lgjr
Minor Hero
Posts: 60
|
Post by lgjr on Nov 1, 2024 8:26:48 GMT -5
Hello,
I'm using beta 3.11 and this was the first time dirac didn’t correct my subs to the target curve. Same approximate mic placement and levels. All the other speakers were fine. Anyone else using this beta version.
|
|
|
|
Post by marcl on Nov 1, 2024 8:44:45 GMT -5
Hello, I'm using beta 3.11 and this was the first time dirac didn’t correct my subs to the target curve. Same approximate mic placement and levels. All the other speakers were fine. Anyone else using this beta version. Wow that looks like a pretty significant release (for a change!). Pretty obscure finding it too because was not in the regular Beta list until 2min after I originally posted this. I'll test it today. (But wouldn't you know, I spent like three hours yesterday doing multiple calibrations  ) Meanwhile, a couple questions ... Did you do your measurements with 3.11, or did you convert an older project or use measurements from an older project? Did you uninstall the previous version before installing 3.11 (always good practice)? Which previous version worked as expected? Was it 3.10.3 or 3.10.10? Here's the link, for others who feel lucky  Dirac Live Beta 3.11helpdesk.dirac.com/en/dirac-live/Dirac-Live-Beta-311-Software-ChangelogFeatures
Delay alignment update: The delay compensation for speaker distance alignment has previously been slightly confusing in that the shown delay values have not corresponded exactly to the measured acoustic delay. This has been due to that part of the delay alignment was included as a pre-amble of zeros in the Dirac Live filter itself, leaving only the remaining part of the alignment to be applied outside of the Dirac Live filter. A consequence of this partitioning was that if the user manually disabled the delay alignment, then the part of the delay inside the Dirac Live filters would remain being applied, resulting in a misalignment of the speakers.
Target curve update: We’re unifying how the advanced target curves (breakpoint) and simple target curves (high and low shelves) behave as input to the filter design. Under the hood, shelving targets will now be converted to breakpoint targets and follow a same filter design process.
We’re introducing improvements in our measurement functionality:
Microphone recording level: Different microphones have resulted in different recording levels between MacOS and Windows due to different gains being applied on the microphone hardware and on the OS (operating system). For example, the UMIK-1 with USB-C connector has experienced very low recording levels on MacOS. In order to make all microphones work on all OS:s, we’re now only applying gain adjustment within the Dirac Live application.
Improved error detection on all platforms: An improved algorithm for detection of missing samples in the measurement recording is introduced. This will generate significantly less false alarms about measurements being “imprecise” (showing an exclamation mark). The improved algorithm will always be applied, both when using microphones connected to the AVR/receiver device and to the PC/Mac.
Improved recording quality on Windows: The Windows WASAPI driver will be used in Exclusive Mode by default. Exclusive Mode significantly reduces the probability for losing samples during the recording. It is possible to opt-out of Exclusive Mode by setting the environment variable: DAUDIO_WASAPI_NON_EXCLUSIVE = ON
We’ve extended the local logging from the Dirac Live, including received notifications from the connected device, to improve problem traceability.
Bug fixes
We have fixed an issue that caused the mic gain slider to sometimes not being shown in the Volume Calibration page.
Sometimes, a persistent error, such as an unreachable network, could generate multiple error messages. We now avoid generating identical generic error messages.
Known issues
Filter export to devices from Onkyo/Integra/Pioneer take longer time than usual, but it eventually completes.
It is not possible to use the maintenancetool to uninstall the Dirac Live application on Intel-based Macs running MacOS 15.
Workaround 1: It is possible to install a newer version of Dirac Live without first uninstalling.
Workaround 2: It is possible to uninstall by dragging the application to the trash can.
|
|
|
|
Post by okjazz on Nov 1, 2024 9:59:31 GMT -5
Hello, I'm using beta 3.11 and this was the first time dirac didn’t correct my subs to the target curve. Same approximate mic placement and levels. All the other speakers were fine. Anyone else using this beta version. Wow that looks like a pretty significant release (for a change!). Pretty obscure finding it too because was not in the regular Beta list until 2min after I originally posted this. I'll test it today. (But wouldn't you know, I spent like three hours yesterday doing multiple calibrations ) Meanwhile, a couple questions ... Did you do your measurements with 3.11, or did you convert an older project or use measurements from an older project? Did you uninstall the previous version before installing 3.11 (always good practice)? Which previous version worked as expected? Was it 3.10.3 or 3.10.10? Here's the link, for others who feel lucky Dirac Live Beta 3.11helpdesk.dirac.com/en/dirac-live/Dirac-Live-Beta-311-Software-ChangelogFeatures
Delay alignment update: The delay compensation for speaker distance alignment has previously been slightly confusing in that the shown delay values have not corresponded exactly to the measured acoustic delay. This has been due to that part of the delay alignment was included as a pre-amble of zeros in the Dirac Live filter itself, leaving only the remaining part of the alignment to be applied outside of the Dirac Live filter. A consequence of this partitioning was that if the user manually disabled the delay alignment, then the part of the delay inside the Dirac Live filters would remain being applied, resulting in a misalignment of the speakers.
Target curve update: We’re unifying how the advanced target curves (breakpoint) and simple target curves (high and low shelves) behave as input to the filter design. Under the hood, shelving targets will now be converted to breakpoint targets and follow a same filter design process.
We’re introducing improvements in our measurement functionality:
Microphone recording level: Different microphones have resulted in different recording levels between MacOS and Windows due to different gains being applied on the microphone hardware and on the OS (operating system). For example, the UMIK-1 with USB-C connector has experienced very low recording levels on MacOS. In order to make all microphones work on all OS:s, we’re now only applying gain adjustment within the Dirac Live application.
Improved error detection on all platforms: An improved algorithm for detection of missing samples in the measurement recording is introduced. This will generate significantly less false alarms about measurements being “imprecise” (showing an exclamation mark). The improved algorithm will always be applied, both when using microphones connected to the AVR/receiver device and to the PC/Mac.
Improved recording quality on Windows: The Windows WASAPI driver will be used in Exclusive Mode by default. Exclusive Mode significantly reduces the probability for losing samples during the recording. It is possible to opt-out of Exclusive Mode by setting the environment variable: DAUDIO_WASAPI_NON_EXCLUSIVE = ON
We’ve extended the local logging from the Dirac Live, including received notifications from the connected device, to improve problem traceability.
Bug fixes
We have fixed an issue that caused the mic gain slider to sometimes not being shown in the Volume Calibration page.
Sometimes, a persistent error, such as an unreachable network, could generate multiple error messages. We now avoid generating identical generic error messages.
Known issues
Filter export to devices from Onkyo/Integra/Pioneer take longer time than usual, but it eventually completes.
It is not possible to use the maintenancetool to uninstall the Dirac Live application on Intel-based Macs running MacOS 15.
Workaround 1: It is possible to install a newer version of Dirac Live without first uninstalling.
Workaround 2: It is possible to uninstall by dragging the application to the trash can.
Mrcl,
Have you heard anything about the Dolby Atmos Music Target Curve which is Dolby's recommendation for the music studio frequency response? If so, do you know if such curve is available to download somewhere and if it can be applied with the Emotiva's processors?
|
|
|
|
Post by marcl on Nov 1, 2024 10:30:27 GMT -5
Wow that looks like a pretty significant release (for a change!). Pretty obscure finding it too because was not in the regular Beta list until 2min after I originally posted this. I'll test it today. (But wouldn't you know, I spent like three hours yesterday doing multiple calibrations ) Meanwhile, a couple questions ... Did you do your measurements with 3.11, or did you convert an older project or use measurements from an older project? Did you uninstall the previous version before installing 3.11 (always good practice)? Which previous version worked as expected? Was it 3.10.3 or 3.10.10? Here's the link, for others who feel lucky Mrcl,
Have you heard anything about the Dolby Atmos Music Target Curve which is Dolby's recommendation for the music studio frequency response? If so, do you know if such curve is available to download somewhere and if it can be applied with the Emotiva's processors?
I have heard of it but am not familiar with all the specifics. I'm looking into it now and will respond later today. But I will say this ... it looks very "Harman-esque" in the sense that it has some bass rise below 100Hz and treble rolls off substantially above 2KHz. It is intended as the target for monitoring in a studio, and is not applied to the recording. This shape matches what is determined to be the perceived preference of many listeners of all ages and experience when listening to various types of music (though the actual Harman curve did not include classical or acoustic jazz). In theory, if ALL STUDIOS monitored with this curve applied to all their speakers, and if WE applied this curve to all our speakers, then we would hear what the engineer heard in the mixing room. |
|
|
|
Post by marcl on Nov 1, 2024 11:09:57 GMT -5
Hello, I'm using beta 3.11 and this was the first time dirac didn’t correct my subs to the target curve. Same approximate mic placement and levels. All the other speakers were fine. Anyone else using this beta version. Okay I just did a full calibration using 3.11. It all went well and all channels matched the target which for me is essentially flat  How did your results differ from your target? |
|
|
|
Post by marcl on Nov 1, 2024 16:07:09 GMT -5
Wow that looks like a pretty significant release (for a change!). Pretty obscure finding it too because was not in the regular Beta list until 2min after I originally posted this. I'll test it today. (But wouldn't you know, I spent like three hours yesterday doing multiple calibrations ) Meanwhile, a couple questions ... Did you do your measurements with 3.11, or did you convert an older project or use measurements from an older project? Did you uninstall the previous version before installing 3.11 (always good practice)? Which previous version worked as expected? Was it 3.10.3 or 3.10.10? Here's the link, for others who feel lucky
Mrcl,
Have you heard anything about the Dolby Atmos Music Target Curve which is Dolby's recommendation for the music studio frequency response? If so, do you know if such curve is available to download somewhere and if it can be applied with the Emotiva's processors?
I think this tells the story. dolby.my.salesforce.com/sfc/p/#700000009YuG/a/4u000000lFIV/YHgXgrJVKAcelcXkD.DnlRUBZU1CqEMGDKweLRpuu98"The frequency response specification for each speaker (i.e. the published data) should extend from 40 Hz to 18 kHz with no greater variation than +/-3 dB." and here is the recommended target curve ...  So IF a studio applies EQ to all their speakers using this curve (at reference level 85dbC) and they mix using that response, then IF you apply the same curve to your home system - and listen at reference level 85dbC - then you hear what they heard. But ... take a look at this FAQ ... professionalsupport.dolby.com/s/article/Music-Room-Calibration-FAQs?language=en_US
|
|
|
|
Post by marcl on Nov 1, 2024 16:23:44 GMT -5
How do I do my calibration? Here ya go ...  Recline both seats, cover with the weighted blanket and then with the soft fleece blanket (fluffed up around the headrest) and another blanket over the other seat. This is my approximation of the MERTF of the MLP ... "Marc and Elise -Related Transfer Function" Six pieces of blue tape ... three 8" apart across the back of the couch, and three 8" forward marking the first row of measurements with the MLP in the center. The picture shows the mic at the MLP ... first Dirac measurement. After doing the first row I use a tape set to 16" positioned at the back row, to measure 8" forward of the first row ... all at the same level. At the end, I return the mic to the MLP position, which is pretty precise given the tape. I do my post=cal verification from that position. I save measurements before proceeding to Filter Design, so if - after filter design and post measurement - I want to go back and add, delete or remeasure positions and then do Filter Design again, I can do that (very rarely happens). More often I just go back and reload a project and tweak a Target Curve or Curtain position and recalculate. |
|
lgjr
Minor Hero
Posts: 60
|
Post by lgjr on Nov 2, 2024 4:56:12 GMT -5
Measured with 3.11 and 3.10 would not uninstall. Did another run yesterday with same results. It seem like dirac is inverting the measurements. Curtains go from 16.5 to 95. Attachments:
|
|
|
|
Post by marcl on Nov 2, 2024 5:36:14 GMT -5
Measured with 3.11 and 3.10 would not uninstall. Did another run yesterday with same results. It seem like dirac is inverting the measurements. Curtains go from 16.5 to 95. Hmm ... trying to understand what your expectation is vs what we see. So what I see is that the measurement shows response of your right sub fairly flat but elevated 16-35, then a 10db dip 35-100, then back up to a flat response similar to the 16-35 before rolling off. You're only asking Dirac to correct from 16-100 and it shows the predicted response lifting that wide dip, but pulling the response from 16-35 down to 0db. Important to note that the orange is predicted and not necessarily achievable, especially being that much boost below 100Hz. How does this differ from your expectation, or what you actually saw with 3.10? Have you measured with REW after calibration to see the actual correction? Is your expectation that Dirac would have the section from 16-35 3db higher, matching the target curve? Are you using any PEQ in the sub itself? Does that account for the flat-ish response between 16-35 in the measurement? You might try not using the recommended target curve setting, and instead use the custom target curve with points. Select -o- at the upper right of the sub group and create your own flat target at +2.8db. |
|
|
|
Post by marcl on Nov 2, 2024 7:12:49 GMT -5
Dirac with Multiple Subs Refresher ...
This topic came up on one of the Dirac FBook Groups and relates to some current discussion so I thought maybe refreshing some best practices would be useful. ttocs and others please chime in if I missed anything. Remember ... 1. All three sub outputs are mono. There is no actual difference between them (unless you designate center sub=LFE). 2. Typically (center sub NOT configured LFE), the LFE channel content and all the Bass Management content for all the small speakers are mixed into a mono signal and sent to all configured sub outputs. 3. But if you configure two or three sub outputs, then Dirac will see them separately, measure them separately, and correct them separately. 4. Since the content of all of the sub outputs is identical, when they actually play the resulting sub response is the sum of all of those outputs with Dirac calibration applied to them separately, not together. 5. Therefore, the summed response of the subs is unpredictable and will not match any of the Dirac target curves. How do we fix this? The two common solutions are to either connect all the subs to a single output using Y cables or daisy chaining them; or, connect a single sub output to a miniDSP and drive the multiple subs from its output. Pros and cons ... • If you use the Y cable or daisy chain approach, Dirac will see the multiple subs as only one, it will measure them together and correct them together. This will give a much better result than connecting them separately. You have the ability to do some optimization by adjusting levels of each sub and maybe phase (or at least polarity) on each sub. But you will not be able to precisely time-align them to each other. So Dirac will see the summed subs at some average distance, set that distance for the single sub output, and align all the other speakers to that. • If you use the miniDSP approach, it allows you to optimize the combined response of the multiple subs (and directly drive up to 4) including gain, time-alignment and polarity. With the subs aligned to each other, Dirac will measure and correct their response and time-align them to the other speakers as though they were one sub. Clearly the miniDSP approach is the better solution. So then the question comes up, should you do any Parametric Equalization (PEQ) in the miniDSP or in your sub's DSP app either before or after Dirac calibration? No. There is no good reason to do this and a few good reasons NOT to do this. But despite this ... I've tried it as have many others, and some insist on doing it. But here are the issues ... 1. miniDSP and sub DSP app PEQ use only IIR filters while Dirac uses a combination of IIR and FIR filters in a proprietary "mixed phase" approach that also includes many more available filter stages than the 10 provided in the miniDSP. Dirac will do a better job. 2. Dirac analyzes the multiple position measurements to determine which deviations from the target CAN be corrected with EQ filters. This results in better seat-to-seat response and avoids excessive boost that can cause distortion. miniDSP and sub DSP will let you do whatever you want ... including trying to correct cancelations that can't be fixed with EQ. 3. Any time you use an IIR filter there will be an effect on phase around the filter frequency. Adding filters before or after Dirac calibration can affect the response and alter the calibration in unpredictable ways. 4. Best practice is to disable ALL PEQ in miniDSP and sub DSP and let Dirac do ALL the EQ correction. Only use miniDSP and sub controls for time/phase/polarity alignment between subs ... and do that before Dirac calibration. Okay now someone may rightly point out that there is a more sophisticated way to use the miniDSP which does use FIR filters. Among these approaches would be Multi-Sub Optimizer which calculates filters in an iterative algorithm. Many people have success with this approach. But it does not make sense to do MSO AND Dirac. It's multiple layers of filters, processing and latency trying to accomplish what Dirac can do very well on its own. Dirac alone is the simplest approach. And yes, if we someday have access to Dirac Live Bass Control (DLBC) that eliminates the need for the miniDSP approach and will provide the best solution ... someday. p.s. despite having said all this, I went ahead and ignored my own advice last week. I have a new implementation of open baffle dipole servo subwoofers, and I made an adjustment to signal routing and the huge 15db peak that I have around 40Hz got even a bit pointier. The result when Dirac does measurements is that this huge peak limits the maximum volume that Dirac can use for measurement. That peak will clip where the rest of the measurements are at much lower levels. When I originally integrated the subs I went ahead and used the single analog PEQ band on the sub plate amp to completely eliminate the 40Hz peak. But I found that Dirac had trouble through that range and I could not get smooth response. So, I turned off the PEQ and the Dirac results were better. So this time I thought how about if I try just a little ... maybe just pull it down 5db and let Dirac do the rest. Nope ... didn't work. I believe the reason is the phase shift introduced when you put a filter in place and then try to run an automated algorithm across the broader response. It's like that phase anomaly at that frequency is a rough patch of road that the Dirac algorithm is affected by and can't correct. So I learned. Again. Maybe for the fourth time in the past three years. p.p.s. I mentioned above the scenario where center sub is configured as LFE vs Mono. This feature (which I think is unique to our processors) allows us to direct LFE only to the center sub output. Bass Management will be directed to the left and right sub outputs if at least one is configured. If no other subs are configured then Bass Management is directed to speakers configured as large. If there are no large speakers or other subs ... don't use the center sub LFE option. But .... if you use the option, everything above applies. Multiple subs for LFE should all be connected to the center sub output. And if you use multiple subs for Bass Management, connect all of them to the left sub output. |
|
lgjr
Minor Hero
Posts: 60
|
Post by lgjr on Nov 2, 2024 9:21:45 GMT -5
Hello,
On every other occasion of of running dirac the subs matched the target curve, even in 3.10.10. I didn't move the subs are change the levels, just remeasured with 3.11. I installed 3.97 and uninstalled and installed 3.11 same results. Installed 3.10 but it wouldn't open the file from newer version. I haven't ever used REW. Not using peq on subs. Need to learn REW someday. There is a problem with uninstalling 3.10 and 3.11. I had to install a older version to uninstall dirac.
Thanks for your help.
|
|
|
|
Post by marcl on Nov 2, 2024 11:35:08 GMT -5
Hello, On every other occasion of of running dirac the subs matched the target curve, even in 3.10.10. I didn't move the subs are change the levels, just remeasured with 3.11. I installed 3.97 and uninstalled and installed 3.11 same results. Installed 3.10 but it wouldn't open the file from newer version. I haven't ever used REW. Not using peq on subs. Need to learn REW someday. There is a problem with uninstalling 3.10 and 3.11. I had to install a older version to uninstall dirac. Thanks for your help. Okay. It's typical that measurements taken with a new 3.X version won't load on an older version. Like a 3.10.10 would probably run on 3.10.3. But it makes sense that a measurement from 3.11 would not run on 3.10.x or older. I was able to uninstall 3.10.10 okay. Are you using a Mac? They say in the release notes "It is not possible to use the maintenancetool to uninstall the Dirac Live application on Intel-based Macs running MacOS 15." So we're still talking about the predicted correction looking different from what it was in 3.10.10. I think it's worth trying the custom target rather than the default blue line targets. I never use that method, but I have read elsewhere that there have been some issues with how it works. Here's what my Center Sub measurement and Target look like from yesterday's calibration with 3.11. See the little green dot in the Center Sub Group that enables the custom Target Curve. Right click to add or delete points. A little hard to see, but the predicted response does match the Target.  |
|
ttocs
Global Moderator 
I always have a wonderful time, wherever I am, whomever I'm with. (Elwood P Dowd)
Posts: 8,158
|
Post by ttocs on Nov 2, 2024 11:53:23 GMT -5
Hello, On every other occasion of of running dirac the subs matched the target curve, even in 3.10.10. I didn't move the subs are change the levels, just remeasured with 3.11. I installed 3.97 and uninstalled and installed 3.11 same results. Installed 3.10 but it wouldn't open the file from newer version. I haven't ever used REW. Not using peq on subs. Need to learn REW someday. There is a problem with uninstalling 3.10 and 3.11. I had to install a older version to uninstall dirac. Thanks for your help. Okay. It's typical that measurements taken with a new 3.X version won't load on an older version. Like a 3.10.10 would probably run on 3.10.3. But it makes sense that a measurement from 3.11 would not run on 3.10.x or older. I was able to uninstall 3.10.10 okay. Are you using a Mac? They say in the release notes "It is not possible to use the maintenancetool to uninstall the Dirac Live application on Intel-based Macs running MacOS 15." So we're still talking about the predicted correction looking different from what it was in 3.10.10. I think it's worth trying the custom target rather than the default blue line targets. I never use that method, but I have read elsewhere that there have been some issues with how it works. Here's what my Center Sub measurement and Target look like from yesterday's calibration with 3.11. See the little green dot in the Center Sub Group that enables the custom Target Curve. Right click to add or delete points. A little hard to see, but the predicted response does match the Target. View AttachmentPlease also note that in Marc's example the target is lower than 0dB, not boosted to any +x.xdB. I use the same practice, whereby I drop the target by an amount that gets it to within an area where the subwoofer, or any speaker for that matter, is very capable of providing. Boosting bad, Cutting good. |
|
|
|
Post by marcl on Nov 2, 2024 12:15:36 GMT -5
Okay. It's typical that measurements taken with a new 3.X version won't load on an older version. Like a 3.10.10 would probably run on 3.10.3. But it makes sense that a measurement from 3.11 would not run on 3.10.x or older. I was able to uninstall 3.10.10 okay. Are you using a Mac? They say in the release notes "It is not possible to use the maintenancetool to uninstall the Dirac Live application on Intel-based Macs running MacOS 15." So we're still talking about the predicted correction looking different from what it was in 3.10.10. I think it's worth trying the custom target rather than the default blue line targets. I never use that method, but I have read elsewhere that there have been some issues with how it works. Here's what my Center Sub measurement and Target look like from yesterday's calibration with 3.11. See the little green dot in the Center Sub Group that enables the custom Target Curve. Right click to add or delete points. A little hard to see, but the predicted response does match the Target. View AttachmentPlease also note that in Marc's example the target is lower than 0dB, not boosted to any +x.xdB. I use the same practice, whereby I drop the target by an amount that gets it to within an area where the subwoofer, or any speaker for that matter, is very capable of providing. Boosting bad, Cutting good. So in my case if I wanted a boost at the bottom end I could bring the target up 3db below 50Hz ... but boosting that 60Hz dip? Not happnin' |
|
|
|
Post by marcl on Nov 2, 2024 12:50:46 GMT -5
Here's how Dirac can make you crazy sometimes. ONE HERT! The only difference between these three LFE measurements is the position of the Left Curtain. Exact same flat Target Curve for all three. Dashed line - Left Curtain 11Hz Dotted line - Left Curtain 12Hz Solid line - Left Curtain 13Hz   And ... next time I measure ... the correct Curtain position MAY be different. Same issue with my L/R channels which also play the subs down below 20Hz. Sometimes 12Hz works, sometimes 11 or 13. Dirac has known issues with corrections being allowed below 20Hz ... but if you're patient you can make it work. |
|
lgjr
Minor Hero
Posts: 60
|
Post by lgjr on Nov 3, 2024 0:59:01 GMT -5
Using windows 11 laptop
|
|
|
|
Post by marcl on Nov 3, 2024 8:17:16 GMT -5
|
|
|
|
Post by webmst007 on Nov 3, 2024 23:11:29 GMT -5
Marc,
Firstly thank you for your efforts here and valuable insights. They are always a great read.
I get where you're coming from regarding multiple subs but I 'm not sure I agree with your MSO comments regarding the management of multiple subs entirely.
Lets take a 7.4.4 setup and an RMC1 in a standard home theatre room with 2 rows of seats.
Let's take the example of 4 subs in a typical Harmon layout (one in each corner or half way along the walls).
So lets feed them out of the single centre sub output set to LFE + Bass.
And lets say we have speakers which are all set to small so most of the bass goes to this single output which feeds a miniDSP.
If managed well, REW will allow a pretty accurate phase / time alignment plus relative volume of the 4 subs to be achieved.
Now running Dirac after that acting on this single channel will only have limited ability to tame the combined output of these subs.
Dirac cannot manage how these 4 subs interact with each other - only how the combined sound from all 4 can be managed as it relates to the overall sound.
This is because without the precise interactions of these subs being measured and managed as they work in unison in the room, a true picture of their individual and combined output is not being addressed.
That's why MSO with the G3P processors (or RMC1+ without DLBC) plus a miniDSP is a good way to add more accurate multiple sub management while we wait for DLBC.
Marc, I do agree that in general any EQ can introduce issues with phase, but MSO does calculate relative phase and volume into account when processing a set of EQ, delay, and volume calculations for a group of subs.
MSO reads the Individual EQ assessment of each sub in situ and then calculates the optimum performance of the combined array.
The really good side of MSO is that the final results can be recalculated and loaded to a miniDSP to test in REW any number of times without having to re-measure the subs. Parameters inside MSO allow for many versions of the final EQ, delay, and volume parameters to be calculated. Uploading these into a miniDSP is simple which allows a number of sets to be tested in the real world with the sub array.
|
|
|
|
Post by marcl on Nov 4, 2024 2:41:57 GMT -5
Marc, Firstly thank you for your efforts here and valuable insights. They are always a great read. I get where you're coming from regarding multiple subs but I 'm not sure I agree with your MSO comments regarding the management of multiple subs entirely. Lets take a 7.4.4 setup and an RMC1 in a standard home theatre room with 2 rows of seats. Let's take the example of 4 subs in a typical Harmon layout (one in each corner or half way along the walls). So lets feed them out of the single centre sub output set to LFE + Bass. And lets say we have speakers which are all set to small so most of the bass goes to this single output which feeds a miniDSP. If managed well, REW will allow a pretty accurate phase / time alignment plus relative volume of the 4 subs to be achieved. Now running Dirac after that acting on this single channel will only have limited ability to tame the combined output of these subs. Dirac cannot manage how these 4 subs interact with each other - only how the combined sound from all 4 can be managed as it relates to the overall sound. This is because without the precise interactions of these subs being measured and managed as they work in unison in the room, a true picture of their individual and combined output is not being addressed. That's why MSO with the G3P processors (or RMC1+ without DLBC) plus a miniDSP is a good way to add more accurate multiple sub management while we wait for DLBC. Marc, I do agree that in general any EQ can introduce issues with phase, but MSO does calculate relative phase and volume into account when processing a set of EQ, delay, and volume calculations for a group of subs. MSO reads the Individual EQ assessment of each sub in situ and then calculates the optimum performance of the combined array. The really good side of MSO is that the final results can be recalculated and loaded to a miniDSP to test in REW any number of times without having to re-measure the subs. Parameters inside MSO allow for many versions of the final EQ, delay, and volume parameters to be calculated. Uploading these into a miniDSP is simple which allows a number of sets to be tested in the real world with the sub array. Thanks! I know you have LOTS of experience with MSO so your input on this is really valuable. I can see how in theory the combination as you describe it could be put to good use, though I would like to see that quantified too, to see what the combination does to the final result vs MSO or Dirac alone. Would you characterize the result of using MSO first and then Dirac as an overall improvement in amplitude/phase/impulse response that can be measured? Is the benefit more about reducing seat to seat variation vs response at the MLP? Or is the benefit more subjective vs quantitative? |
|
|
|
Post by marcl on Nov 4, 2024 2:53:50 GMT -5
Mrcl,
Have you heard anything about the Dolby Atmos Music Target Curve which is Dolby's recommendation for the music studio frequency response? If so, do you know if such curve is available to download somewhere and if it can be applied with the Emotiva's processors?
Post yesterday from Sean Olive ...PhD Senior Fellow and Acoustic Research Fellow at Harman International. Sean worked with Floyd Toole and others at Harman and of course has done extensive research on human perception of sound from loudspeakers and headphones.  Anthony Grimani responds ...  Channa DeSilva (TechnoDad) comments ...  |
|
)
