Lsc
Emo VIPs
Posts: 3,337
|
Post by Lsc on Jan 14, 2021 21:43:01 GMT -5
Well, Dirac should have issues.. The direct sound isnβt even coming towards the mic when calibrating. How would Dirac calculate distance on a bi-pole? Phase? Volume? Then we have the multiple direction and out of phase sound of di-poles. In this case bi-poles are better than di-poles I guess.. I agree with Markc. Doing Dirac on di/bi-poles is counter productive. Nowadays there are even voices coming from the sides so having di/bi-poles sure will sound strange in those cases. So after an initial XMC-2 Dirac run that included Magnepans for the fronts only (5.1.4 setup), I had the dreaded imaging problem in stereo playing fronts only (image was shifted very far right). For my second Dirac calibration I switched out the Magnepans for a pair of decent but old normal box speakers (ACI Saphire IIIs), re-ran everything and the imaging problem is now fixed, the new Dirac sounds quite good for movies and stereo and much different than the user equalization (that I didn't do anything to). HOWEVER, when I checked the levels on the new Dirac filter, all speakers are set to 0 db. After the first Dirac run, most of the levels had been adjusted. I just did a simple check using a phone app and all of the speakers are playing at about the same volume, but doesn't Dirac normally adjust all of the levels? If so, any thoughts on what I did wrong? The levels are not touched by Dirac. If you had previously adjusted the levels, thatβs the new βzeroβ post Dirac.
|
|
|
Post by leonski on Jan 14, 2021 23:20:36 GMT -5
Well, Dirac should have issues.. The direct sound isnβt even coming towards the mic when calibrating. How would Dirac calculate distance on a bi-pole? Phase? Volume? Then we have the multiple direction and out of phase sound of di-poles. In this case bi-poles are better than di-poles I guess.. I agree with Markc. Doing Dirac on di/bi-poles is counter productive. Nowadays there are even voices coming from the sides so having di/bi-poles sure will sound strange in those cases. So after an initial XMC-2 Dirac run that included Magnepans for the fronts only (5.1.4 setup), I had the dreaded imaging problem in stereo playing fronts only (image was shifted very far right). For my second Dirac calibration I switched out the Magnepans for a pair of decent but old normal box speakers (ACI Saphire IIIs), re-ran everything and the imaging problem is now fixed, the new Dirac sounds quite good for movies and stereo and much different than the user equalization (that I didn't do anything to). HOWEVER, when I checked the levels on the new Dirac filter, all speakers are set to 0 db. After the first Dirac run, most of the levels had been adjusted. I just did a simple check using a phone app and all of the speakers are playing at about the same volume, but doesn't Dirac normally adjust all of the levels? If so, any thoughts on what I did wrong? I had a similar problem with some Original MG-1 Magnepans. Maybe late '70s vintage? BTW, I kept them for over 20 years after buying them used. They were marked at the factory 'left' and 'right'. I tried them in a very small room. Certainly no larger than maybe 11' x 11' with an 8' ceilling. After setup? Sound came from upper left corner of the room. I swapped speaker L/R and problem solved. I confirmed speakers were properly wired Both Times. I'd recommend a few tests. Try Tweeter IN or OUT. Make Certain speakers are properly wired. Out of phase panels will throw an image 'off side'. Will or CAN a Dirac setup change phase on the output of a single channel? Does anyone know more about Dirac and panels?
|
|
|
Post by waynestv on Jan 15, 2021 9:24:26 GMT -5
So after an initial XMC-2 Dirac run that included Magnepans for the fronts only (5.1.4 setup), I had the dreaded imaging problem in stereo playing fronts only (image was shifted very far right). For my second Dirac calibration I switched out the Magnepans for a pair of decent but old normal box speakers (ACI Saphire IIIs), re-ran everything and the imaging problem is now fixed, the new Dirac sounds quite good for movies and stereo and much different than the user equalization (that I didn't do anything to). HOWEVER, when I checked the levels on the new Dirac filter, all speakers are set to 0 db. After the first Dirac run, most of the levels had been adjusted. I just did a simple check using a phone app and all of the speakers are playing at about the same volume, but doesn't Dirac normally adjust all of the levels? If so, any thoughts on what I did wrong? The levels are not touched by Dirac. If you had previously adjusted the levels, thatβs the new βzeroβ post Dirac. Thanks and very good to know. It makes sense because I had adjusted levels before the first Dirac run (on Preset 1) but did not adjust levels for second Dirac (on Preset 2). I think I did show that substituting the forward firing speakers for the Magnepans during the Dirac run solved my imaging problem.
|
|
|
Post by megash0n on Jan 15, 2021 11:30:01 GMT -5
The levels are not touched by Dirac. If you had previously adjusted the levels, thatβs the new βzeroβ post Dirac. Thanks and very good to know. It makes sense because I had adjusted levels before the first Dirac run (on Preset 1) but did not adjust levels for second Dirac (on Preset 2). I think I did show that substituting the forward firing speakers for the Magnepans during the Dirac run solved my imaging problem. I want to make sure I'm not misunderstanding. Some of you are replacing your speakers with temporary speakers to run Dirac calculations, and then putting your "real" speakers back in place using said filters made from different speakers?
|
|
|
Post by marcl on Jan 15, 2021 13:22:55 GMT -5
Since I don't know how the EDNIB does its translation to the processor I can't rule out that it is malfunctioning, but I would doubt it is the problem. Likewise, I doubt the mic is the issue, except as it relates to the frequency response curve. I've got a CSL UMIK-1, an OEM miniDSP UMIK-1, as well as the EMM-1, and there's a little difference between the UMIKs but not a lot, and the EMM-1 has a different curve at the extremes but less than the spec of any speaker's frequency response tolerance. The bass ends up being more pronounced with the EMM-1, so I prefer either of the UMIKs. marcl had a problem with Dirac 3 seeing a reflection that it was sensitive to, and he tracked it to a second reflection I believe, and his room was already very well treated with everything imaginable. My reflection issue I had when starting to use the XMC-2 and Dirac 3 was from the rear wave from my dipole speakers. I added some panels and that solved the issue. Dipole speakers?! Can Dirac even measure dipole speakers correctly? For modern multi channel recordings, beyond legacy DD/DTS 5.1, it is recommended to use direct-radiating speakers. Regarding mics, alot have had issues with EMM-1 that have been fixed with the switching to a UMIK mic. Strange bass readings and early drop in the treble often before 10khz. Actually, I have 9 dipole speakers in my room . Dirac works great. I have a LOT of diffusion and bass traps, and as has been mentioned I did have a problem with a side wall reflection recently. I do have one unique aspect to my setup in that I'm using what is called the "Rooze" setup for my main L/R. They are angled 45 degrees toward reflective side walls, and the direct sound is actually a bounce. I know it sounds crazy, but in my room, and with about a year and a half of tweaking it sounds great. But ... Dirac does get confused with the low level rear wave and direct (not bounced) front wave leaking to the microphone locations. So I do two things to mitigate this: first, I take the MLP measurement with 2x4ft absorption panels blocking the rear and direct front waves from the main speakers. This allows Dirac to get the impulse response and distance settings correct. I then remove the absorbers for the remaining measurements. The second thing I do is a relatively tight measurement pattern of only 7 or 8 positions. The positions are modified slightly from the Dirac diagram because my couch back is high and I want to minimize direct sound from the front speakers. Impulse alignment is dead on, and frequency response is excellent.
|
|
|
Post by waynestv on Jan 15, 2021 14:32:59 GMT -5
Thanks and very good to know. It makes sense because I had adjusted levels before the first Dirac run (on Preset 1) but did not adjust levels for second Dirac (on Preset 2). I think I did show that substituting the forward firing speakers for the Magnepans during the Dirac run solved my imaging problem. I want to make sure I'm not misunderstanding. Some of you are replacing your speakers with temporary speakers to run Dirac calculations, and then putting your "real" speakers back in place using said filters made from different speakers? Yes that is what I did in order to eliminate Dirac from shifting my stereo image to the far right. The only other change I made was to measure more positions (13 vs. 8). It sounds as good or better for movies now but much better for music in stereo. I'll be interested to hear any more ideas on how to fix that imaging issue and would then re-run Dirac again with my Magnepan fronts.
|
|
|
Post by leonski on Jan 15, 2021 14:35:35 GMT -5
Rooze got great play over on Audio Asylum.
Another technique would be to TIME LIMIT the readings. Depending on how far speakers are from front wall, you can set the measuring tool to exclude reflections based on Time.
Sound travels about 1 foot per Milisecond so that would be a starting point based on speaker distance (times 2) to front wall.
Will Dirac allow this?
Have you experimented with Tweeters IN or OUT on the 3.7s?
|
|
|
Post by marcl on Jan 15, 2021 15:41:32 GMT -5
Rooze got great play over on Audio Asylum. Another technique would be to TIME LIMIT the readings. Depending on how far speakers are from front wall, you can set the measuring tool to exclude reflections based on Time. Sound travels about 1 foot per Milisecond so that would be a starting point based on speaker distance (times 2) to front wall. Will Dirac allow this? Have you experimented with Tweeters IN or OUT on the 3.7s? No Dirac won't let you adjust time windows. Its measurements are managed by the application and we have no insight into time windows or anything like that. I have tried many configurations ... tweeters in and out (or as I prefer to call it, WOOFERS in and out, since tweeter position is more determined by listening position distance and "making the triangle"). I've done this in the conventional setup with various degrees of toe-in, and with distances from the side wall from a couple inches to two or three feet (13ft wide room). And in experimenting with the Rooze configuration I've had the tweeters in close to the wall so they are about one foot from the reflecting surface, and also out closer to the listening position which places the tweeters just over two feet from the reflecting surface. And also ... experimented with added baffles, currently using two oak boards 1'x6'x1". I had them connected to each other mounted to one side of the speaker. Currently the setup has one board on each side of the speaker with the tweeter closer to the wall, such that the board is just an inch from the wall. This last configuration results in the best time alignment with Dirac, and best imaging. There is definitely "leakage" of the back wave toward the listening position as well as the direct front wave, though both are very off-axis and lower level. I've determined these effects using temporary absorbers and REW measurements. I have a LOT of 2D diffusers in the front area of the room, but the Rooze setup is subject to odd reflections since the rear wave is more "exposed". I tracked down an issue where the rear wave from the right speaker had a path to the left reflective wall and then to the right wall and to the listening position. That out of phase rear wave was delayed 18ms, was clearly visible on the ETC plot, and really messed up the right channel imaging. I tracked it down using temporary absorbers, then I fixed it with an absorber on the right wall beside the MLP. Interestingly, the left speaker rear wave did not cause a problem because after its first reflection it went right out the door! Here's a diagram of the reflection problem and the fix. Also the most recent Impulse response (LCR and Subs) with Dirac alignment.
|
|
|
Post by leonski on Jan 15, 2021 19:07:40 GMT -5
If you turn on only ONE speaker and place your ear close to the edge of the frame and move slowly from to back you should find Null where the front and back wave cance.
For limitieing time window? Perhaps REW would work AFTER you did your best work at the Dirac adjustments...
time window is important since you can seperate the speaker from the room to an extent. That's why professional speaker reviews that actuallly bother to measure anything will frequently resort to 'close mics'
And don't forget that in Maggie Land you have 4 possible orientations, the idea of which will apply to the Rooze. All Maggies below the 20 level are single ended. So the POLE PIECE side and the mylar are different. So the suggestion woulld be to place tweeters in or out and THAN place the pole piece to the listener or as Magnepan now prefers, Anyway.....I think Magnepan made this change to intent 20 to 30 years ago. My original MG-1 panels were intened to have the Pole Piece facing the listener. My 1.6s have the Mylar facing the listener One possible reason for tweeter in or out preference is that the crossover is not symetric. In my 1.6s, it is a 2nd order Low Pass and 1st order High Pass. This puts the tweeter and woofer 90 degrees out, thru the passband. In this case 'degrees' equals time. So to correct, you put one side a slight bit further away, thru in/ out and 'toe'.
|
|
|
Post by megash0n on Jan 15, 2021 22:10:47 GMT -5
I want to make sure I'm not misunderstanding. Some of you are replacing your speakers with temporary speakers to run Dirac calculations, and then putting your "real" speakers back in place using said filters made from different speakers? Yes that is what I did in order to eliminate Dirac from shifting my stereo image to the far right. The only other change I made was to measure more positions (13 vs. 8). It sounds as good or better for movies now but much better for music in stereo. I'll be interested to hear any more ideas on how to fix that imaging issue and would then re-run Dirac again with my Magnepan fronts. I'm not sure what to say here. Regardless of whether or not it sounds better, you've pretty much negated the entire point of using Dirac. Dirac is measuring a specific speakers, it's placement in the room, and the room itself. Changing the speakers for measurement is removing one of the most critical aspects of the entire process.
|
|
|
Post by marcl on Jan 16, 2021 2:22:43 GMT -5
Yes that is what I did in order to eliminate Dirac from shifting my stereo image to the far right. The only other change I made was to measure more positions (13 vs. 8). It sounds as good or better for movies now but much better for music in stereo. I'll be interested to hear any more ideas on how to fix that imaging issue and would then re-run Dirac again with my Magnepan fronts. I'm not sure what to say here. Regardless of whether or not it sounds better, you've pretty much negated the entire point of using Dirac. Dirac is measuring a specific speakers, it's placement in the room, and the room itself. Changing the speakers for measurement is removing one of the most critical aspects of the entire process. I certainly agree that - in the end - Dirac needs to be run with the actual speakers in their actual positions. But one thing that does make sense with this as a diagnostic tool is that room correction's primary intent is to correct room resonances, and those resonances will be substantially the same below the transition frequency for any speaker placed in a given location. Below the transition frequency ... which is typically around 200Hz. I've run at least 100 Dirac calibrations with my Magnepan 3.7's in a lot of configurations and have only run into imaging issues a couple times. Most of those issues have related to my unconventional "Rooze" setup that I described in an earlier post. But I've traced some imaging issues to reflections during the calibration process confusing Dirac, and these could be related to measuring over too wide an area. To correct these problems I didn't do fewer measurements, I just pulled them a little closer together. Around 9 measurements should be sufficient. I believe someone suggested calibrating with one measurement and listening to see if the imaging is correct at the MLP. I'd agree with that. If it's good, then add measurements expanding about a foot away from the MLP and see if there's a point where it changes.
|
|
|
Post by markc on Jan 16, 2021 4:19:03 GMT -5
Itβs going to be 7.1.2. Right now itβs 7.1 - Iβm procrastinating hooking up my Atmos speakers. The surround back has made the surround sound experience more enveloping. Itβs actually added a lot to the movie watching experience. I do think I need to rerun Dirac and figure out what to do with the rear surround dispersion. Procrastinate no more! What the surround back channels did for surround, .2 ceiling speakers will develop further, but also add something that is tough for other speakers to create. Just make sure you set them as Top Middle and in the Dolby recommended zone. Whilst you can create phantom rear audio from side surrounds (much like stereo music creating a phantom centre), none of the standard channels can create a phantom ceiling origin for audio. Overhead Atmos speakers really do create a feeling of a dome of sound, as especially the illusion of sounds passing front to rear (or rear to front) passing over you, or even through you when the side surrounds are used to transition the audio.
|
|
|
Post by marcl on Jan 16, 2021 7:57:36 GMT -5
Itβs going to be 7.1.2. Right now itβs 7.1 - Iβm procrastinating hooking up my Atmos speakers. The surround back has made the surround sound experience more enveloping. Itβs actually added a lot to the movie watching experience. I do think I need to rerun Dirac and figure out what to do with the rear surround dispersion. Procrastinate no more! What the surround back channels did for surround, .2 ceiling speakers will develop further, but also add something that is tough for other speakers to create. Just make sure you set them as Top Middle and in the Dolby recommended zone. Whilst you can create phantom rear audio from side surrounds (much like stereo music creating a phantom centre), none of the standard channels can create a phantom ceiling origin for audio. Overhead Atmos speakers really do create a feeling of a dome of sound, as especially the illusion of sounds passing front to rear (or rear to front) passing over you, or even through you when the side surrounds are used to transition the audio. I tried 7.1.2 and 5.1.4. I've read that Dolby recommends that if you have 9 speakers, Atmos benefits more from 5.1.4. But agreed ... with native Atmos and even with upmixed 2 or 5.1, the tops make a big difference ... sometimes shockingly good!
|
|
|
Post by waynestv on Jan 16, 2021 9:13:01 GMT -5
I'm not sure what to say here. Regardless of whether or not it sounds better, you've pretty much negated the entire point of using Dirac. Dirac is measuring a specific speakers, it's placement in the room, and the room itself. Changing the speakers for measurement is removing one of the most critical aspects of the entire process. I certainly agree that - in the end - Dirac needs to be run with the actual speakers in their actual positions. But one thing that does make sense with this as a diagnostic tool is that room correction's primary intent is to correct room resonances, and those resonances will be substantially the same below the transition frequency for any speaker placed in a given location. Below the transition frequency ... which is typically around 200Hz. I've run at least 100 Dirac calibrations with my Magnepan 3.7's in a lot of configurations and have only run into imaging issues a couple times. Most of those issues have related to my unconventional "Rooze" setup that I described in an earlier post. But I've traced some imaging issues to reflections during the calibration process confusing Dirac, and these could be related to measuring over too wide an area. To correct these problems I didn't do fewer measurements, I just pulled them a little closer together. Around 9 measurements should be sufficient. I believe someone suggested calibrating with one measurement and listening to see if the imaging is correct at the MLP. I'd agree with that. If it's good, then add measurements expanding about a foot away from the MLP and see if there's a point where it changes. Many thanks for all of the feedback. I do plan to spend more time trying to figure out the imaging issue with the Magnepans (MG 12s)to be able to run Dirac with them. I had assumed it was the bipole reflections but will check for other issues. A few months ago I did hear a very slight buzz on only one song at a very low frequency - could be the beginning of a delamination problem and now wondering if that showed up during the Dirac run. I'll try the one measurement and then expand from there.
|
|
|
Post by marcl on Jan 16, 2021 10:21:54 GMT -5
|
|
|
Post by ttocs on Jan 16, 2021 10:37:13 GMT -5
I have one, just didn't know where to look. Not a fan of their website navigation.
|
|
Lsc
Emo VIPs
Posts: 3,337
|
Post by Lsc on Jan 16, 2021 11:17:12 GMT -5
That's odd.Β The kit is the interface that allows Dirac software on your PC to communicate with the XMC-2 to run calibration measurements and download filters.Β The current firmware in the XMC-2 works with the currently available software from Dirac.Β We have seen some crashes with Dirac 3.0.13, but it works.Β 3.0.11 is more stable.Β So you should be able to connect the EDNIB and download Dirac (recommend 3.0.11) and use it, as long as your Mac address has been registered with Dirac as confirmation of your license. You should give it a try. Not sure how one would download a "previous" version since Dirac doesn't keep the previous available, do they? I haven't seen old verisons on their web site. Itβs literally right below the download link of the latest version π¬.
|
|
|
Post by ttocs on Jan 16, 2021 11:28:24 GMT -5
Not sure how one would download a "previous" version since Dirac doesn't keep the previous available, do they? I haven't seen old verisons on their web site. Itβs literally right below the download link of the latest version π¬. Yep, see it now. Like I said, not a fan of their web site. I looked right past the faded out text, never saw it. User error.
|
|
|
Post by homer23 on Jan 16, 2021 14:17:30 GMT -5
Apologies in advance for the long post as well if this has been part of an earlier post but wanted to share my findings.
So after a ton of hours and trial and error I have narrowed down the issue of the XMC-2 either loosing Audio, Stuttering orrrr just plainly locked up. If the Unit does any one of these you will need to Hard boot as I am sure so many of you have done in the past. But I found something weird. When recovering from a lock up and the unit comes back online you should Choose the User filter then go back to your Dirac filter of choice. I have found when the unit gets locked up and recovers from a hard reboot the filters don't load properly.
Now What I have found that allows me to have stability is the following..
In my use case I have DiRAC 2.1 enabled firmware along with a few filters. The unit would be on standby and after turning the unit on by simply pressing the power button (No long boot sequence) and start performing a ton of demos were your going back and forth a lot thatβs what eventually causes the system to lock up or loose audio. I have also had this happen just by simply popping in another movie. The second thing that causes it is when you have the system on standby for a few days. Now this for sure yields the most amount of problems. Has to be something with the OS running in sleep or standby mode.
Now if you are using the User profile which drops DiRAC filters all together it doesnβt really seem like an issue. I had it running like this for a long time with very little issues. Now a tell tell sign is the volume starts to decrease over time and the Dynamics start to roll off a bit. Basically the system sounds a bit flat. Sure enough every time I start to notice this the system looses audio or locks up. So, the fix is to simply not to place your system on stand by when your done listening. Keep your finger on the power button and freshly boot the unit whenever u want to use it. I understand this wouldn't roll well with the folks who need Video Pass thru but This has yielded the best results in terms of loading the DiRAC filters properly, sound and stability. I can run thru a ton of movies, switch back and forth from diff HDMI ports and have had no issues. But once I am done the Unit is turned off by keeping my finger on the power button until i hear the relay click. Fresh boots take a little longer but itβs Rock solid after that.
Like I said this seems to be more DiRAC related. And, emotiva has a new firmware upload coming up here shortly. Hopefully!!
Dirac makes a world of a difference in sound so Iβll have to live with it. For those interested I got mine tuned by Terry Ellis (https://www.pursuitperfectsystem.com/). It took some time but the end result was really night and day. Lastly, For those who have been complaining/experiencing Dirac lowering their Volume after the initial load up from a PC just reboot the unit and just like that your volume wont be as loud as when you were using the "User" Profile but louder than the initial load.
|
|
|
Post by richter250 on Jan 16, 2021 16:36:04 GMT -5
Homer23. Thanks so much for your research!
β But I found something weird. When recovering from a lock up and the unit comes back online you should Choose the User filter then go back to your Dirac filter of choice. I have found when the unit gets locked up and recovers from a hard reboot the filters don't load properly. β
I have noticed that sometimes things just donβt sound right compared to other times. Slightly distorted and just weird. Seems to affect the center channel most often in my system. I had not noticed that it occurred specifically after recovering from a lock up. Sure enough I had a lockup last night and the system has sounded wrong all day. After reading your post I loaded the user preset and then went back to a Dirac preset and all is well! Thanks so much!
|
|