|
|
Post by hsamwel on Aug 6, 2021 4:33:29 GMT -5
You may be surprised to know that many people just run it, and it works, and they like the results. And, for some others, they aren't all that impressed with the results, and they turn it off again. The purpose of room correction is to correct various flaws and imperfections in your room acoustics and your speakers.
a) if your room and speakers are perfect, or near perfect, to begin with, then there may be nothing to correct, and so no audible improvement
b) not all flaws can be corrected using electronic room correction c) all room correction involves compromises so, depending on the situation, there is a slight chance that the net result will be worse rather than better
All of this is IN ADDITION to any flaws in the specific room correction solution you're using. HOWEVER, when all is said and done, it can't hurt to try, and you can always turn it off, or reset your processor to factory defaults, if you are unhappy with the results you get. (But then, if you're happy with how everything sounds now, maybe you don't need room correction.)
Keith, you forgot one thing. Some people also like the ”flawed” sound.. I have a friend who has two big subs that brawls loudly. Obviously have some big peaks. When we calibrated them to play more evenly and smoother. He didn’t like it at all. But me and another friend who are used to calibrated sound thought it sounded great. |
|
cawgijoe
Emo VIPs 
"When you come to a fork in the road, take it." - Yogi Berra
Posts: 5,032
|
Post by cawgijoe on Aug 6, 2021 7:19:02 GMT -5
You may be surprised to know that many people just run it, and it works, and they like the results. And, for some others, they aren't all that impressed with the results, and they turn it off again. The purpose of room correction is to correct various flaws and imperfections in your room acoustics and your speakers.
a) if your room and speakers are perfect, or near perfect, to begin with, then there may be nothing to correct, and so no audible improvement
b) not all flaws can be corrected using electronic room correction c) all room correction involves compromises so, depending on the situation, there is a slight chance that the net result will be worse rather than better
All of this is IN ADDITION to any flaws in the specific room correction solution you're using. HOWEVER, when all is said and done, it can't hurt to try, and you can always turn it off, or reset your processor to factory defaults, if you are unhappy with the results you get. (But then, if you're happy with how everything sounds now, maybe you don't need room correction.)
Keith, you forgot one thing. Some people also like the ”flawed” sound.. I have a friend who has two big subs that brawls loudly. Obviously have some big peaks. When we calibrated them to play more evenly and smoother. He didn’t like it at all. But me and another friend who are used to calibrated sound thought it sounded great. Yeah...agree with this....I have a friend who hates a calibrated TV. He prefers "Vivid". When he got a new set, I came over and adjusted his settings with recommendations from Rtings and also used the Disney Wow disc. I thought the picture looked really good. The next time I was over, he had chosen Vivid. I didn't say a word. |
|
Lsc
Emo VIPs
Posts: 3,434
|
Post by Lsc on Aug 6, 2021 9:55:16 GMT -5
I think I’ll give the Storm Audio’s speaker positions a try. I’ve always done the picture of the 9 point so far. Even though I got a great result last night it’s worth giving it a shot. Yeah, it seems I was right.. All my measurements have gotten about the same result no matter where the extra positions have been.. Although I have always measured a cube around the head as Dirac shows in the picture. StormAudio says it doesn’t matter, with a disclaimer though. I think most that changes when moving the mic positions around is how Dirac calculates the bass.. Depending on room and seating position it can probably be good or bad. Which StormAudio says in the video as well. Always save the latest project and the one your satified with. This way you can always go back and use it if another experiment/change doesn’t work out. Yes the bass from the sub is integrated better with the rest of the system. The imaging is also much improved and there a level of cohesiveness that wasn’t there before. Thanks - I hate doing these measurements but it’s a necessary evil. |
|
|
|
Post by AudioHTIT on Aug 6, 2021 10:23:15 GMT -5
… Although I have always measured a cube around the head as Dirac shows in the picture. StormAudio says it doesn’t matter, with a disclaimer though … I had always gotten the impression from the Dirac diagram to try and make the left right, high low measurements as symmetrical as possible; what I found interesting in the Storm video was how they recommend not doing that, and making the measurements in different axes. I’ll be fighting my OCD tendencies for symmetry, but will do this the next time I measure. Someone mentioned a robotic arm to hold and move the mic in a programmed and repeatable fashion while measuring … I want one of those! 🤖 |
|
ttocs
Global Moderator 
I always have a wonderful time, wherever I am, whomever I'm with. (Elwood P Dowd)
Posts: 8,113
|
Post by ttocs on Aug 6, 2021 10:37:48 GMT -5
Someone mentioned a robotic arm to hold and move the mic in a programmed and repeatable fashion while measuring … I want one of those! 🤖 Me too! Let's not forget about the "moving mic method" of measuring. This method is principally aimed at mixing workstations with just a couple of speakers, but it has very repeatable success. Play pink noise and move the mic around the area listening area. |
|
KeithL
Administrator
Posts: 10,246
|
Post by KeithL on Aug 6, 2021 10:49:42 GMT -5
I have no problem whatsoever with people who actually prefer inaccurate or colored sound. I personally prefer to have my system reproduce whatever is in the recording as accurately as possible.
However I AM willing to accept the fact that many recordings have flaws or in a more general sense just don't sound very good. And, if a particular recording I really like doesn't sound good, I sometimes make corrections to it in an audio editing program. That way I can listen to both the original and my modified copy... But, more importantly, I can make individual corrections, only when I deem them necessary... rather than try to find a coloration that "makes everything sound better".
(And modern editing programs have far more options, and far more detailed control, than you're ever going to achieve with system tweaks.)
Most of the serious problems I see involve people who actually do not prefer the most accurate sound possible...
But, for whatever reason, they are unwilling to admit that they prefer some specific sort of coloration or distortion... And, as a result, they embark on a seemingly endless quest to "prove" that the coloration they prefer is "really more accurate" or "really closer to the original"... You may be surprised to know that many people just run it, and it works, and they like the results. And, for some others, they aren't all that impressed with the results, and they turn it off again. The purpose of room correction is to correct various flaws and imperfections in your room acoustics and your speakers.
a) if your room and speakers are perfect, or near perfect, to begin with, then there may be nothing to correct, and so no audible improvement
b) not all flaws can be corrected using electronic room correction c) all room correction involves compromises so, depending on the situation, there is a slight chance that the net result will be worse rather than better
All of this is IN ADDITION to any flaws in the specific room correction solution you're using. HOWEVER, when all is said and done, it can't hurt to try, and you can always turn it off, or reset your processor to factory defaults, if you are unhappy with the results you get. (But then, if you're happy with how everything sounds now, maybe you don't need room correction.)
Keith, you forgot one thing. Some people also like the ”flawed” sound.. I have a friend who has two big subs that brawls loudly. Obviously have some big peaks. When we calibrated them to play more evenly and smoother. He didn’t like it at all. But me and another friend who are used to calibrated sound thought it sounded great. |
|
|
|
Post by marcl on Aug 6, 2021 10:50:23 GMT -5
Someone mentioned a robotic arm to hold and move the mic in a programmed and repeatable fashion while measuring … I want one of those! 🤖 Me too! Let's not forget about the "moving mic method" of measuring. This method is principally aimed at mixing workstations with just a couple of speakers, but it has very repeatable success. Play pink noise and move the mic around the area listening area. Moving mic would make sense for averaging. One of the often overlooked aspects of the way Dirac uses the measurements, is they are not averaging over all the seating locations. In one of the podcast interviews that I shared the Dirac guy explains looking for minimum phase and non-minimum phase anomalies so as to apply different types of filters (or no correction) to each. But yeah, a robotic arm would be nice. Closest I got was attaching a spring-balanced boom to the desk behind the couch so the extensions of it were always from the same point. I gave up on that the third time it hit me in the head. Now I use a long boom in a single position so it's easy to replicate within a couple inches. I also wish REW would measure a whole series like Dirac does, instead of just one speaker at a time. |
|
KeithL
Administrator
Posts: 10,246
|
Post by KeithL on Aug 6, 2021 10:51:21 GMT -5
The short answer is that it SHOULDN'T matter ... since all you're doing is choosing an output level for the test tones.
The levels you set IN THE DIRAC LIVE SOFTWARE determine the level at which the tests are run. They are not supposed to have any effect whatsoever on the final gain set in the resulting filters. The TRIM LEVELS THAT YOU SET BEFORE RUNNING DIRAC (in the processor's Speaker Preset) are also supposed to be ignored by Dirac Live when it runs the tests and calculates its filters. So the Dirac Live software will set the gain for each channel as it sees fit, based on its measurements and calculations, and IGNORE any Trim Levels you had set before it ran the tests.
HOWEVER, in the current firmware, the Trim Levels you set before running Dirac Live will remain set and in effect after you run it. And, after running Dirac Live, those manual Trim Levels will be added to the results produced by Dirac Live, causing them to be incorrect. Another way of saying that is that:
"The filters produced by Dirac Live are calculated to produce the correct levels WITH THE MANUAL LEVEL TRIMS SET TO 0 dB - regardless of what the manual Trim Levels were set to before you ran Dirac Live." Therefore feel free to us the manual Level Trims to tweak Dirac's results... But, after running Dirac Live, you should initially set them all back to 0 dB, "to hear what Dirac thinks you should hear", and then go from there...
I'm very clear on exactly how the Speaker Levels are applied after Dirac but not during Dirac calibration. With the current firmware, even when using a Dirac Filter, my subwoofer channels are adjusted to at or near the lowest level possible with the processor. Dirac vs User is different by only a little. However, the speaker channels are much more consistent with a Dirac Filter than with User. So, getting back to the seesaw issue. In regards to the Master Output Slider and Channel Sliders does it matter which one is higher if the resulting SPL is the same? If all the Channel Sliders are exactly at the same level, does it matter if they are as low as they can go so the Master Output can be all the way up? |
|
|
|
Post by marcl on Aug 6, 2021 12:58:38 GMT -5
The short answer is that it SHOULDN'T matter ... since all you're doing is choosing an output level for the test tones.
I'm very clear on exactly how the Speaker Levels are applied after Dirac but not during Dirac calibration. With the current firmware, even when using a Dirac Filter, my subwoofer channels are adjusted to at or near the lowest level possible with the processor. Dirac vs User is different by only a little. However, the speaker channels are much more consistent with a Dirac Filter than with User. So, getting back to the seesaw issue. In regards to the Master Output Slider and Channel Sliders does it matter which one is higher if the resulting SPL is the same? If all the Channel Sliders are exactly at the same level, does it matter if they are as low as they can go so the Master Output can be all the way up? Okay so ... if the Volume settings choose the output level of the test tones, and the measured response of the playback of the test tones is what Dirac uses to align the levels of all the speakers ... then how is the Volume setting not directly related to the final level settings that Dirac determines? Actually, I'll agree that somehow it isn't ... or at least it kinda was but now doesn't seem to be. I think the Volume settings are just to prevent the signal from being too low to accurately measure (relative to the noise floor) or clipped. So they can be all over the place as long as they're within the window. Then Dirac takes the lowest level speaker measurement and normalizes all the others down to that level. This results in perceived volume in the processor being usually lower after Dirac calibration. But we used to be able to affect this by purposely setting sub volumes higher (6 to 10db) than the other speakers, as long as we didn't get a clipping warning on the subs or low level warning on the other speakers. This worked until at least June 28. I did dozens of calibrations and proved that setting the sub Volume higher caused it to be measured at the same level as the L/R and resulted in higher output levels. Then sometime between June 28 and July 8 it changed and this technique no longer works. Same Dirac 3.0.14, same FW 2.3 (as it worked with 2.2 and 2.1 also). |
|
KeithL
Administrator
Posts: 10,246
|
Post by KeithL on Aug 6, 2021 14:08:40 GMT -5
Here's the way it's supposed to work (and I'm not interested in arguing about how reality may differ at the moment). The Dirac Live software is intended to create filters that correct for differences in your hardware and room acoustics. Therefore, any USER SETTINGS that you make on your processor before running Dirac Live, like using the level trims to adjust relative speaker levels, should be ignored. Then, as you say, the level adjustments you make while running the Dirac Live software are really intended merely to adjust the levels of the test tones themselves. The test tones must be loud enough to avoid the noise floor by a reasonable margin, yet not loud enough to cause clipping, or be dangerous to your speakers. Beyond that, any adjustments you make at that point, or any differences between the exact levels you set for various speakers, are NOT supposed to affect the final test results. (The Dirac Live software "nulls them out" or "ignores them" or "corrects for them".) Dirac then calculates the corrections that it wants to make in its filters for each channel. Those corrections are then compared to various "global limits"... For example, the Dirac Live filters are limited to a maximum boost, at any given frequency, of about 6 dB. Therefore, if Dirac Live has chosen to boost a certain range of frequencies, in a certain channel, by +9 dB, it must reduce the overall level of that channel by 3 dB. (So it can boost those frequencies in that channel by 9 dB relative to the others - while still avoiding exceeding the overall limit of "+6 dB of boost".) Then, since we still want all of the channels to be the same level relative to each other, it must then reduce the overall level of the other channels by the same 3 dB. And this "cross checking" must be done between all the channels.
All of this is done with the "baseline assumption" that the user trims are set to 0 dB. And any adjustments you wish to make between the levels of the individual channels will be made using the user level trims - AFTER Dirac Live has done its work. And, as you can see, this is all quite complicated, and things must be corrected and compensated for at multiple levels... BUT, assuming it all works as intended, and assuming you set the user trims back to 0 dB before or after running Dirac Live.... Dirac Live is going to "batch the level between channels" and you are supposed to then use the user level trims if you wish to adjust them. (And, as far as we know, the Dirac Target Curves DO NOT affect the overall level of each channel, but ONLY its frequency response.)
The fact that this may have acted differently in previous versions of firmware or of the Dirac Software merely indicates that either: 1) something we previously wrong and has been corrected 2) somebody changed their mind about some detail of how it all works
Note that there are various things which cannot be clearly defined either way. For example, we can expect Dirac Live to "make all the channels to play at the same level" ASSUMING THAT THE FREQUENCY RESPONSE OF EACH CHANNEL IS FLAT. However, if you have "more interesting target curves" there is some question of where we match those levels. (If you've created a Target Curve with a 3 dB bump at 50 Hz on one channel, do we match the top of that 50 Hz bump to the highest level of the other channels, or do we match the average level of that channel to the average level of the other channels?) The short answer is that it SHOULDN'T matter ... since all you're doing is choosing an output level for the test tones.
Okay so ... if the Volume settings choose the output level of the test tones, and the measured response of the playback of the test tones is what Dirac uses to align the levels of all the speakers ... then how is the Volume setting not directly related to the final level settings that Dirac determines? Actually, I'll agree that somehow it isn't ... or at least it kinda was but now doesn't seem to be. I think the Volume settings are just to prevent the signal from being too low to accurately measure (relative to the noise floor) or clipped. So they can be all over the place as long as they're within the window. Then Dirac takes the lowest level speaker measurement and normalizes all the others down to that level. This results in perceived volume in the processor being usually lower after Dirac calibration. But we used to be able to affect this by purposely setting sub volumes higher (6 to 10db) than the other speakers, as long as we didn't get a clipping warning on the subs or low level warning on the other speakers. This worked until at least June 28. I did dozens of calibrations and proved that setting the sub Volume higher caused it to be measured at the same level as the L/R and resulted in higher output levels. Then sometime between June 28 and July 8 it changed and this technique no longer works. Same Dirac 3.0.14, same FW 2.3 (as it worked with 2.2 and 2.1 also). |
|
Lsc
Emo VIPs
Posts: 3,434
|
Post by Lsc on Aug 8, 2021 9:25:43 GMT -5
… Although I have always measured a cube around the head as Dirac shows in the picture. StormAudio says it doesn’t matter, with a disclaimer though … I had always gotten the impression from the Dirac diagram to try and make the left right, high low measurements as symmetrical as possible; what I found interesting in the Storm video was how they recommend not doing that, and making the measurements in different axes. I’ll be fighting my OCD tendencies for symmetry, but will do this the next time I measure. Someone mentioned a robotic arm to hold and move the mic in a programmed and repeatable fashion while measuring … I want one of those! 🤖 I did the same until someone mentioned Storm’s tutorial and it’s remarkable how different the sound is. The imaging is so much better…I’m putting a couple Atmos speakers up today so I’ll need to remeasure…so fingers crossed. |
|
LCSeminole
Global Moderator
Res firma mitescere nescit.
Posts: 20,830
|
Post by LCSeminole on Aug 8, 2021 12:52:09 GMT -5
Over the next few days, I'll be cleaning up this threads "OFF TOPIC" posts. This should help members from having to go through the clutter and easier for everyone to follow the "ON TOPIC" discussion, as well as easier to find posts that pertain to particular subjects. Thanks for everyone's cooperation in not posting "OFF TOPIC" posts.
|
|
|
|
Post by jjkessler on Aug 10, 2021 21:15:41 GMT -5
Might be a dumb question but, why wouldn’t you want to run Dirac measurements while sitting in your favorite listening position? Why wouldn’t you want the correction to occur with you in the listening space.
If you correct the room, then sit in the space and listen, wouldn’t you be altering the corrected response ?
|
|
ttocs
Global Moderator
I always have a wonderful time, wherever I am, whomever I'm with. (Elwood P Dowd)
Posts: 8,113
|
Post by ttocs on Aug 10, 2021 22:23:03 GMT -5
Might be a dumb question but, why wouldn’t you want to run Dirac measurements while sitting in your favorite listening position? Why wouldn’t you want the correction to occur with you in the listening space. If you correct the room, then sit in the space and listen, wouldn’t you be altering the corrected response ? Yes, if the room is measured "empty", then as soon as the first person sits down the room is altered, but how much? I've actually done it with me sitting in the MLP, but it's been a while. The first challenge is allowing the mic to "see" all the speakers with my head/body possibly in the way. If the mic is in front of my nose then I'm blocking the rear channels. But being reclined below the mic would work. I don't remember exactly what the difference was, but I don't recall it being worth the effort. Maybe I'll try again and refresh my memory. |
|
richb
Sensei
Oppo Beta Group - Audioholics Reviewer
Posts: 889
|
Post by richb on Aug 11, 2021 9:01:09 GMT -5
Might be a dumb question but, why wouldn’t you want to run Dirac measurements while sitting in your favorite listening position? Why wouldn’t you want the correction to occur with you in the listening space. If you correct the room, then sit in the space and listen, wouldn’t you be altering the corrected response ? It is an interesting question. I only correct for room modes that are fairly consistent. - Rich |
|
ttocs
Global Moderator
I always have a wonderful time, wherever I am, whomever I'm with. (Elwood P Dowd)
Posts: 8,113
|
Post by ttocs on Aug 11, 2021 9:32:02 GMT -5
Might be a dumb question but, why wouldn’t you want to run Dirac measurements while sitting in your favorite listening position? Why wouldn’t you want the correction to occur with you in the listening space. If you correct the room, then sit in the space and listen, wouldn’t you be altering the corrected response ? It is an interesting question. I only correct for room modes that are fairly consistent.- Rich Good point. We control how much latitude Dirac gets to utilize. I'll let Dirac correct all the Small channels which have "normal" drivers, and also let it have it's way - mostly - with the Center channel. But for the Front Left & Right, I pull the upper Curtain down to around 400-500Hz, depending on how I feel that day, so it's mostly just the bass being adjusted. |
|
|
|
Post by okjazz on Aug 11, 2021 19:07:14 GMT -5
The Dirac Live download website shows Emotiva among the supported brands for the Dirac Live App (Apple and Android). Have you ever used the App? If yes, what is your take on it, and how does it perform compared to the PC or Mac based Dirac Live software? www.dirac.com/live/downloads/ |
|
|
|
Post by marcl on Aug 12, 2021 12:06:53 GMT -5
It is an interesting question. I only correct for room modes that are fairly consistent.- Rich Good point. We control how much latitude Dirac gets to utilize. I'll let Dirac correct all the Small channels which have "normal" drivers, and also let it have it's way - mostly - with the Center channel. But for the Front Left & Right, I pull the upper Curtain down to around 400-500Hz, depending on how I feel that day, so it's mostly just the bass being adjusted. + richb I encourage everyone to listen to the recent interviews with two executives from Dirac. Dirac decides through a pretty sophisticated algorithm (more sophisticated than us looking at the measurement) which issues are minimum phase or non-minimum phase. Dirac decides what can be corrected and what cannot. And Dirac uses different types of filters to accomplish both the amplitude and impulse response corrections. So there's no need for us to try to micromanage what Dirac corrects. This is the reason why Dirac requires us to take many measurements ... not really to cover the actual seating area, but to give them multiple views of the room response in order to make the minimum/non-minimum phase determination. www.youtube.com/watch?v=YLfT7jwTm2w&t=1saudiophilestyle.libsyn.com/dirac-research-interview |
|
uwe
Minor Hero
RMC-1
Posts: 20
|
Post by uwe on Aug 12, 2021 13:08:06 GMT -5
Good point. We control how much latitude Dirac gets to utilize. I'll let Dirac correct all the Small channels which have "normal" drivers, and also let it have it's way - mostly - with the Center channel. But for the Front Left & Right, I pull the upper Curtain down to around 400-500Hz, depending on how I feel that day, so it's mostly just the bass being adjusted. + richb I encourage everyone to listen to the recent interviews with two executives from Dirac. Dirac decides through a pretty sophisticated algorithm (more sophisticated than us looking at the measurement) which issues are minimum phase or non-minimum phase. Dirac decides what can be corrected and what cannot. And Dirac uses different types of filters to accomplish both the amplitude and impulse response corrections. So there's no need for us to try to micromanage what Dirac corrects. This is the reason why Dirac requires us to take many measurements ... not really to cover the actual seating area, but to give them multiple views of the room response in order to make the minimum/non-minimum phase determination. www.youtube.com/watch?v=YLfT7jwTm2w&t=1saudiophilestyle.libsyn.com/dirac-research-interviewFor anyone interested in how Dirac Live works internally, this early paper may also be of interest: on_room_correction.pdf (379.61 KB) diracdocs.com/on_room_correction.pdf |
|
|
|
Post by AudioHTIT on Aug 13, 2021 12:20:37 GMT -5
Over the next few days, I'll be cleaning up this threads "OFF TOPIC" posts. This should help members from having to go through the clutter and easier for everyone to follow the "ON TOPIC" discussion, as well as easier to find posts that pertain to particular subjects. Thanks for everyone's cooperation in not posting "OFF TOPIC" posts. I want to thank LCSeminole , not only for helping to clean up this thread up, but for all the work he does around here. You may have noticed that this thread's page count has gone down quite a bit, I can’t let LC take all the credit, if your favorite post is now missing, I may be to blame. There were quite a few off topic posts that makes it much harder for those looking for help to find it, and some have even used page count as a reason to blame Dirac. After scanning virtually every post, and reading many, I have to say there’s an incredible amount of knowledge, experience, and help going on here; the page count does not reflect that everyone is having trouble, but rather that this is a complex topic and process, that is not yet (nor may ever be) ‘plug and play’, and that some very detailed and elaborate help is being given I want to specifically thank marcl and ttocs who’s enthusiasm for this process and our hobby has been a tremendous benefit, education, and help to us all; geebo, hsamwel, LCSeminole, KeithL, and others have been quick to lend a hand as well. I hope we can keep this going to make the process easier for newcomers, and more effective for the experienced user. Please think twice before making an OT post — does it really further the discussion? — sometimes it does, but often it’s then replied to and a digression adds pages of unrelated talk that belongs elsewhere (sometimes a related thread like measurements might be a better place). Also consider using the 'Like' button before quoting a long post and then just adding a "+1", I think a lot of Likes on a good post is a great way to judge how helpful it is. Speaking of OT, the OP (that’s me) has updated the 3rd post with additional acronyms, FAQs, and general clean up. It does help if you refer people to the first three posts instead of answering the same questions over and over; please take a look at them, I will try to keep those posts current. If you have a FAQ you think would help out, feel free to write it up and PM me; even if you’re not sure of how to phrase it, or the exact answer, I can help put something together. So, let’s press on with the Dirac talk and help, thanks everyone—Bruce The OP (1st Post)Resources & Links (2nd Post)Acronyms and FAQ (3rd Post) |
|
