|
|
Post by okjazz on Nov 4, 2024 11:58:36 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. View AttachmentAnthony Grimani responds ... View AttachmentChanna DeSilva (TechnoDad) comments ... View AttachmentI think Mr. Grimani' s response describes very close my current target curve. I won't bother with the Dolby Atmos Music Target Curve anymore, given that even Dolby Engineers use it only for the initial approval of their paperwork by the headquarter and use totally different curve which is field-customized (per Channa's comments). Thank you Marcl for the followup. |
|
|
|
Post by marcl on Nov 5, 2024 13:12:44 GMT -5
Proof at last! I'm going to do a Dirac calibration on a Primare AVR at Capital Audio Fest on Thursday. To be sure that I'll be able to connect everything together and validate the Dirac license in the show hotel room, I set up my laptop at home today with the laptop, EDNIB and XMC-2 off the local network and connected via a small travel router. Once they saw each other I connected the laptop to the internet using my iPhone hotspot, and Dirac opened, validated the license and was ready to go. Whew!  Then I turned off the iPhone hotspot and within a couple minutes I got a message that Dirac failed validation, restore internet connection, some features may not function correctly. Hmmm. I proceeded with Volume Calibration and did one set of measurements and it seemed to work okay. But when I went to Filter Design it did the Initializing but did not say anything about Phase Correction as it normally does. I went to check the corrected impulse responses and they were not there, and then it showed a message that Phase Correction was not completed ... restore internet ... and there was now a button to retry Phase Correction. Didn't work. I turned the iPhone hotspot back on and Dirac calculated the corrected impulse responses for all channels. I verified that it does do filter design without the internet ... if I changed targets and curtains it recalculated. But not Phase Correction. So ... Dirac DOES need to maintain a connection to Sweden for more than just license validation. |
|
|
|
Post by webmst007 on Nov 5, 2024 17:15:45 GMT -5
I've tried that same test a few times and always get the same result.
If I disconnect after tests the first time I run filters I don't get them and it asks me to reconnect.
If I run re-connected I get filters and phase etc as usual.
What I think happens after that is if I make changes in Dirac after testing a new set of filter results by more sound tests in REW, and then try to re calculate the filters, it will recalculate the filters but not the phase - I think!
I'll try that again soon to confirm.
I'll post some sub stuff with MSO as well.
|
|
|
|
Post by geebo on Nov 5, 2024 18:17:59 GMT -5
With 3.11 I got an error when I went from measurements to filter creation. I backed out and went back into filter creation and there was a button to retry Phase Correction which I clicked and it went through fine. Never saw that behavior before.
|
|
|
|
Post by marcl on Nov 5, 2024 18:34:56 GMT -5
With 3.11 I got an error when I went from measurements to filter creation. I backed out and went back into filter creation and there was a button to retry Phase Correction which I clicked and it went through fine. Never saw that behavior before. 3.11 is a different animal in many ways. Probably in good ways in the long run. But methods that I have used for years are not working. I spent all day d***** around with target curves and left curtain positions and I ended up with good results ... but very different setup parameters from what has worked in the past. One thing that is not working so well is my "if you can't raise the river, lower the bridge" method ... dropping the target curve to -3 or -6 or whatever it takes to get to the bottom of a cancelation. Dirac does not seem to like pulling down peaks 10-15db ... at least not anymore. It also defaults to the paddle curve depicted as points in the point curve mode. That's fine ... I use my own targets ... but that is new. This is the first time I ever saw this phase correction thing, because I have always had a continuous internet correction. Is it calculating the phase correction and downloading from the server? Is it just reverifying the license at that moment? Is it a bug? Who knows. I'm early into my 3.11 experience but stuff is definitely different. In the end, I got very good results today. |
|
|
|
Post by geebo on Nov 5, 2024 19:54:15 GMT -5
With 3.11 I got an error when I went from measurements to filter creation. I backed out and went back into filter creation and there was a button to retry Phase Correction which I clicked and it went through fine. Never saw that behavior before. 3.11 is a different animal in many ways. Probably in good ways in the long run. But methods that I have used for years are not working. I spent all day d***** around with target curves and left curtain positions and I ended up with good results ... but very different setup parameters from what has worked in the past. One thing that is not working so well is my "if you can't raise the river, lower the bridge" method ... dropping the target curve to -3 or -6 or whatever it takes to get to the bottom of a cancelation. Dirac does not seem to like pulling down peaks 10-15db ... at least not anymore. It also defaults to the paddle curve depicted as points in the point curve mode. That's fine ... I use my own targets ... but that is new. This is the first time I ever saw this phase correction thing, because I have always had a continuous internet correction. Is it calculating the phase correction and downloading from the server? Is it just reverifying the license at that moment? Is it a bug? Who knows. I'm early into my 3.11 experience but stuff is definitely different. In the end, I got very good results today. I had an active internet connection when I got the error and the subsequent Phase Correction retry button. I, too, am very pleased with the end result. |
|
|
|
Post by marcl on Nov 10, 2024 8:19:39 GMT -5
I had an interesting experience last week, and it illustrated how different manufacturers can implement Dirac in different ways. I was asked to help set up an all-Magnepan Dolby Atmos room at Capital Audio Fest. How that went is a long story, but for now I'll say I was very pleased with the sound when I left Thursday night. Part of my task was to do a Dirac calibration using a Primare SPA25 AVR. This product has 9 channels of Class D power onboard so it can drive a 5.1.4 or 7.1.2 system ... and 7.1.4 with the addition of a stereo external amp. It comes with the bandwidth-limited Dirac version which works to 500Hz, and the full version can be purchased for the usual $99. Here are some differences between how the Primare works vs Emotiva: - I searched for a way to Enable Dirac the way we do, to do a calibration. I could not find anything like that. So I started Dirac on my laptop and it saw the SPA25 and it connected! ... and apparently put the processor in calibration mode, but ...
- I got no sound in Volume Calibration. I restarted and reconnected, but nothing. The mic was reading room noise but no calibration sounds ... until someone raised the volume on the processor itself and then the calibration sounds played! It turns out this is by design. When you enter calibration mode it zeroes the volume and you have to bring up the master on the processor until all channels are playing, and then fine tune with the Dirac sliders. In my case the Perlisten subs were WAY too loud compared to the 9 Magnepans, so we lowered the levels on the subs themselves until I could measure all at -25db. This took about four tries going back and forth from Measurement to get an acceptably high actual measurement level.
- The other difference was that we normally see the mic level at a position near the top of the scale that says 100%. The Primare version has the mic default to the middle which says 0db.
- I did the 6-position domino measurement pattern at the MLP expanding the space between measurements to about 12" ... and proceeded to Filter Design, and I found when checking the impulse responses that the subs were opposite polarity to the other speakers. I would not have had a way to know this otherwise. Dirac corrected the problem. And BTW ... there are two sub outputs on the SPA25 but they are connected together so Dirac only sees one sub ... as should be the implementation in ALL processors without DLBC .. IMO.
- The subs had some peaks and nulls around +/-5db or so ... curiously squared off rather than pointy. I pulled the target curve down to flat to eliminate the null and the predicted response went to 16Hz. After loading the filter I listened to some familiar Atmos music and brought the sub level up to just a touch above flat ... still very uniform response ... which I wanted to sound good with a variety of music as well as movies. I was very pleased with the sound balance, and Atmos object steering was excellent!
- I was not able to get them to spring for the full Dirac license and that was a shame because these new Magnepans were not broken in and there was some harshness above 500Hz. Next morning I went in with the intention of tweaking some levels, particularly the center channel which had to be crossed at 120Hz. The four beds and four tops were crossed at 80Hz and 100Hz respectively. I knew the dealer would boost the bass no matter what I said, and those high crossover points could result in some muddy dialogue. But ...
- I found that Primare locks out the level trims when a Dirac filter is enabled! I had no control over individual channel levels at all. This might have been in my favor to ensure uniform response of the system ... but no, it was still possible to increase the levels on the subs themselves, alas.
So this was an interesting experience and I was very pleased with the sound of the room and the Dirac results given the short time I had and limited ability to optimize. I must add that, alas, what I heard Thursday night was not what people would hear throughout the show ... but that's the audio business for ya ...  |
|
lgjr
Minor Hero
Posts: 60
|
Post by lgjr on Nov 12, 2024 9:12:54 GMT -5
Hello,
In 3.11, where is everyone setting the mic gain? Do I leave it at the default of 0db or raise it to the top?
|
|
|
|
Post by marcl on Nov 12, 2024 9:50:56 GMT -5
Hello, In 3.11, where is everyone setting the mic gain? Do I leave it at the default of 0db or raise it to the top? It really does depend on your system but generally we all get good measurements leaving it alone ... 0db or 100% whatever it says. Raising the master gain should be able to get you far enough above the noise floor of the room. I set the sub level first - usually at -15 to -18 - then set the rest to around -24. If your sub is too loud at those values, then lower the master gain and try to keep that 6-8db difference. After the measurement you should see the sub actually measured at about the same level as other speakers. I think 3.11 is really good, and more tolerant of measurement levels. |
|
|
|
Post by danielb on Nov 12, 2024 14:28:23 GMT -5
.... then lower the master gain and try to keep that 6-8db difference. Hello marcl I don't understand where the difference of approx. 6-8 dB between normal speakers and subwoofers comes from that you describe in the Dirac measurement (level setting). What I have done: - Measurement with REW and filter set - then level measurement of all speakers to 75 dB using REW - Adjustment of the subwoofer level until a “straight” line is created together with the center This results in a level for me (I take the center as an example because I have set it as bass-limited) - SPEAKER_CENTER: 8.0 - SPEAKER_C_SUBWOOFER: 5.0 This results in the following control measurement with REW:  Measurement with Dirac 3.11 beta (I leave all levels set in the RMC-1L unchanged) Dirac level settings of all speakers from approx. -15dB to -18dB  Control/comparison measurement with REW/Dirac (Compared to the left front speaker)  Both times (REW, Dirac) the subwoofer plays significantly louder on its own, but together with the center it fits again. That's why I actually have these questions: - Why is the subwoofer so much louder on its own, but together with a bass-limited loudspeaker it's right again - this also applies to the surround loudspeakers. - Where do the 6-8 dB's mentioned above come from? - Am I misunderstanding it, misinterpreting it , or am I doing something wrong? Best regards Daniel |
|
|
|
Post by marcl on Nov 12, 2024 17:28:16 GMT -5
.... then lower the master gain and try to keep that 6-8db difference. Hello marcl I don't understand where the difference of approx. 6-8 dB between normal speakers and subwoofers comes from that you describe in the Dirac measurement (level setting). What I have done: - Measurement with REW and filter set - then level measurement of all speakers to 75 dB using REW - Adjustment of the subwoofer level until a “straight” line is created together with the center This results in a level for me (I take the center as an example because I have set it as bass-limited) - SPEAKER_CENTER: 8.0 - SPEAKER_C_SUBWOOFER: 5.0 This results in the following control measurement with REW: View AttachmentMeasurement with Dirac 3.11 beta (I leave all levels set in the RMC-1L unchanged) Dirac level settings of all speakers from approx. -15dB to -18dB View AttachmentControl/comparison measurement with REW/Dirac (Compared to the left front speaker) View AttachmentBoth times (REW, Dirac) the subwoofer plays significantly louder on its own, but together with the center it fits again. That's why I actually have these questions: - Why is the subwoofer so much louder on its own, but together with a bass-limited loudspeaker it's right again - this also applies to the surround loudspeakers. - Where do the 6-8 dB's mentioned above come from? - Am I misunderstanding it, misinterpreting it , or am I doing something wrong? Best regards Daniel Hi Daniel ... a couple answers and a question ... The reason many of us have found that it works best to have the sub Volume calibration level higher than the rest of the speakers is that it seems to result in better Dirac levels after calibration. Usually the sub requires a lot of amplitude reduction and that sometimes makes the playback level a lot lower. May not be so for all systems. If you're asking why the LFE level in the REW measurement is higher than the bass level of the small speakers which also play through the subs, there is a simple explanation. The LFE channel in a multichannel system is boosted +10db by the processor. This is by design. So what you measure on output #4 in REW will be +10db compared to the other channels assuming all the levels are set correctly. When you check levels with the Level noise in the RMC-1, they will all play at the same level without the boost. Now my question ... in your measurement you labeled the traces "REW" and "Dirac". Is that meant to indicate REW is without Dirac prior to calibration, and the Dirac measurement is with Dirac filters after calibration? They look very similar. |
|
|
|
Post by danielb on Nov 13, 2024 4:45:21 GMT -5
Hello marcl I don't understand where the difference of approx. 6-8 dB between normal speakers and subwoofers comes from that you describe in the Dirac measurement (level setting). What I have done: - Measurement with REW and filter set - then level measurement of all speakers to 75 dB using REW - Adjustment of the subwoofer level until a “straight” line is created together with the center This results in a level for me (I take the center as an example because I have set it as bass-limited) - SPEAKER_CENTER: 8.0 - SPEAKER_C_SUBWOOFER: 5.0 This results in the following control measurement with REW: View AttachmentMeasurement with Dirac 3.11 beta (I leave all levels set in the RMC-1L unchanged) Dirac level settings of all speakers from approx. -15dB to -18dB View AttachmentControl/comparison measurement with REW/Dirac (Compared to the left front speaker) View AttachmentBoth times (REW, Dirac) the subwoofer plays significantly louder on its own, but together with the center it fits again. That's why I actually have these questions: - Why is the subwoofer so much louder on its own, but together with a bass-limited loudspeaker it's right again - this also applies to the surround loudspeakers. - Where do the 6-8 dB's mentioned above come from? - Am I misunderstanding it, misinterpreting it , or am I doing something wrong? Best regards Daniel Hi Daniel ... a couple answers and a question ... The reason many of us have found that it works best to have the sub Volume calibration level higher than the rest of the speakers is that it seems to result in better Dirac levels after calibration. Usually the sub requires a lot of amplitude reduction and that sometimes makes the playback level a lot lower. May not be so for all systems. If you're asking why the LFE level in the REW measurement is higher than the bass level of the small speakers which also play through the subs, there is a simple explanation. The LFE channel in a multichannel system is boosted +10db by the processor. This is by design. So what you measure on output #4 in REW will be +10db compared to the other channels assuming all the levels are set correctly. When you check levels with the Level noise in the RMC-1, they will all play at the same level without the boost. Now my question ... in your measurement you labeled the traces "REW" and "Dirac". Is that meant to indicate REW is without Dirac prior to calibration, and the Dirac measurement is with Dirac filters after calibration? They look very similar. Hello marcl REW means: Equalization -> User - Measured with REW - REW Filter set, control measurement with REW Dirac means: Equalization -> Dirac Preset - Control measurement performed with REW Yes, curves look pretty much the same. This is also the first time I can say that Dirac sounds good, before there was just no comparison to REW, REW was much better. I'm also not sure what's really better now, or what I've just gotten used to. |
|
|
|
Post by marcl on Nov 13, 2024 8:06:40 GMT -5
Hi Daniel ... a couple answers and a question ... The reason many of us have found that it works best to have the sub Volume calibration level higher than the rest of the speakers is that it seems to result in better Dirac levels after calibration. Usually the sub requires a lot of amplitude reduction and that sometimes makes the playback level a lot lower. May not be so for all systems. If you're asking why the LFE level in the REW measurement is higher than the bass level of the small speakers which also play through the subs, there is a simple explanation. The LFE channel in a multichannel system is boosted +10db by the processor. This is by design. So what you measure on output #4 in REW will be +10db compared to the other channels assuming all the levels are set correctly. When you check levels with the Level noise in the RMC-1, they will all play at the same level without the boost. Now my question ... in your measurement you labeled the traces "REW" and "Dirac". Is that meant to indicate REW is without Dirac prior to calibration, and the Dirac measurement is with Dirac filters after calibration? They look very similar. Hello marcl REW means: Equalization -> User - Measured with REW - REW Filter set, control measurement with REW Dirac means: Equalization -> Dirac Preset - Control measurement performed with REW Yes, curves look pretty much the same. This is also the first time I can say that Dirac sounds good, before there was just no comparison to REW, REW was much better. I'm also not sure what's really better now, or what I've just gotten used to. Okay I understand now! Before Dirac was available on the XMC-2 in mid-2020 I used REW and User PEQ for a few months. I had used Dirac V1 on the XMC-1. I found I could not get as good results with REW/User PEQ but it was okay ... it did most of the correction but of course Dirac also does phase and impulse correction. So in my experience Dirac will sound better once you have a good calibration. It's hard to do a quick A/B comparison though because of levels. But also look at your impulse response and distortion and waterfall plots in REW and see if you see a difference. It's really great that you verify with REW. So you might try different patterns of measurement and target curves. I think Dirac should be able to get you flatter response below 100Hz. Some of us have had good results with six measurements like this:  If your subs have peaks and cancelations you might try a target like this:  It will result in your sub level being lower, but then you just raise it in the processor Levels. |
|
|
|
Post by lhracing on Nov 13, 2024 9:59:15 GMT -5
I came across what looks like a problem with my Front channels phase after running Dirac. I rarely run REW after running Dirac but I did the other day and the results showed a rather large null on my Fronts at the crossover frequency. I then inverted my 3 subs on the MiniDSP and reran the fronts, the null went away and the output came up some below the crossover frequency. I decided to check the Center and it had a null at it's crossover point so I changed the settings in the MiniDSP back, reran the Center and the null went away. I swapped the wires on my Fronts, reran REW on the Fronts and Center. The results showed no nulls and sounded good.
Has anyone else experienced something similar?
Is having 3 subs with a MiniDSP setup as a single "virtual" sub confusing Dirac?
Should I change the front wiring back, ignore REW and assume Dirac got is right?
|
|
lgjr
Minor Hero
Posts: 60
|
Post by lgjr on Nov 13, 2024 9:59:48 GMT -5
Hello, In 3.11, where is everyone setting the mic gain? Do I leave it at the default of 0db or raise it to the top? It really does depend on your system but generally we all get good measurements leaving it alone ... 0db or 100% whatever it says. Raising the master gain should be able to get you far enough above the noise floor of the room. I set the sub level first - usually at -15 to -18 - then set the rest to around -24. If your sub is too loud at those values, then lower the master gain and try to keep that 6-8db difference. After the measurement you should see the sub actually measured at about the same level as other speakers. I think 3.11 is really good, and more tolerant of measurement levels. Thanks, I think I had the subs at about the same level. Will try what you suggested this weekend. System sounds great right now. |
|
|
|
Post by marcl on Nov 13, 2024 10:27:48 GMT -5
I came across what looks like a problem with my Front channels phase after running Dirac. I rarely run REW after running Dirac but I did the other day and the results showed a rather large null on my Fronts at the crossover frequency. I then inverted my 3 subs on the MiniDSP and reran the fronts, the null went away and the output came up some below the crossover frequency. I decided to check the Center and it had a null at it's crossover point so I changed the settings in the MiniDSP back, reran the Center and the null went away. I swapped the wires on my Fronts, reran REW on the Fronts and Center. The results showed no nulls and sounded good. Has anyone else experienced something similar? Is having 3 subs with a MiniDSP setup as a single "virtual" sub confusing Dirac? Should I change the front wiring back, ignore REW and assume Dirac got is right? I have experienced this thing with a channel or pair measuring with a null at the crossover many times. For a while - due to an unusual positioning of my dipole planar fronts - Dirac kept flipping the polarity of the fronts and there was a null and I switched the wires and the null went away. Then next time I ran Dirac it did the same thing and I switched the wires ... etc. I fixed that with some absorbers to keep Dirac from getting confused. That was a very unique situation though, but it illustrates the point that Dirac can get confused. I'll preface this next bit with the fact that my L/R Fronts are set to large, and I use an external active crossover to cross them to my two subs at 60Hz. Center sub is set to LFE and it goes directly to both subs. So Bass Management goes to the Fronts ... which includes subs below 60Hz ... so ... The weird thing is it comes and goes - maybe with different Dirac versions even ... not sure. My Center used to have that crossover dip, now it doesn't. Like right now all my channels are fine at the crossover, except the L/R Surrounds which have a dip and it's at 100Hz while the crossover is at 60Hz. 60 gives me the least dip. And, I'm no longer using the miniDSP to integrate subs. So this is not a simple issue. My only guess is that Dirac is doing something with phase correction and something in the measurement is causing this problem. Are you doing any PEQ in the miniDSP or just time alignment? Definitely trust the REW measurement. Check polarity in Dirac to see if it's flipping in the correction ... then check polarity in the REW impulse response. It is just weird ... |
|
|
|
Post by danielb on Nov 13, 2024 10:53:29 GMT -5
Okay I understand now! Before Dirac was available on the XMC-2 in mid-2020 I used REW and User PEQ for a few months. I had used Dirac V1 on the XMC-1. I found I could not get as good results with REW/User PEQ but it was okay ... it did most of the correction but of course Dirac also does phase and impulse correction. So in my experience Dirac will sound better once you have a good calibration. It's hard to do a quick A/B comparison though because of levels. But also look at your impulse response and distortion and waterfall plots in REW and see if you see a difference. It's really great that you verify with REW. So you might try different patterns of measurement and target curves. I think Dirac should be able to get you flatter response below 100Hz. Some of us have had good results with six measurements like this: View AttachmentIf your subs have peaks and cancelations you might try a target like this: View AttachmentIt will result in your sub level being lower, but then you just raise it in the processor Levels. It looks pretty much the same to me - that's probably why it sounds the same Waterfall of the left front speaker with REW:  Waterfall of the left front speaker with Dirac standard target curve:  Waterfall of the left front speaker with REW Dirac standard target curve in the bass range reduced from +6.5 dB to + 3.0 dB  This setting sounds better than the standard curve |
|
|
|
Post by marcl on Nov 13, 2024 11:06:02 GMT -5
Okay I understand now! Before Dirac was available on the XMC-2 in mid-2020 I used REW and User PEQ for a few months. I had used Dirac V1 on the XMC-1. I found I could not get as good results with REW/User PEQ but it was okay ... it did most of the correction but of course Dirac also does phase and impulse correction. So in my experience Dirac will sound better once you have a good calibration. It's hard to do a quick A/B comparison though because of levels. But also look at your impulse response and distortion and waterfall plots in REW and see if you see a difference. It's really great that you verify with REW. So you might try different patterns of measurement and target curves. I think Dirac should be able to get you flatter response below 100Hz. Some of us have had good results with six measurements like this: View AttachmentIf your subs have peaks and cancelations you might try a target like this: View AttachmentIt will result in your sub level being lower, but then you just raise it in the processor Levels. It looks pretty much the same to me - that's probably why it sounds the same Waterfall of the left front speaker with REW: View AttachmentWaterfall of the left front speaker with Dirac standard target curve: View AttachmentWaterfall of the left front speaker with REW Dirac standard target curve in the bass range reduced from +6.5 dB to + 3.0 dB View Attachment This setting sounds better than the standard curve I like a flat Dirac target. I believe those recommended targets with the raised bass are a marketing concession to the fact that many people don't understand that room correction will pull down big bass resonances ... maybe people say "Dirac killed my bass" because they no longer hear the boom. I think from flat to +3db makes sense for accuracy. Interesting big 30Hz resonance in the Dirac measurements ... somebody cutting the lawn outside? Helicopters?  |
|
|
|
Post by danielb on Nov 13, 2024 11:28:27 GMT -5
Interesting big 30Hz resonance in the Dirac measurements ... somebody cutting the lawn outside? Helicopters? My stomach, I was hungry after all the measuring |
|
|
|
Post by lhracing on Nov 13, 2024 12:32:41 GMT -5
Are you doing any PEQ in the miniDSP or just time alignment? Definitely trust the REW measurement. Check polarity in Dirac to see if it's flipping in the correction ... then check polarity in the REW impulse response. It is just weird ...[/quote] On my minidsp I am only doing time alignment, no PEQ. I will just leave "my wires crossed"  for now. A side note, I cannot seem to get a good download of 3.11. My laptop does not like the files. |
|
for now.