|
|
Post by AudioHTIT on Jun 26, 2020 21:02:09 GMT -5
The off topic posts have been removed, I hope we can get back to good Dirac discussion here.
|
|
|
|
Post by AudioHTIT on Jun 27, 2020 9:45:47 GMT -5
Added Harmon target curves and Mehlau guides to 2nd post (resources) Dirac Resources PostHere’s Mehlau’s description of DL3. Version 3 has been released on 03.06.2020. The user interface is virtually the same as in version 2 but it comes with a new feature, "Bass Control" (although technically "Bass Control" was already available in 2.5.x). "Bass Control" encompasses a number of tools for optimizing the crossover region between sub(s)/sats and seat to seat differences when using multiple subs. "Bass Control" requires the AVR/P to run Dirac's own bass management DSP code. If the hardware supports it additional controls become available in the Dirac Live app. "Off": Regular single speaker optimization. Bass management has to be configured by the user in the the AVR/P. "Upmix Only": Dirac bass management replaces the AVR/P's bass management. Visualize the crossover splice between sub(s)/sats. "Full Bass Optimisation": Same as "Upmix Only" but adds optimization of the crossover splice if a single sub is used (also applies to AVR/P's that split a single subwoofer signal to multiple outputs). If the AVR/P supports individually addressable subwoofer outputs "Bass Control" additionally reduces seat to seat differences besides optimizing the crossover splice. |
|
|
|
Post by marcl on Jun 28, 2020 12:42:00 GMT -5
I watched the Dirac talk on the Hi-Fi Summit webcast. It was presented by Jakob Ågren, Director of Product Management at Dirac.
It was informative for those not familiar with digital room correction, and explained issues of reflections and impulse response and variations around a room as well as the difference between minimum phase and linear phase (Dirac attempts to correct to minimum phase). As far as any insight into Dirac 3.0, differences from Dirac 1.0, Dirac Bass Control ..... nothing really. I and others asked questions but he didn't get to them.
A tiny bit of insight into the measurement process though. He said the reason they do another sweep of the first speaker at the end of each measurement is to zero out clock differences between the PC and the Processor. Like REW does when you have a timing reference. He also said they do different kinds of correction in the impulse response as it gets a few milliseconds away from 0. More correction early to get it onto 0, and then average corrections across all measurements as time goes on, because the other measurements will have more variable reflections.
He talked about measurement patterns - minimum of 5 but preferably 9 and able to do 13 or 17. As before, prioritize position 1 as the prime listening position even if it's off-axis. Distance/timing is determined from this position. Other positions can be as varied as you want based on your listening area, but always encouraged to be a couple feet apart.
Something I found curious was a statement that Dirac has good noise rejection during the measurement process. Maybe 3.0 is better, but with 1.0 the house had to be silent. Even a little creak of a floor board (not even an actual footstep) if my wife walked upstairs would cause the measurement to fail. Agren said steady noise like HVAC is more of a problem during measurements. That may be the case, but I found any transient to be deadly. With 3.0 I plan to do all my calibrations between 3 and 5am with the AC off :-)
|
|
|
|
Post by bluescale on Jun 28, 2020 20:53:48 GMT -5
Something I found curious was a statement that Dirac has good noise rejection during the measurement process. Maybe 3.0 is better, but with 1.0 the house had to be silent. Even a little creak of a floor board (not even an actual footstep) if my wife walked upstairs would cause the measurement to fail. Agren said steady noise like HVAC is more of a problem during measurements. That may be the case, but I found any transient to be deadly. With 3.0 I plan to do all my calibrations between 3 and 5am with the AC off :-) I never foudn this to be a problem. I wonder if you had your mic sensitivity set too high, maybe? |
|
KeithL
Administrator 
Posts: 10,273
|
Post by KeithL on Jun 29, 2020 1:54:22 GMT -5
I would say that the User Interface is A LOT nicer...
And a lot of the ambiguous areas and limitations have been improved...
When you're discussing real-time acoustic measurements terms like "noise rejection" are a tricky subject. Dirac is measuring the frequency response of each speaker at a given location using a single frequency sweep. This means that ANYTHING that alters the measurement from what it should be is going to affect the results. The best you can do is to minimize the effect random noises have on the result... or warn the user to re-run measurements that are suspect... I would note that the new version of Dirac Live DOES make it relatively simple to discard a measurement and repeat it if you suspect a problem... (So, if an extraneous sound occurs during a measurement, it's simple enough to re-do it.)
And, yes, any relatively continuous noise, like from an air conditioner or refrigerator, is unavoidably going to be a problem. (You can successfully null against an absolutely consistent noise contribution... but even slight variations make its effect unpredictable.)
Reasonable assumption that it supports all the speakers that the unit supports. I've read the manual and the differences are relatively small. Yes, the 9/13/17 point options for measurement instead of the 9 point with three patterns from 1.0. Subs are handled the same as in 1.0. They're measured and corrected separately from each other, essentially treated like all the other speakers. And the UI is new, but no real different functionality. It's like "Dirac Full" on the XMC-1. On the Emotiva side we get to save multiple Dirac filter sets, which is important to some folks. Not clear what changes Dirac may have made to the algorithms for filter creation and impulse response correction. There's a talk by someone from Dirac Saturday during this Hi-Fi Summit: thehifisummit.com/2020q2seminars/Maybe we'll get some insight. I watched the Dirac talk on the Hi-Fi Summit webcast. It was presented by Jakob Ågren, Director of Product Management at Dirac. It was informative for those not familiar with digital room correction, and explained issues of reflections and impulse response and variations around a room as well as the difference between minimum phase and linear phase (Dirac attempts to correct to minimum phase). As far as any insight into Dirac 3.0, differences from Dirac 1.0, Dirac Bass Control ..... nothing really. I and others asked questions but he didn't get to them. A tiny bit of insight into the measurement process though. He said the reason they do another sweep of the first speaker at the end of each measurement is to zero out clock differences between the PC and the Processor. Like REW does when you have a timing reference. He also said they do different kinds of correction in the impulse response as it gets a few milliseconds away from 0. More correction early to get it onto 0, and then average corrections across all measurements as time goes on, because the other measurements will have more variable reflections. He talked about measurement patterns - minimum of 5 but preferably 9 and able to do 13 or 17. As before, prioritize position 1 as the prime listening position even if it's off-axis. Distance/timing is determined from this position. Other positions can be as varied as you want based on your listening area, but always encouraged to be a couple feet apart. Something I found curious was a statement that Dirac has good noise rejection during the measurement process. Maybe 3.0 is better, but with 1.0 the house had to be silent. Even a little creak of a floor board (not even an actual footstep) if my wife walked upstairs would cause the measurement to fail. Agren said steady noise like HVAC is more of a problem during measurements. That may be the case, but I found any transient to be deadly. With 3.0 I plan to do all my calibrations between 3 and 5am with the AC off :-) |
|
|
|
Post by marcl on Jun 29, 2020 7:15:06 GMT -5
Something I found curious was a statement that Dirac has good noise rejection during the measurement process. Maybe 3.0 is better, but with 1.0 the house had to be silent. Even a little creak of a floor board (not even an actual footstep) if my wife walked upstairs would cause the measurement to fail. Agren said steady noise like HVAC is more of a problem during measurements. That may be the case, but I found any transient to be deadly. With 3.0 I plan to do all my calibrations between 3 and 5am with the AC off :-) I never foudn this to be a problem. I wonder if you had your mic sensitivity set too high, maybe? With the XMC-1 and Dirac 1 the instructions were to bring the input level up to -24db, which is what I did. I notice in the Dirac 2 manual it says -30db. Ambient in my room (with HVAC on, refrigerator in the kitchen 12ft away, and a couple computers) is about 47dbC. I set input and output levels very carefully, but always thought the measurement sweeps could have been louder. |
|
|
|
Post by AudioHTIT on Jun 29, 2020 10:03:09 GMT -5
I never foudn this to be a problem. I wonder if you had your mic sensitivity set too high, maybe? With the XMC-1 and Dirac 1 the instructions were to bring the input level up to -24db, which is what I did. I notice in the Dirac 2 manual it says -30db. Ambient in my room (with HVAC on, refrigerator in the kitchen 12ft away, and a couple computers) is about 47dbC. I set input and output levels very carefully, but always thought the measurement sweeps could have been louder. I think your attempt to make things as quiet as possible is the best method/strategy (I also have to get my dogs outside or in the bedroom) and with the likelihood of people having more speakers with the G3Ps, setting mic and speaker levels will take longer to get right. From what I’ve read, noise will still be an issue. In the Mehlau section of the second post I’ve linked a method he used for setting mic and speaker levels that seems better than just setting the mic to a specific number. |
|
KeithL
Administrator
Posts: 10,273
|
Post by KeithL on Jun 29, 2020 11:44:27 GMT -5
One philosophy suggests that the ideal situation would be for the room to be totally quiet... but that's not realistic. Another suggests that you should aim to replicate your normal listening situation... to take the normal background noises into account. However, any attempt to measure the background noise level, or take it into account, assumes that it is consistent, which is also often not the case.
(It also adds a lot of time to the overall time required to calibrate things.)
The compromise we chose was to specify some starting numbers that work well for most users under most typical conditions.
The main goals are for the level to:
- be loud enough that the background noise won't have much of an effect on the measurements
- be quiet enough that it won't be too uncomfortable for you or risk damaging your speakers
- remain within this range, for all of your speakers, over the entire audible frequency range
The biggest issue is to avoid noise that might change, either in level or character, over the time of a single sweep. This includes both transient noises and things like refrigerators that might switch on and off during the sweep. (This is why it's so handy that Dirac Live lets you discard and re-run a sweep if you suspect that it has been compromised by extraneous noise.)
With the XMC-1 and Dirac 1 the instructions were to bring the input level up to -24db, which is what I did. I notice in the Dirac 2 manual it says -30db. Ambient in my room (with HVAC on, refrigerator in the kitchen 12ft away, and a couple computers) is about 47dbC. I set input and output levels very carefully, but always thought the measurement sweeps could have been louder. I think your attempt to make things as quiet as possible is the best method (I also have to get my dogs outside or in the bedroom) and with the likelihood of people having more speakers with the G3Ps, setting mic and speaker levels will take longer to get right. From what I’ve read, noise will still be an issue. In the Mehlau section of the second post I’ve linked a method he used for setting mic and speaker levels that seems better than just setting the mic to a specific number. |
|
|
|
Post by hsamwel on Jul 1, 2020 3:51:57 GMT -5
Will PEQ be active before Dirac? I have seen that it is on others manufacturers Dirac solution.
I have a BIG bump in the 20-25hz (8-9dB) area that I’d like to take down before Dirac does it’s magic.
The bad thing is that I also have quite big slump (5-6dB) in 30-35hz area..
Now I use the PEQ to take these down and up to make my subs play pretty flat from 18-120hz.
I have a couple of smaller 2-3dB bumps in 75hz and 125hz, the latter being outside my crossover
which I have at 90hz all over.
|
|
|
|
Post by AudioHTIT on Jul 1, 2020 5:45:35 GMT -5
Will PEQ be active before Dirac? I have seen that it is on others manufacturers Dirac solution. I have a BIG bump in the 20-25hz (8-9dB) area that I’d like to take down before Dirac does it’s magic. The bad thing is that I also have quite big slump (5-6dB) in 30-35hz area.. Now I use the PEQ to take these down and up to make my subs play pretty flat from 18-120hz. I have a couple of smaller 2-3dB bumps in 75hz and 125hz, the latter being outside my crossover which I have at 90hz all over. I believe PEQ will only be active in the ‘User’ slot, and not simultaneously with Dirac. In the end if all the gain or EQ comes out of the processor the results should be similar (though you have less control). I also have a dip in the mid 30’s that I’m using my sub’s eq to help with, but I still have to worry about overdriving the sub. Yours sounds like it’s doing pretty well. |
|
|
|
Post by jhurst on Jul 1, 2020 7:54:49 GMT -5
I have an RMC-1 on the way with XMC trade in. I’m so confused. Will we be getting Dirac 2 or 3? Is there a better option like Dirac vs Dirac live with XMC?
|
|
|
|
Post by AudioHTIT on Jul 1, 2020 8:15:33 GMT -5
I have an RMC-1 on the way with XMC trade in. I’m so confused. Will we be getting Dirac 2 or 3? Is there a better option like Dirac vs Dirac live with XMC? You will use the current version of Dirac Live on the laptop or desktop you use to take measurements and build filters, currently that’s Dirac 3.0.0, but it will undoubtedly be revised over time. Only the filters it creates are uploaded to the RMC, so they are somewhat version independent. You might check out the resources and FAQ in the 2nd and 3rd post of this thread. |
|
|
|
Post by marcl on Jul 1, 2020 8:30:24 GMT -5
Will PEQ be active before Dirac? I have seen that it is on others manufacturers Dirac solution. I have a BIG bump in the 20-25hz (8-9dB) area that I’d like to take down before Dirac does it’s magic. The bad thing is that I also have quite big slump (5-6dB) in 30-35hz area.. Now I use the PEQ to take these down and up to make my subs play pretty flat from 18-120hz. I have a couple of smaller 2-3dB bumps in 75hz and 125hz, the latter being outside my crossover which I have at 90hz all over. I believe PEQ will only be active in the ‘User’ slot, and not simultaneously with Dirac. In the end if all the gain or EQ comes out of the processor the results should be similar (though you have less control). I also have a dip in the mid 30’s that I’m using my sub’s eq to help with, but I still have to worry about overdriving the sub. Yours sounds like it’s doing pretty well. Correct that we won't have PEQ available when using a Dirac slot. It's better to have Dirac do the whole job. Dirac's filter algorithms are more sophisticated than PEQ. Regarding cancellations (nulls, dips) in the bass frequencies below 200Hz .... conventional wisdom says it's impossible to correct these with PEQ because the more you boost the direct signal the deeper the out of phase reflection gets at the listening position. All you do is cause a lot of distortion with little measurable effect. Somehow, Dirac manages to fill is some of the cancellations with no strain on the amps and speakers and no serious distortion. Maybe it has to do with some phase manipulation in the filters, but it definitely accomplishes something that PEQ can't. Having a miniDSP on my sub output when I had the XMC-1, I tried knocking down the big peaks in the miniDSP PEQ first and letting Dirac smooth out the overall response. I saw no advantage to doing that. Dirac did an excellent job on its own. |
|
|
|
Post by marcl on Jul 1, 2020 11:47:33 GMT -5
Correct that we won't have PEQ available when using a Dirac slot. It's better to have Dirac do the whole job. Dirac's filter algorithms are more sophisticated than PEQ. Regarding cancellations (nulls, dips) in the bass frequencies below 200Hz .... conventional wisdom says it's impossible to correct these with PEQ because the more you boost the direct signal the deeper the out of phase reflection gets at the listening position. All you do is cause a lot of distortion with little measurable effect. Somehow, Dirac manages to fill is some of the cancellations with no strain on the amps and speakers and no serious distortion. Maybe it has to do with some phase manipulation in the filters, but it definitely accomplishes something that PEQ can't. Having a miniDSP on my sub output when I had the XMC-1, I tried knocking down the big peaks in the miniDSP PEQ first and letting Dirac smooth out the overall response. I saw no advantage to doing that. Dirac did an excellent job on its own. Just wish I had the opportunity to put your experience to the test. This is bordering on the ridiculous 😡🤬. It may be relevant that in Dirac you can manipulate the target curve any way you want. You can add control points to the curve and tweak the target response graphically in a way that you can't do in REW. Measuring the actual response after applying Dirac filters - with REW, at the exact center of the Dirac measurements - you can adjust the target curve and run the Dirac filters again. Lather, rinse, repeat ... until you get exactly what makes you smile :-) |
|
KeithL
Administrator
Posts: 10,273
|
Post by KeithL on Jul 1, 2020 12:17:53 GMT -5
Dirac Live is just the full name of the product. The XMC-1 used a customized version of Dirac Live which was officially called "Dirac Live for Emotiva". (At the time it was Dirac Live 1.x)
Now it's been standardized so it's all just Dirac Live. As of a few weeks ago the official version number was at v2.53 . (You only really see the version number in specific discussions about the software itself.)
The latest version, which includes a few minor updates, is now officially v3 . Our current processors will come with the current version. (When you install it the first thing you do is to download the current version.)
If you're curious -
The XMC-1 ONLY works with the original version customized for them - which is a version of Dirac Live 1.x
The XMC-1 came with a Lite version of Dirac Live and the Full version was a $100 upgrade. The Certified Refurbished XMC-1 now comes without Dirac Live and you have to purchase the Full version separately for $100. The RMC-1, RMC-1L, and XMC-2 ONLY work on the current / new version.
The current version of Dirac Live has all of the features of the UPGRADED version of the original one... plus other improvements. (And the new processors now offer three separate Dirac Live filter sets in each of their two Speaker Presets.) I have an RMC-1 on the way with XMC trade in. I’m so confused. Will we be getting Dirac 2 or 3? Is there a better option like Dirac vs Dirac live with XMC? |
|
Lsc
Emo VIPs 
Posts: 3,435
|
Post by Lsc on Jul 3, 2020 6:27:00 GMT -5
I believe PEQ will only be active in the ‘User’ slot, and not simultaneously with Dirac. In the end if all the gain or EQ comes out of the processor the results should be similar (though you have less control). I also have a dip in the mid 30’s that I’m using my sub’s eq to help with, but I still have to worry about overdriving the sub. Yours sounds like it’s doing pretty well. Correct that we won't have PEQ available when using a Dirac slot. It's better to have Dirac do the whole job. Dirac's filter algorithms are more sophisticated than PEQ. Regarding cancellations (nulls, dips) in the bass frequencies below 200Hz .... conventional wisdom says it's impossible to correct these with PEQ because the more you boost the direct signal the deeper the out of phase reflection gets at the listening position. All you do is cause a lot of distortion with little measurable effect. Somehow, Dirac manages to fill is some of the cancellations with no strain on the amps and speakers and no serious distortion. Maybe it has to do with some phase manipulation in the filters, but it definitely accomplishes something that PEQ can't. Having a miniDSP on my sub output when I had the XMC-1, I tried knocking down the big peaks in the miniDSP PEQ first and letting Dirac smooth out the overall response. I saw no advantage to doing that. Dirac did an excellent job on its own. That’s good to know. I wasn’t planning on buying a miniDSP anyway but this confirms that I don’t need it. |
|
|
|
Post by marcl on Jul 3, 2020 10:40:19 GMT -5
Correct that we won't have PEQ available when using a Dirac slot. It's better to have Dirac do the whole job. Dirac's filter algorithms are more sophisticated than PEQ. Regarding cancellations (nulls, dips) in the bass frequencies below 200Hz .... conventional wisdom says it's impossible to correct these with PEQ because the more you boost the direct signal the deeper the out of phase reflection gets at the listening position. All you do is cause a lot of distortion with little measurable effect. Somehow, Dirac manages to fill is some of the cancellations with no strain on the amps and speakers and no serious distortion. Maybe it has to do with some phase manipulation in the filters, but it definitely accomplishes something that PEQ can't. Having a miniDSP on my sub output when I had the XMC-1, I tried knocking down the big peaks in the miniDSP PEQ first and letting Dirac smooth out the overall response. I saw no advantage to doing that. Dirac did an excellent job on its own. That’s good to know. I wasn’t planning on buying a miniDSP anyway but this confirms that I don’t need it. Well here's the catch .... you don't need the miniDSP for EQ, but ... If you have more than one sub, and they are connected to the separate sub outputs (L, R, and/or C) then Dirac will do distance and impulse response correction separately, which is fine. But, Dirac will also do EQ separately. When the multiple subs play together (send LFE from REW to measure) the combined response will be different. Until/If we get Bass Control, it's better to use a miniDSP connected to one sub output, and do distance and gain in the miniDSP to align the subs. Then let Dirac see it as one sub with respect to the other speakers and do its thing overall ... distance, impulse, EQ. Or a simpler alternative if your subs are about equidistant from the listening position ... use a Y cable. |
|
|
|
Post by cwt on Jul 3, 2020 13:07:30 GMT -5
Until/If we get Bass Control, it's better to use a miniDSP connected to one sub output, and do distance and gain in the miniDSP to align the subs. Then let Dirac see it as one sub with respect to the other speakers and do its thing overall ... distance, impulse, EQ. I wouldnt be investing in a mini dsp at this time marcl ;even though the minidsp does a similar amalgamation as say audyssey for multiple subs . Hesitant to say this following sounds like a fait accompli ; or just wishful thinking ; maybe Keith can add further info ? emotivalounge.proboards.com/post/1022634/thread |
|
|
|
Post by marcl on Jul 3, 2020 17:02:50 GMT -5
Until/If we get Bass Control, it's better to use a miniDSP connected to one sub output, and do distance and gain in the miniDSP to align the subs. Then let Dirac see it as one sub with respect to the other speakers and do its thing overall ... distance, impulse, EQ. I wouldnt be investing in a mini dsp at this time marcl ;even though the minidsp does a similar amalgamation as say audyssey for multiple subs . Hesitant to say this following sounds like a fait accompli ; or just wishful thinking ; maybe Keith can add further info ? emotivalounge.proboards.com/post/1022634/threadYes, Keith said what I said ... Dirac calibrates each channel individually. That includes the subs. But you don't want to EQ subs individually. You want to EQ the response of all subs playing at once. The reason for multiple subs is to help cancel room modes and give a more uniform bass response across listening positions. You have to measure them together and EQ the combined response. I've done it (with XMC-1, Dirac 1.0) with subs calibrated separately, together with a Y cable, and with a miniDSP. The latter is most accurate. |
|
|
|
Post by cwt on Jul 4, 2020 8:33:47 GMT -5
Yes, Keith said what I said ... Dirac calibrates each channel individually. That includes the subs. But you don't want to EQ subs individually. You want to EQ the response of all subs playing at once. The reason for multiple subs is to help cancel room modes and give a more uniform bass response across listening positions. You have to measure them together and EQ the combined response. I've done it (with XMC-1, Dirac 1.0) with subs calibrated separately, together with a Y cable, and with a miniDSP. The latter is most accurate. Yes marcl ; to elaborate I was referencing the likelihood of DBC incorporation making a minidsp mute [ 1 less box] . Anyway this post juxtaposing DBC with the emo pre pro's is convincing enough to me ..[ now dirac 2.0 is implemented the same among multiple platforms]  emotivalounge.proboards.com/post/1022648/thread |
|
