|
|
Post by marcl on Jan 27, 2022 15:53:32 GMT -5
I know your 3.7's reflect off the side wall and you must be creative to measure them properly with Dirac. If you have a standard placement for the Maggies do you have to do anything differently, relative to Dirac, than you would with other speakers? Many years ago.....maybe 5 decades! I lived in apartments and had a good stereo. In those days? JBL4311 copies made by RSL here in Southern California. They were regarded well enough that more than a few pair found their way into Studio Use. I had the 3600s which had a better (dome) tweeter. 12" 3 way was the norm back than for a large home speaker. But what do to about the neighbors? When Moving to a new flat? I'd convince the neighbors I was HARD OF HEARING. Then when I turned it UP, they'd know why. The TRICK? Do NOT abuse this privilage. I'd NEVER crank after maybe 7pm or 8pm and NEVER before maybe NOON. You had to really control yourself so you didn't respond to a noise or voice you should NOT hear! Try THAT if you can. Works the charm..... NOBODY ever thought to question a partially DEAF Audiophile.......I don't know why? For MY Nickel? Bloated bass MAY be the result of excess overlap between mains and sub. My mains / sub have NO overlap except as they naturally roll off. So the sub crosses at 45hz or so @24db octave and the MAINS cut off about 55hz @12db octave. No Conflict..... I'm in an end unit so I have neighbors front and back. Right now two single women, one retired. If they have a plumbing emergency or need help with anything, they call me ... so we're cool! But when I first moved in at 28 I had a big drum set and despite considerable soundproofing I guess the lows got through the walls sometimes. The older lady who lived in the back at the time told me it didn't bother her but if it did she'd bang on the wall. I though ... oh wow, so THAT'S what that noise was!. So I told her ... weird, I never heard any banging ... so she stopped  |
|
|
|
Post by marcl on Jan 27, 2022 17:10:05 GMT -5
News from Flavio ... woohooo!
flax said:
Thank you all for sharing your thoughts/wishes!
It goes without saying that, as due, they are taken into full consideration
We're backing away from the 20hz cutoff change, setting compensation for < 20 Hz is being reinstated
|
|
|
|
Post by fbczar on Jan 27, 2022 20:29:29 GMT -5
I know your 3.7's reflect off the side wall and you must be creative to measure them properly with Dirac. If you have a standard placement for the Maggies do you have to do anything differently, relative to Dirac, than you would with other speakers? No issues with conventional speaker placement. Dirac does a great job without any trickery. In fact I have a CC5 center with DWM, MC1 surrounds, and MMGW top front Atmos speakers. Dirac handles thwm just fin, corrects impulse response and aligns them perfectly. I would think correcting the impulse response on the Maggies is a big advantage for Dirac over other forms of EQ. |
|
|
|
Post by fbczar on Jan 27, 2022 21:10:57 GMT -5
News from Flavio ... woohooo! flax said: Thank you all for sharing your thoughts/wishes! It goes without saying that, as due, they are taken into full consideration We're backing away from the 20hz cutoff change, setting compensation for < 20 Hz is being reinstated You have clout. Thankfully. |
|
|
|
Post by marcl on Jan 28, 2022 6:56:45 GMT -5
News from Flavio ... woohooo! flax said: Thank you all for sharing your thoughts/wishes! It goes without saying that, as due, they are taken into full consideration We're backing away from the 20hz cutoff change, setting compensation for < 20 Hz is being reinstated You have clout. Thankfully. Well, maybe I added to the voices but it was the regulars over on AVS that made the most noise for sure. Now hopefully we get a 3.2.2 very soon. And to the other point about impulse response correction for Maggies .... well first of all they do remarkably well on their own because the bass panels have so much less inertia than a big heavy cone speaker. The uncorrected impulse response is not bad, and the difference in response time between the woofers and tweeters in the Maggie panels should be somewhat compensated for by the toe-in. For example, they recommend that the tweeter of a 3.7 be one inch further from the MLP than the center of the woofer. Here's a comparison of the uncorrected impulse response of my L/R 3.7s as measured by Dirac, and as measured with REW.  And this is with the speakers in the Rooze configuration which causes the measured impulse to look a little sloppy due to all the unusual reflections. However, the Rooze does place the tweeter farther from the wooofer than normal. |
|
|
|
Post by marcl on Jan 28, 2022 7:06:12 GMT -5
|
|
|
|
Post by marcl on Jan 28, 2022 8:40:09 GMT -5
|
|
|
|
Post by marcl on Jan 28, 2022 11:34:21 GMT -5
I just updated to 3.2.2. I have saved measurements-only from the last calibration I did with 3.1.2 (measurements prior to preprocessing and filter design). Using the same target curves, I ran filter design using 3.2.2. Having the saved project from 3.1.2, I used REW to run a full set of measurements, without moving the mic, to compare processing of 3.1.2 and 3.2.2. The results were interesting. For the most part results were nearly identical. The one major exception was the left front channel which with the new 3.2.2 now has a 10db dip around 20Hz. Measuring L+R results in a 5db rolloff below 25Hz. My subwoofer also responds down to 16Hz and though the 3.2.2 response is not identical to 3.1.2, it does not have this significant dip at 20Hz. This is most interesting given the 20Hz issue that has just been addressed, so I'm suspicious if maybe the left channel processing is incorrect in 3.2.2. I reported this over on AVS also so maybe Flavio will see it.    |
|
|
|
Post by fbczar on Jan 28, 2022 12:47:31 GMT -5
You have clout. Thankfully. Well, maybe I added to the voices but it was the regulars over on AVS that made the most noise for sure. Now hopefully we get a 3.2.2 very soon. And to the other point about impulse response correction for Maggies .... well first of all they do remarkably well on their own because the bass panels have so much less inertia than a big heavy cone speaker. The uncorrected impulse response is not bad, and the difference in response time between the woofers and tweeters in the Maggie panels should be somewhat compensated for by the toe-in. For example, they recommend that the tweeter of a 3.7 be one inch further from the MLP than the center of the woofer. Here's a comparison of the uncorrected impulse response of my L/R 3.7s as measured by Dirac, and as measured with REW. <button disabled="" class="c-attachment-insert--linked o-btn--sm">Attachment Deleted</button> And this is with the speakers in the Rooze configuration which causes the measured impulse to look a little sloppy due to all the unusual reflections. However, the Rooze does place the tweeter farther from the wooofer than normal. What about the subwoofer impulse response relative to the Maggies? Any substantial benefit there? I see what you mean by the Maggies impulse response. My big TDL Reference Standards have a nice post correction response, but not as good as your Maggies have pre-correction. The Rooze configuration presents an entirely different way Maggies can be configured. Is a Rooze configuration better in a small room or is your choice to use it based on the shape of the room? |
|
|
|
Post by fbczar on Jan 28, 2022 12:52:16 GMT -5
I just updated to 3.2.2. I have saved measurements-only from the last calibration I did with 3.1.2 (measurements prior to preprocessing and filter design). Using the same target curves, I ran filter design using 3.2.2. Having the saved project from 3.1.2, I used REW to run a full set of measurements, without moving the mic, to compare processing of 3.1.2 and 3.2.2. The results were interesting. For the most part results were nearly identical. The one major exception was the left front channel which with the new 3.2.2 now has a 10db dip around 20Hz. Measuring L+R results in a 5db rolloff below 25Hz. My subwoofer also responds down to 16Hz and though the 3.2.2 response is not identical to 3.1.2, it does not have this significant dip at 20Hz. This is most interesting given the 20Hz issue that has just been addressed, so I'm suspicious if maybe the left channel processing is incorrect in 3.2.2. I reported this over on AVS also so maybe Flavio will see it. <button disabled="" class="c-attachment-insert--linked o-btn--sm">Attachment Deleted</button><button disabled="" class="c-attachment-insert--linked o-btn--sm">Attachment Deleted</button><button disabled="" class="c-attachment-insert--linked o-btn--sm">Attachment Deleted</button> I am going to leave everything alone until you fix this |
|
|
|
Post by marcl on Jan 28, 2022 13:50:33 GMT -5
Well, maybe I added to the voices but it was the regulars over on AVS that made the most noise for sure. Now hopefully we get a 3.2.2 very soon. And to the other point about impulse response correction for Maggies .... well first of all they do remarkably well on their own because the bass panels have so much less inertia than a big heavy cone speaker. The uncorrected impulse response is not bad, and the difference in response time between the woofers and tweeters in the Maggie panels should be somewhat compensated for by the toe-in. For example, they recommend that the tweeter of a 3.7 be one inch further from the MLP than the center of the woofer. Here's a comparison of the uncorrected impulse response of my L/R 3.7s as measured by Dirac, and as measured with REW. <button disabled="" class="c-attachment-insert--linked o-btn--sm">Attachment Deleted</button> And this is with the speakers in the Rooze configuration which causes the measured impulse to look a little sloppy due to all the unusual reflections. However, the Rooze does place the tweeter farther from the wooofer than normal. What about the subwoofer impulse response relative to the Maggies? Any substantial benefit there? I see what you mean by the Maggies impulse response. My big TDL Reference Standards have a nice post correction response, but not as good as your Maggies have pre-correction. The Rooze configuration presents an entirely different way Maggies can be configured. Is a Rooze configuration better in a small room or is your choice to use it based on the shape of the room? Here is the subwoofer impulse (two subs aligned in the miniDSP and corrected by Dirac), vs the L/R front response. So the design is a little complicated with the L/R going to a 3.7 and a DWM each side, and the combined L+R going to the miniDSP to feed the subs below 50Hz. However .... the way REW does the impulse measurement the part of the L/R that goes to the subs doesn't affect what the impulse graph looks like. The sub impulse response is about what you'd see uncorrected, but a little cleaner and of course aligned perfectly with the L/R.  Rooze is a wild "out of the box" idea and I never would have thought it would work, but I tried it and it was so good with a very crude implementation that I chased it for several months including some room modifications. And yes, I think the benefits are realized most in a small room. Mine is 12.5ft wide, the trailing edge of the speakers is about 5ft from the front wall, and the MLP is 9ft from the leading edge. In a much wider room you would have to sit really far away. But I actually can get imaging that appears wider than the room itself. |
|
|
|
Post by leonski on Jan 28, 2022 15:51:22 GMT -5
You have clout. Thankfully. Well, maybe I added to the voices but it was the regulars over on AVS that made the most noise for sure. Now hopefully we get a 3.2.2 very soon. And to the other point about impulse response correction for Maggies .... well first of all they do remarkably well on their own because the bass panels have so much less inertia than a big heavy cone speaker. The uncorrected impulse response is not bad, and the difference in response time between the woofers and tweeters in the Maggie panels should be somewhat compensated for by the toe-in. For example, they recommend that the tweeter of a 3.7 be one inch further from the MLP than the center of the woofer. Here's a comparison of the uncorrected impulse response of my L/R 3.7s as measured by Dirac, and as measured with REW. <button disabled="" class="c-attachment-insert--linked o-btn--sm">Attachment Deleted</button> And this is with the speakers in the Rooze configuration which causes the measured impulse to look a little sloppy due to all the unusual reflections. However, the Rooze does place the tweeter farther from the wooofer than normal. A lot of the issue with maggies is the crossover. In my 1.6s? The difference between the LP and HP filters is 90 degrees and is in effect, a time delay. That's why putting the tweeter ahead or behind the main panel....using TOE IN......helps..... Other maggies, especially the 3-way 3 series UP can have a compounded version of the same issue..... In my 1.6s? The tweeter LEADS the woofer section by a very tiny amount... |
|
|
|
Post by marcl on Jan 30, 2022 15:25:17 GMT -5
I've seen a weird dip with 3.2.2 using measurements/project from 3.1.2, and only for my Left Front channel. Here's some specific data. I have a file of just measurements from 9/23 3.1.2 (before preprocessing) and I have a project file with filters from 9/23 3.1.2. I opened each with 3.2.2 and it took a while to open. The project file no longer had target curves or curtain positions so I loaded the targets I had used on 9/23 and I set the curtains as I had them then. I saved filters. I opened the measurements file and did the same. Here is what my L/R Fronts look like with the target curve and curtains.  So without moving the UMIK-1 I measured the left channel for each of the following filter sets made from the same 9/23 measurements: Orange 9/23 project, filters with 3.1.2 Green 9/23 measurements, filters with 3.2.2 Blue 9/23 project, filters with 3.2.2  The original 9/23 with 3.1.2 is pretty flat in the low bass. The other two have very different response below 40Hz, and they're different from each other. This issue is only with the Left channel. Right channel measures the same with all three versions of filters. And look at the measured response in Dirac. With such a peak and no dips in that range I would expect Dirac to correct to dead flat. |
|
|
|
Post by hsamwel on Feb 1, 2022 13:13:42 GMT -5
Installed Dirac 3.2.2 on my Windows 11 laptop.
I loaded my saved projects.. All my curves were gone! Also curtain settings was reset.
Dirac opened all my saved projects with the default Dirac curve.
These were all saved with 3.1.2 in september.
Had to redo them all and save them again.
|
|
|
|
Post by marcl on Feb 1, 2022 13:49:18 GMT -5
Installed Dirac 3.2.2 on my Windows 11 laptop. I loaded my saved projects.. All my curves were gone! Also curtain settings was reset. Dirac opened all my saved projects with the default Dirac curve. These were all saved with 3.1.2 in september. Had to redo them all and save them again. Yes I found the same. But I have all my curves saved to a folder so I opened them again and reset the curtains. Even before 3.2.2 I think it always reset the curtains. Have you done REW after to compare same project from 3.1.2 to 3.2.2? I saw some weirdness on the left front only. |
|
|
|
Post by hsamwel on Feb 1, 2022 18:39:40 GMT -5
Installed Dirac 3.2.2 on my Windows 11 laptop. I loaded my saved projects.. All my curves were gone! Also curtain settings was reset. Dirac opened all my saved projects with the default Dirac curve. These were all saved with 3.1.2 in september. Had to redo them all and save them again. Yes I found the same. But I have all my curves saved to a folder so I opened them again and reset the curtains. Even before 3.2.2 I think it always reset the curtains. Have you done REW after to compare same project from 3.1.2 to 3.2.2? I saw some weirdness on the left front only. I will do both a complete new Dirac calibration and REW control measurement as I have had some changes done in my room. New furniture and some other changes. I have not checked the current recalculation. |
|
|
|
Post by marcl on Feb 1, 2022 18:41:47 GMT -5
Yes I found the same. But I have all my curves saved to a folder so I opened them again and reset the curtains. Even before 3.2.2 I think it always reset the curtains. Have you done REW after to compare same project from 3.1.2 to 3.2.2? I saw some weirdness on the left front only. I will do both a complete new Dirac calibration and REW control measurement as I have had some changes done in my room. New furniture and some other changes. I have not checked the current recalculation. I'm doing a full recal tomorrow too. It will be with FW 2.5 and 3.2.2. |
|
|
|
Post by hsamwel on Feb 1, 2022 19:06:38 GMT -5
Installed Dirac 3.2.2 on my Windows 11 laptop. I loaded my saved projects.. All my curves were gone! Also curtain settings was reset. Dirac opened all my saved projects with the default Dirac curve. These were all saved with 3.1.2 in september. Had to redo them all and save them again. Yes I found the same. But I have all my curves saved to a folder so I opened them again and reset the curtains. Even before 3.2.2 I think it always reset the curtains. Have you done REW after to compare same project from 3.1.2 to 3.2.2? I saw some weirdness on the left front only. Btw.. I can’t save a single curve to use.. Each speaker group in my setup has different curves. I mostly steer how the bass falls of depending on how my room response is with each speaker. I try not to have any speaker over work the low end. If it was only the curtains that reset it would be fine.. Having all my curves reset is atleast 1 hours of work, if not more. |
|
|
|
Post by marcl on Feb 1, 2022 19:40:56 GMT -5
Yes I found the same. But I have all my curves saved to a folder so I opened them again and reset the curtains. Even before 3.2.2 I think it always reset the curtains. Have you done REW after to compare same project from 3.1.2 to 3.2.2? I saw some weirdness on the left front only. Btw.. I can’t save a single curve to use.. Each speaker group in my setup has different curves. I mostly steer how the bass falls of depending on how my room response is with each speaker. I try not to have any speaker over work the low end. If it was only the curtains that reset it would be fine.. Having all my curves reset is atleast 1 hours of work, if not more. You can save a target curve for each speaker group. I have seven. You select save target then select the group and give the target the name of the group. I save them in a folder dated when I did the calibration. Unfortuately you have to load them one at a time too for each group. I wish I could just point to a folder and tell it to load targets and load them all. But it's faster than having to do them over. |
|
|
|
Post by marcl on Feb 3, 2022 10:25:51 GMT -5
I did a full Dirac calibration with FW 2.5 and DL 3.2.2. and the results were excellent! Both measurements and listening tests suggest this combination is solid and working well. The Bass Management bug - which in my case raised bass levels as much as 8-9db with FW 2.3 - is now down in the midst of the normal room +/- response. It's actually +3db as it was with FW 2.2 and before, but hardly noticeable in practical terms. Note that in my system, I have Center Sub=LFE and bass management goes to the two Large Fronts. I've been experimenting with a cork head at the MLP (see further discussion in the Measurements thread). This is not for any sort of "binaural" measurement, but rather to use the head to block sound - especially reflections - from the opposite side. I used the cork head (AKA "Agent 13") in conjunction with the Dirac measurements. I took 11 measurements with four of them being high/low forward of MLP and two high behind MLP, the usual. The MLP measurement was at the "forehead" line of sight to all speakers. Then I did one measurement at each ear, intentionally blocking sound from the opposite side. And finally a measurement 1ft away from each ear, at ear level. Attachment DeletedAttachment DeletedAttachment DeletedThe results were very interesting! Using all 11 measurements I got a big dip at 40Hz in the L/R channels, an obvious overcorrection because my room has a huge PEAK at 40Hz. Experimenting removing combinations of the four measurements near the head resulted in interesting combinations of either the left or right channel (or both) having strange cancellations. The curious conclusion was to use a near-ear measurement on one side of the head, and the ear +1ft measurement on the opposite side. 9 measurements total. Here's the response with Variable Smoothing, along with the Dirac measurement of the Front L/R for reference of the starting point.   Listening tests .... so far I'd say this calibration - in conjunction with addition of two strategically placed absorbers - results in the best sound yet in terms of clarity and soundstage. p.s. Prior to Dirac calibration Agent 13 was used in a series of about 40-50 measurements to assess a reflection I was seeing at 8ms with left and right channel measurements (as measured in the ETC plot in the impulse response of each channel). The absorbers were placed to eliminate this reflection. The result was better impulse response correction for the L/R channels and I believe this contributed to improved soundstage in the final result. More listening tests to follow, with and without the two new absorbers. |
|
