Lsc
Emo VIPs
Posts: 3,348
|
Post by Lsc on Apr 10, 2020 21:42:32 GMT -5
REW can do all sorts of interesting things... but it's main benefit is in measuring frequency response and calculating correction filters for it. REW measures each channel separately... one at a time... And then calculates "standard PEQ-based corrections" to adjust the frequency response. Just use the "Distance" settings on the processor to adjust the relative timing.
(Trying to control and rely on things like the latency in your computer audio drivers is really complex... and unnecessary.) If you're really obsessive about the timing... here's something you could try. Record a monaural audio file with a single sharp loud "tick" sound on it... like a metronone.
Play that file through your processor in All Stereo. At your prime listening position the tick should sound sharp and clear... with no doubling... and no echoes. If you hear multiple ticks, or a sort of blurry tick, or echoes... then adjust your distance settings until the ticks sound sharp and clear. You could put a microphone at your listening position and look at the sound on your computer screen with your favorite audio editor.
The tick is PLAYING at exactly the same time on all your speakers. Therefore, when the sound is arriving from all of them at the same time, you'll see a single spike on the screen... and your speakers are time aligned. (Due to the room, and your speakers, you won't get it perfect... but, the closer, the better.)
I've attached a little bit of advice about using REW with your processor... NOTE:
While all of the other channels work properly... the manual PEQ settings in the subwoofer channel do not work in firmware v1.9 and some previous versions. This has been addressed in beta versions of the firmware and will be working properly in the next release. This is an area I know just enough to be dangerous... And most the time I am wrong. After thinking through it, I'd say the main issue would just be having your speakers in proper time alignment. This would be more critical on speakers that roll off to your subs, but most important on LCR I would think. Rew allows you to specify a timing speaker as well. I understand a lot of this when I'm watching videos, but putting it into practice is another thing. Thanks Keith! I’m doing the frequency correction using REW. It wasn't too bad changing out the XLR cables to measure my Width and Height speakers, but I had to change the XML name in each of these files. I entered the distances once. Hopefully this is the last time I will be using REW until Dirac.
|
|
|
Post by megash0n on Apr 11, 2020 8:56:27 GMT -5
REW can do all sorts of interesting things... but it's main benefit is in measuring frequency response and calculating correction filters for it. REW measures each channel separately... one at a time... And then calculates "standard PEQ-based corrections" to adjust the frequency response. Just use the "Distance" settings on the processor to adjust the relative timing.
(Trying to control and rely on things like the latency in your computer audio drivers is really complex... and unnecessary.) If you're really obsessive about the timing... here's something you could try. Record a monaural audio file with a single sharp loud "tick" sound on it... like a metronone.
Play that file through your processor in All Stereo. At your prime listening position the tick should sound sharp and clear... with no doubling... and no echoes. If you hear multiple ticks, or a sort of blurry tick, or echoes... then adjust your distance settings until the ticks sound sharp and clear. You could put a microphone at your listening position and look at the sound on your computer screen with your favorite audio editor.
The tick is PLAYING at exactly the same time on all your speakers. Therefore, when the sound is arriving from all of them at the same time, you'll see a single spike on the screen... and your speakers are time aligned. (Due to the room, and your speakers, you won't get it perfect... but, the closer, the better.)
I've attached a little bit of advice about using REW with your processor...
NOTE:
While all of the other channels work properly... the manual PEQ settings in the subwoofer channel do not work in firmware v1.9 and some previous versions. This has been addressed in beta versions of the firmware and will be working properly in the next release. This is an area I know just enough to be dangerous... And most the time I am wrong. After thinking through it, I'd say the main issue would just be having your speakers in proper time alignment. This would be more critical on speakers that roll off to your subs, but most important on LCR I would think. Rew allows you to specify a timing speaker as well. I understand a lot of this when I'm watching videos, but putting it into practice is another thing. Thank you for the document Keith. I do have a difference in opinion on the timing importance. Any speakers that need to be in alignment, like fronts to subs for example, will have significant phase issues around the crossover point, and each harmonic there after if applicable. I will agree that you can "just use the distance controls in the RMC", because that is all we've been given to work with. In order to use these controls, one must do math to convert the time delay to a distance number. Or, just "guess". REW and other tools provide many features that allow you to figure these things out. It had built in timing chirps. I don't know how one would listen to a chirp played in All Stereo and distinguish between a time delay issue or sound reflection/ reverberation. You'd have to use REW again to rule those things out so you knew it was a time issue or a reflection issue. Those distance numbers in the processor aren't the for you to use a tape measure. Maybe if we were outside with no walls.. This advice would work better, but I respectfully disagree with concept that timing isn't that important.
|
|
|
Post by megash0n on Apr 11, 2020 9:17:14 GMT -5
REW can do all sorts of interesting things... but it's main benefit is in measuring frequency response and calculating correction filters for it. REW measures each channel separately... one at a time... And then calculates "standard PEQ-based corrections" to adjust the frequency response. Just use the "Distance" settings on the processor to adjust the relative timing.
(Trying to control and rely on things like the latency in your computer audio drivers is really complex... and unnecessary.) If you're really obsessive about the timing... here's something you could try. Record a monaural audio file with a single sharp loud "tick" sound on it... like a metronone.
Play that file through your processor in All Stereo. At your prime listening position the tick should sound sharp and clear... with no doubling... and no echoes. If you hear multiple ticks, or a sort of blurry tick, or echoes... then adjust your distance settings until the ticks sound sharp and clear. You could put a microphone at your listening position and look at the sound on your computer screen with your favorite audio editor.
The tick is PLAYING at exactly the same time on all your speakers. Therefore, when the sound is arriving from all of them at the same time, you'll see a single spike on the screen... and your speakers are time aligned. (Due to the room, and your speakers, you won't get it perfect... but, the closer, the better.)
I've attached a little bit of advice about using REW with your processor... NOTE:
While all of the other channels work properly... the manual PEQ settings in the subwoofer channel do not work in firmware v1.9 and some previous versions. This has been addressed in beta versions of the firmware and will be working properly in the next release. Thanks Keith! I’m doing the frequency correction using REW. It wasn't too bad changing out the XLR cables to measure my Width and Height speakers, but I had to change the XML name in each of these files. I entered the distances once. Hopefully this is the last time I will be using REW until Dirac. Do you have any of your speakers set to small? If so, then you are integrating your subs, at some crossover point, with that speaker you set to small. Assuming that your LCR are set to small, you would want to start with proper sub integration with these speakers. There's a concept of integrating multiple subs together in the room, and then secondly, integrating that "virtual sub" with your mains. This almost certainly will have a timing or delay component. It is my opinion, one gathered by following numerous individuals who install home theaters and listening rooms for a living, that you should start here. After everything is "aligned" using the left channel for example, then you can take better measurements and apply filters. The kicker here is that the RMC only allows for a distance setting which has to be converted from a "time delay". There is some math, but you also can just keep playing with the distance until you pull a measurement that appears to be in alignment during your sub integration. REW has a few ways to actually show timing issues, but it's a bit over my head. I'm sure I'll get some negative commentary on something I've said here, but again.. I'll trust the people the get paid specifically for their knowledge on this matter.
|
|
Lsc
Emo VIPs
Posts: 3,348
|
Post by Lsc on Apr 11, 2020 11:24:28 GMT -5
Do you have any of your speakers set to small? If so, then you are integrating your subs, at some crossover point, with that speaker you set to small. Assuming that your LCR are set to small, you would want to start with proper sub integration with these speakers. There's a concept of integrating multiple subs together in the room, and then secondly, integrating that "virtual sub" with your mains. This almost certainly will have a timing or delay component. It is my opinion, one gathered by following numerous individuals who install home theaters and listening rooms for a living, that you should start here. After everything is "aligned" using the left channel for example, then you can take better measurements and apply filters. The kicker here is that the RMC only allows for a distance setting which has to be converted from a "time delay". There is some math, but you also can just keep playing with the distance until you pull a measurement that appears to be in alignment during your sub integration. REW has a few ways to actually show timing issues, but it's a bit over my head. I'm sure I'll get some negative commentary on something I've said here, but again.. I'll trust the people the get paid specifically for their knowledge on this matter. Great question and what you are saying makes total sense. I make my REW measurements simple as this is a stop gap until Dirac arrives. My goal is to just EQ up to 500hz for each speaker independently. There is no blending with subs in my quick REW run. I’m hoping Dirac BM will help me here. Using preset 2, I set all the speakers to large and I leave the distance to 0. I do take the time to equalize the volume level since that makes REW easier. After I do all my measurements, I load them to preset 1 in the XMC2. Then I set the crossovers and distances. Then just level the volume and I’m done. There is a clear improvement in my system after doing the REW. But I’m not doing anything fancy. Just a simple use of the PEQ.
|
|
|
Post by marcl on Apr 11, 2020 11:52:41 GMT -5
My two bytes on REW. First, thanks for the writeup Keith ... and good to see acknowledgment of the PEQ issue with subs. I was chasing my tail a while and ended up using my miniDSP for time alignment and PEQ. Works best with multiple subs anyway. I do look at Impulse Response and Phase in REW to double check time alignment and to be sure all is well at the crossover points. Right now I have front L/R set to Large with Enhanced Bass on and a 60Hz crossover. REW is essential to tweak this to get the best crossover from my Magnepan 3.7's to the subs/DWM.
And then there's ATMOS. My system is 5.1.4 and of course the only way to measure the height speakers with REW is to swap cables. When I did this I used Preset2 with distances and levels set, did all the height channel measurements using the Surround Left channel, and changed the distance and level values on that channel to match the required values for each height channel as I swapped the cables. Worked fine. But I swear I could not see consistent results loading the filters from REW to the XMC-2. What I measured after loading the filters did not make consistent sense, so I gave up on it.
Main thing is the range below 100Hz is handled by the miniDSP for now.
|
|
Lsc
Emo VIPs
Posts: 3,348
|
Post by Lsc on Apr 11, 2020 12:03:17 GMT -5
But I swear I could not see consistent results loading the filters from REW to the XMC-2. What I measured after loading the filters did not make consistent sense, so I gave up on it. This might be obvious to some...but when you swap cables you have to change the label in the XML inside the file since there is nothing in REW for the heights and wide speakers. For instance my frontWidth (leftWidth / rightWidth), I used my leftFront and rightRight to take the measurements. After I save the filter settings as leftWidth and rightWidth, I had to open the file and change the xml name as it says “leftFront or rightFront”. So I obviously changed that to “leftWidth or rightWidth”. Hopefully this is clear.
|
|
|
Post by megash0n on Apr 11, 2020 13:13:53 GMT -5
Do you have any of your speakers set to small? If so, then you are integrating your subs, at some crossover point, with that speaker you set to small. Assuming that your LCR are set to small, you would want to start with proper sub integration with these speakers. There's a concept of integrating multiple subs together in the room, and then secondly, integrating that "virtual sub" with your mains. This almost certainly will have a timing or delay component. It is my opinion, one gathered by following numerous individuals who install home theaters and listening rooms for a living, that you should start here. After everything is "aligned" using the left channel for example, then you can take better measurements and apply filters. The kicker here is that the RMC only allows for a distance setting which has to be converted from a "time delay". There is some math, but you also can just keep playing with the distance until you pull a measurement that appears to be in alignment during your sub integration. REW has a few ways to actually show timing issues, but it's a bit over my head. I'm sure I'll get some negative commentary on something I've said here, but again.. I'll trust the people the get paid specifically for their knowledge on this matter. Great question and what you are saying makes total sense. I make my REW measurements simple as this is a stop gap until Dirac arrives. My goal is to just EQ up to 500hz for each speaker independently. There is no blending with subs in my quick REW run. I’m hoping Dirac BM will help me here. Using preset 2, I set all the speakers to large and I leave the distance to 0. I do take the time to equalize the volume level since that makes REW easier. After I do all my measurements, I load them to preset 1 in the XMC2. Then I set the crossovers and distances. Then just level the volume and I’m done. There is a clear improvement in my system after doing the REW. But I’m not doing anything fancy. Just a simple use of the PEQ. I gotcha. I often find I spend hours digging into the details to end up frustrated because I feel like I'm missing something. REW is a powerful tool. I'm with you though, I just kinda gave up learning more and more in hopes that Dirac will be here soon. If I were going to be a professional, I would probably continue learning, but I'll end up forgetting everything I learn today by the time I mess with it again in 6 months. I'll probably hire a professional once I finish a few more projects.
|
|
|
Post by marcl on Apr 11, 2020 13:38:49 GMT -5
But I swear I could not see consistent results loading the filters from REW to the XMC-2. What I measured after loading the filters did not make consistent sense, so I gave up on it. This might be obvious to some...but when you swap cables you have to change the label in the XML inside the file since there is nothing in REW for the heights and wide speakers. For instance my frontWidth (leftWidth / rightWidth), I used my leftFront and rightRight to take the measurements. After I save the filter settings as leftWidth and rightWidth, I had to open the file and change the xml name as it says “leftFront or rightFront”. So I obviously changed that to “leftWidth or rightWidth”. Hopefully this is clear. Yes I forgot to mention that. I was meticulous about the naming for both the filename and the name of the channel in the XML file. And first did an export of filters to get a sample XML file for each to be sure of the naming since this is not fully documented elsewhere.
|
|
|
Post by Gary Cook on Apr 11, 2020 15:09:30 GMT -5
My approach is to fix the room acoustics first, as best I can within the physical limitations. Then, and only then, do I work on correcting the little issues left with REW, Audyssey, DIRAC or whatever I have at the time. In general the corrections are very minor. To my ears an acoustically bad room always sounds bad, no matter what correction methodology and process is applied. Conversely an acoustically good room without correction always sounds better to me than an acoustically poor room no matter how much correction is applied.
Cheers Gary
|
|
|
Post by marcl on Apr 12, 2020 7:52:19 GMT -5
Here's a basic problem I found when trying to make PEQ filters with REW. Assume a 5.1 system with all speakers set to Small and crossing to the sub at 40Hz for mains and 100Hz for the other three. Assume highest priority for correction is below 200Hz.
1 - Measure each speaker full range individually with each set to Small and sub enabled - each speaker's filters will be correct, as will the sub; but when you play the speakers the sub will get corrected twice ... the sub's own filter, plus the filter for each speaker correction; bass response will be over-corrected 2 - Measure each speaker only in the range above its crossover point, set to Small with sub enabled - each speaker's filters will be correct, as will the sub; but the crossover point between each speaker and the sub will not be corrected properly 3 - Measure the sub first and load correction filter, then measure each speaker full range set to Small with the SUB SET TO NONE - obviously won't work because low frequencies will be directed to L/R 4 - Measure the sub first and load correction filter, then measure each speaker full range set to Large with the sub still enabled, but TURN OFF POWER TO THE SUB - unclear how correction will be applied across the crossover frequency, but likely some over-correction 5 - Measure the sub first and load correction filter, then measure each speaker full range set to Small - this should correct the speaker properly including its own interaction with the sub at the crossover frequency
I think #5 is the only viable choice. But none of my experiments made sense because the sub filter isn't working. Am I thinking right?
|
|
|
Post by megash0n on Apr 12, 2020 8:29:13 GMT -5
Here's a basic problem I found when trying to make PEQ filters with REW. Assume a 5.1 system with all speakers set to Small and crossing to the sub at 40Hz for mains and 100Hz for the other three. Assume highest priority for correction is below 200Hz. 1 - Measure each speaker full range individually with each set to Small and sub enabled - each speaker's filters will be correct, as will the sub; but when you play the speakers the sub will get corrected twice ... the sub's own filter, plus the filter for each speaker correction; bass response will be over-corrected 2 - Measure each speaker only in the range above its crossover point, set to Small with sub enabled - each speaker's filters will be correct, as will the sub; but the crossover point between each speaker and the sub will not be corrected properly 3 - Measure the sub first and load correction filter, then measure each speaker full range set to Small with the SUB SET TO NONE - obviously won't work because low frequencies will be directed to L/R 4 - Measure the sub first and load correction filter, then measure each speaker full range set to Large with the sub still enabled, but TURN OFF POWER TO THE SUB - unclear how correction will be applied across the crossover frequency, but likely some over-correction 5 - Measure the sub first and load correction filter, then measure each speaker full range set to Small - this should correct the speaker properly including its own interaction with the sub at the crossover frequency I think #5 is the only viable choice. But none of my experiments made sense because the sub filter isn't working. Am I thinking right? In my opinion, option 5 is the best of these. IF the sub filters worked, you would run sweeps from 15 hz up to 200hz for example with the LR turned off or unplugged.. Go thru all of your sub adjustments to make sure all your subs are integrated into you room properly. Apply any necessary filters to achieve room curve or flatness of your choosing. Then, run the same on LR set to large for comparison, and to determine ideal crossover point. Adjust delay or distance to ensure proper time alignment and integration between sub/virtual sub and mains. Pick appropriate crossover point that works well between the room, sub and mains. Set speakers to small. Run sweep from 15 to 200 or 400 or whatever your goal is. Set REW to create filters in and around the crossover point and up thru say... 400hz. This is my basic, elementary understanding of how to do it. In practice, I struggle with the details and implementation. I've put my minidsp back into play because peq doesn't work on the RMC nor is there a time delay setting outside of manipulating distance.
|
|
|
Post by marcl on Apr 12, 2020 9:20:32 GMT -5
Here's a basic problem I found when trying to make PEQ filters with REW. Assume a 5.1 system with all speakers set to Small and crossing to the sub at 40Hz for mains and 100Hz for the other three. Assume highest priority for correction is below 200Hz. 1 - Measure each speaker full range individually with each set to Small and sub enabled - each speaker's filters will be correct, as will the sub; but when you play the speakers the sub will get corrected twice ... the sub's own filter, plus the filter for each speaker correction; bass response will be over-corrected 2 - Measure each speaker only in the range above its crossover point, set to Small with sub enabled - each speaker's filters will be correct, as will the sub; but the crossover point between each speaker and the sub will not be corrected properly 3 - Measure the sub first and load correction filter, then measure each speaker full range set to Small with the SUB SET TO NONE - obviously won't work because low frequencies will be directed to L/R 4 - Measure the sub first and load correction filter, then measure each speaker full range set to Large with the sub still enabled, but TURN OFF POWER TO THE SUB - unclear how correction will be applied across the crossover frequency, but likely some over-correction 5 - Measure the sub first and load correction filter, then measure each speaker full range set to Small - this should correct the speaker properly including its own interaction with the sub at the crossover frequency I think #5 is the only viable choice. But none of my experiments made sense because the sub filter isn't working. Am I thinking right? In my opinion, option 5 is the best of these. IF the sub filters worked, you would run sweeps from 15 hz up to 200hz for example with the LR turned off or unplugged.. Go thru all of your sub adjustments to make sure all your subs are integrated into you room properly. Apply any necessary filters to achieve room curve or flatness of your choosing. Then, run the same on LR set to large for comparison, and to determine ideal crossover point. Adjust delay or distance to ensure proper time alignment and integration between sub/virtual sub and mains. Pick appropriate crossover point that works well between the room, sub and mains. Set speakers to small. Run sweep from 15 to 200 or 400 or whatever your goal is. Set REW to create filters in and around the crossover point and up thru say... 400hz. This is my basic, elementary understanding of how to do it. In practice, I struggle with the details and implementation. I've put my minidsp back into play because peq doesn't work on the RMC nor is there a time delay setting outside of manipulating distance. We're on the same page. So I did my sub PEQ and alignment with miniDSP, but it still seemed the PEQ filters from REW were not being applied right on the other speakers in the XMC-2. It wasn't too bad so I gave up on it. Waiting for Dirac. Regarding time alignment ... distance and time are really the same thing. Roughly 1ms per foot. If you measure the speaker distances exactly and enter them, then run REW with a timing reference, you should see them all line up on 0 in the Impulse Response. If not, just use the 1ms per foot to fine tune.
|
|
|
Post by htnut1975 on Apr 12, 2020 19:09:02 GMT -5
Anyone know why the XMC-2 isn’t available?
|
|
|
Post by hytram on Apr 12, 2020 19:23:22 GMT -5
Anyone know why the XMC-2 isn’t available? Out of stock I would say.. The Covid-19 pandemic has effected supply chains across the board
|
|
Lsc
Emo VIPs
Posts: 3,348
|
Post by Lsc on Apr 12, 2020 19:43:38 GMT -5
In my opinion, option 5 is the best of these. IF the sub filters worked, you would run sweeps from 15 hz up to 200hz for example with the LR turned off or unplugged.. Go thru all of your sub adjustments to make sure all your subs are integrated into you room properly. Apply any necessary filters to achieve room curve or flatness of your choosing. Then, run the same on LR set to large for comparison, and to determine ideal crossover point. Adjust delay or distance to ensure proper time alignment and integration between sub/virtual sub and mains. Pick appropriate crossover point that works well between the room, sub and mains. Set speakers to small. Run sweep from 15 to 200 or 400 or whatever your goal is. Set REW to create filters in and around the crossover point and up thru say... 400hz. This is my basic, elementary understanding of how to do it. In practice, I struggle with the details and implementation. I've put my minidsp back into play because peq doesn't work on the RMC nor is there a time delay setting outside of manipulating distance. We're on the same page. So I did my sub PEQ and alignment with miniDSP, but it still seemed the PEQ filters from REW were not being applied right on the other speakers in the XMC-2. It wasn't too bad so I gave up on it. Waiting for Dirac. Regarding time alignment ... distance and time are really the same thing. Roughly 1ms per foot. If you measure the speaker distances exactly and enter them, then run REW with a timing reference, you should see them all line up on 0 in the Impulse Response. If not, just use the 1ms per foot to fine tune. Wow that’s really good. I should check that impulse response. I’ll do some google search on this since Dirac won’t help until we get Dirac bass management. Who knows when Dirac will have that finished.
|
|
cawgijoe
Emo VIPs
"We made too many of the wrong mistakes." - Yogi Berra
Posts: 4,914
|
Post by cawgijoe on Apr 12, 2020 20:49:51 GMT -5
Anyone know why the XMC-2 isn’t available? Give them a call. Quickest way to find out availability.
|
|
|
Post by ttocs on Apr 12, 2020 21:28:23 GMT -5
Anyone know why the XMC-2 isn’t available? Give them a call. Quickest way to find out availability. . . . and let us know what you find out. just curious.
|
|
Lsc
Emo VIPs
Posts: 3,348
|
Post by Lsc on Apr 13, 2020 11:18:55 GMT -5
Anyone getting a lot of comb filtering effects using 2 subs?
When I use my 2 subs (they are not the same), I get a lot of cancellation and the bass sounds flat and muted. When I disconnect my old SB13, the PB16 pressurizes the room to the point where stuff starts to fall off the walls.
I was planning on buying another PB16 down the road but I don’t think I need it.
|
|
|
Post by foggy1956 on Apr 13, 2020 11:56:43 GMT -5
Anyone getting a lot of comb filtering effects using 2 subs? When I use my 2 subs (they are not the same), I get a lot of cancellation and the bass sounds flat and muted. When I disconnect my old SB13, the PB16 pressurizes the room to the point where stuff starts to fall off the walls. I was planning on buying another PB16 down the road but I don’t think I need it. Did you time align them with REW?
|
|
Lsc
Emo VIPs
Posts: 3,348
|
Post by Lsc on Apr 13, 2020 12:27:23 GMT -5
Anyone getting a lot of comb filtering effects using 2 subs? When I use my 2 subs (they are not the same), I get a lot of cancellation and the bass sounds flat and muted. When I disconnect my old SB13, the PB16 pressurizes the room to the point where stuff starts to fall off the walls. I was planning on buying another PB16 down the road but I don’t think I need it. Did you time align them with REW? No. Unfortunately I don’t know how to time align. Maybe that’ll fix it. I’ll give it a shot before I move the SB13 to the loft.
|
|