my own graphic and parametric eq settings. The only issue I now seem to be facing are a few dips when measuring the stereo image. Perhaps these are phase issues...? approx 1.5kHz-3.5kHz + 7kHz-10kHz + 15kHz-18kHz Thanks again for everyone's help. It is great get all of your advice. Best, M
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Feb 24, 2019 10:22:11 GMT
Nice results with DEQ2496, remember FR is not the most important thing. It's easy to get over obsessed about it.
|
|
|
Post by Michael Lawrence on Feb 24, 2019 13:06:35 GMT
Mike-
Couple of ideas.
1) The peaks and dips from 3k and up in the combined response are very likely a comb filter caused by the mic not being EXACTLY centered between the two loudspeakers. You have a dip at 3k, a peak at 6k, a dip at 9k...your mic was about 2.25 inches off center. Move the mic sideways a bit and those will all change. You may see two arrivals on the Impulse Response tab as corroborating evidence.
2)This is a GREAT example of why AutoEQ leaves something to be desired. I had this conversation with a couple of big wigs in the optimization community not so long ago: computers lack context. They don't know if you're taking a nearfield measurement, or if you're in the back row of an arena. In both of those cases, a flat line would be an abnormal result, but the algos don't know that. An analyzer will always be a tool that gives a human more information to make a decision. So you did the right thing by using your brain instead of the algo.
3)Things are looking really good now, and I suspect sounding much better also. You equalized your sub bumb down to basically nothing. There's no wrong or right here, just questions of expectation. Listen to some music you know really well and make sure it sounds like you think it should. We as humans are used to systems with some amount of LF rollup - it's a natural consequence of being indoors - and so often when we don't hear that, things sound off. So I would say adjust to taste.
It would be interesting to see the response curve of the EQ you're using. You could capture this by taking an REW measurement through just the EQ unit itself, but don't feel obligated to take your whole rig apart just because we're curious.
Cheers.
|
|
|
Post by mikewellington on Mar 29, 2019 17:23:42 GMT
Mike- Couple of ideas. 1) The peaks and dips from 3k and up in the combined response are very likely a comb filter caused by the mic not being EXACTLY centered between the two loudspeakers. You have a dip at 3k, a peak at 6k, a dip at 9k...your mic was about 2.25 inches off center. Move the mic sideways a bit and those will all change. You may see two arrivals on the Impulse Response tab as corroborating evidence. 2)This is a GREAT example of why AutoEQ leaves something to be desired. I had this conversation with a couple of big wigs in the optimization community not so long ago: computers lack context. They don't know if you're taking a nearfield measurement, or if you're in the back row of an arena. In both of those cases, a flat line would be an abnormal result, but the algos don't know that. An analyzer will always be a tool that gives a human more information to make a decision. So you did the right thing by using your brain instead of the algo. 3)Things are looking really good now, and I suspect sounding much better also. You equalized your sub bumb down to basically nothing. There's no wrong or right here, just questions of expectation. Listen to some music you know really well and make sure it sounds like you think it should. We as humans are used to systems with some amount of LF rollup - it's a natural consequence of being indoors - and so often when we don't hear that, things sound off. So I would say adjust to taste. It would be interesting to see the response curve of the EQ you're using. You could capture this by taking an REW measurement through just the EQ unit itself, but don't feel obligated to take your whole rig apart just because we're curious. Cheers. Thanks Michael. Just getting back into this now. I'm not sure how to capture the EQ curve you mentioned - If you tell me how to go about doing this I'll give it a shot. Thanks again for all of your help, M.
|
|
|
Post by Hexspa on Mar 30, 2019 6:52:07 GMT
I think he means running the sine sweep from REQW through the EQ and directly back into the software - basically a loopback with the EQ inside. Similar to capturing a room - which itself is a sort of loop back - you won't be using speakers and a mic but simply a hardware device. In essence, an effects loop.
Hope that makes sense.
|
|
|
Post by mikewellington on Mar 30, 2019 13:58:20 GMT
I think he means running the sine sweep from REQW through the EQ and directly back into the software - basically a loopback with the EQ inside. Similar to capturing a room - which itself is a sort of loop back - you won't be using speakers and a mic but simply a hardware device. In essence, an effects loop. Hope that makes sense. Okay thanks Hexspa - I'll experiment and get back to you. M.
|
|