• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

My first attempt at Dirac this morning (on the 3800)

EWL5

Addicted to Fun and Learning
Joined
Jan 25, 2023
Messages
628
Likes
403
So I found some quiet time to finally incorporate DL (full bandwith) into my 3800. Here are the steps I took:

1) Installed DL onto my Windows 10 laptop. The software found my 3800 on the network and I selected it.
2) I attached the UMIK1 to the laptop and proceeded to download the 90 deg calibration file once I put in the serial number for the mic (the other file is better for stereo music).
3) Directed the software to the .txt calibration file and hit Next.
4) I decided to choose the option for "Wide focus" to get my feet wet and have fewer positions to calibrate.
5) Got a "Msg 210, error code 106" that claimed there was too much ambient sound. I increased the mic input on Windows to 100% and proceeded to increase master volume on the 3800 to -40, which did the trick!
6) I measured only 5 positions, center, 2 to the right of center, and 2 to the left of center.
7) I noticed the subwoofer wasn't emitting any pink noises so I increased the dial to give something for the mic to read. I noticed the amplitudes of the SW during the calibration were a bit higher than the rest of the speakers (will fix in next attempt).
8) I studied the curves for each speaker and proposed filters. I exported them to the 3800.
9) Listened to Rebel Moon (scene where the drednought ship enters the atmosphere) and noticed the mid-bass was missing! I suspected this would happen because the handoff from the LFE to the actual speakers were bad according to the curves.

Current theory: I believe the speaker distances and LFE crossovers are inherited from the prior Audyssey calibration (someone please correct if I'm wrong). As such, Audyssey had determined that my L/C/R speakers should have a 260 Hz crossover and my surrounds should be 60 Hz. I've kept this setting since but was hoping to see what impact it would have on Dirac. Unless someone says otherwise, I'm planning to update the crossovers to 80 Hz for all speakers and rerun Dirac. Thoughts?
 
Since you don't have DLBC, it will in fact rely on the crossovers as they are set in the AVR -- and those seem wonky. Try setting everything to 80 Hz and leaving it there.

Also, here's a good guide on how to optimize your microphone measurements with Dirac:


From the Dirac Live thread on AVS, which is also a great resource:

Thanks for the additonal Dirac reads! As I only have one SW and I know the reason for the missing midbass, I don't feel compelled to get DLBC. I can imagine some folks panicking and getting DLBC after what I just experienced w/o knowing the cause!
 
Second attempt:
After previously ignoring the leveling screen (because of mistaken assumption numbers would be carried over from Audyssey), I followed the procedure on mehlau.net and managed to level the speakers decently above the noise floor. I had changed the SW crossover to 80 Hz in the Manual Setup and also peaked at the Dirac filter (first one and made sure it was still Speaker Preset 2 for the 3800). I accepted the filter and exported to the 3800.

I demoed the drednought scene from Rebel Moon and despite it being markedly better, it was still missing midbass/LFE. Even Billy Eilish's "Therefore I Am" was lacking any punch whatsover! I proceeded to check the crossover setting and was shocked to see the old settings of 260Hz for the fronts and 60 Hz for the rears and top speakers managed to come back! There appears to be a bug where if you "check" the filter page, it will revert to the old crossover settings despite changing and clicking away to save. I believe if the crossover is "saved" before creating the Dirac filter (without going anywhere else), the crossover setting will hold. This is either some kind of nasty bug or gamesmanship to get folks to invest in DLBC by Denon/Dirac. Anyone else experience this bug?

I tried sharing my filter curves but every time I hit the "Snapshot" button, I have no clue where the pic goes (I checked all the obvious places in Windows like "Pictures", etc). I can certainly do screenshots but I figured the snapshot was supposed to give nice curves. Could use help with this!
 
Second attempt:
After previously ignoring the leveling screen (because of mistaken assumption numbers would be carried over from Audyssey), I followed the procedure on mehlau.net and managed to level the speakers decently above the noise floor. I had changed the SW crossover to 80 Hz in the Manual Setup and also peaked at the Dirac filter (first one and made sure it was still Speaker Preset 2 for the 3800). I accepted the filter and exported to the 3800.

I demoed the drednought scene from Rebel Moon and despite it being markedly better, it was still missing midbass/LFE. Even Billy Eilish's "Therefore I Am" was lacking any punch whatsover! I proceeded to check the crossover setting and was shocked to see the old settings of 260Hz for the fronts and 60 Hz for the rears and top speakers managed to come back! There appears to be a bug where if you "check" the filter page, it will revert to the old crossover settings despite changing and clicking away to save. I believe if the crossover is "saved" before creating the Dirac filter (without going anywhere else), the crossover setting will hold. This is either some kind of nasty bug or gamesmanship to get folks to invest in DLBC by Denon/Dirac. Anyone else experience this bug?

I tried sharing my filter curves but every time I hit the "Snapshot" button, I have no clue where the pic goes (I checked all the obvious places in Windows like "Pictures", etc). I can certainly do screenshots but I figured the snapshot was supposed to give nice curves. Could use help with this!
Without DLBC, Audyssey could do a better job managing the bass range, even with one subwoofer. It is not only audible but also visible from my REW measurements.
 
Last edited:
Second attempt:
After previously ignoring the leveling screen (because of mistaken assumption numbers would be carried over from Audyssey), I followed the procedure on mehlau.net and managed to level the speakers decently above the noise floor. I had changed the SW crossover to 80 Hz in the Manual Setup and also peaked at the Dirac filter (first one and made sure it was still Speaker Preset 2 for the 3800). I accepted the filter and exported to the 3800.

I demoed the drednought scene from Rebel Moon and despite it being markedly better, it was still missing midbass/LFE. Even Billy Eilish's "Therefore I Am" was lacking any punch whatsover! I proceeded to check the crossover setting and was shocked to see the old settings of 260Hz for the fronts and 60 Hz for the rears and top speakers managed to come back! There appears to be a bug where if you "check" the filter page, it will revert to the old crossover settings despite changing and clicking away to save. I believe if the crossover is "saved" before creating the Dirac filter (without going anywhere else), the crossover setting will hold. This is either some kind of nasty bug or gamesmanship to get folks to invest in DLBC by Denon/Dirac. Anyone else experience this bug?

I tried sharing my filter curves but every time I hit the "Snapshot" button, I have no clue where the pic goes (I checked all the obvious places in Windows like "Pictures", etc). I can certainly do screenshots but I figured the snapshot was supposed to give nice curves. Could use help with this!
You might want to consider a factory reset, followed by re-uploading your Dirac filters on Slot 1. I've read reports of bugs with losing filter/crossover settings on these receivers when changing quick select, for instance. Some of those issues have been claimed fixed by various D+M firmware updates.
 
Without DLBC, Audyssey could do a better job managing the bass range, even with one subwoofer. It is not only audible but also visible from my REW measurements.
260 Hz crossover on his main front left+right speakers is clearly broken.
 
You might want to consider a factory reset, followed by re-uploading your Dirac filters on Slot 1. I've read reports of bugs with losing filter/crossover settings on these receivers when changing quick select, for instance. Some of those issues have been claimed fixed by various D+M firmware updates.
I'll do that but hope to save the Audyssey settings to USB as it's the best sounding room correction at this point (still reserving judgment).
 
Ahead of tomorrow morning's redo of Dirac, here are my curves. Let me know if any insight can be gleaned from these.

Left Front
1703713630744.png


Right Front
1703713674477.png


Center
1703713735811.png


Surround Right
1703713813915.png


Surround Left
1703713900870.png


Top Front Right
1703713961096.png


Top Front Left
1703714114777.png


Subwoofer
1703714156567.png


Note that rear surrounds and top front speakers are in-ceiling speakers (don't know the brand). My L/C/R is a Martin Logan SLM X3 and the sub is a DefTech Supercube 1.
 
Ahead of tomorrow morning's redo of Dirac, here are my curves. Let me know if any insight can be gleaned from these.

Left Front


Right Front


Center


Surround Right


Surround Left


Top Front Right


Top Front Left


Subwoofer


Note that rear surrounds and top front speakers are in-ceiling speakers (don't know the brand). My L/C/R is a Martin Logan SLM X3 and the sub is a DefTech Supercube 1.

Oh wow, I didn't realize your front LCR was a soundbar. The -3 dB point is 120 Hz. Looks like your sub plays up to 200 Hz, though no idea how well it does up that high. Maybe try setting your crossover to 160 Hz and see what that does.
 
The dialogue in movies after this morning is super clear (just missing the transition LFE frequencies).
 
Made the mistake of saving the settings to USB this morning instead of last night. It's been 40 minutes and it's still going?!!! Does the screensaver interrupt progress?

Before I saved settings, I increased the crossover to 150 Hz (there is no 160Hz as the next step up is 180Hz?!!) and there was marked improvement in the midbass region. I will try 180Hz next and I might be able to get away from needing to redo calibration if I'm happy.

Does the crossover need to be the "correct" setting before Dirac calibration?
 
Made the mistake of saving the settings to USB this morning instead of last night. It's been 40 minutes and it's still going?!!! Does the screensaver interrupt progress?

Before I saved settings, I increased the crossover to 150 Hz (there is no 160Hz as the next step up is 180Hz?!!) and there was marked improvement in the midbass region. I will try 180Hz next and I might be able to get away from needing to redo calibration if I'm happy.

Does the crossover need to be the "correct" setting before Dirac calibration?
So long as you're not using DLBC, you can change the crossover at will. Only if you add DLBC will you need to change the crossover inside the Dirac software. 150 Hz should be a pretty good setting since your subwoofer should still be able to play at that range and you will be 30 Hz past the low-end of your LCR.
 
Ahead of tomorrow morning's redo of Dirac, here are my curves. Let me know if any insight can be gleaned from these.

Left Front
View attachment 337470

Right Front
View attachment 337471

Center
View attachment 337472

Surround Right
View attachment 337473

Surround Left
View attachment 337474

Top Front Right
View attachment 337475

Top Front Left
View attachment 337476

Subwoofer
View attachment 337477

Note that rear surrounds and top front speakers are in-ceiling speakers (don't know the brand). My L/C/R is a Martin Logan SLM X3 and the sub is a DefTech Supercube 1.
The X3's spec -3dB roll off frequency is 120Hz and that's where you should cross them over for optimal minimum phase correction at the LP. Room correction will sound best when it filters out the room from the speaker. Extra bass obtained by room reflections will introduce delays to the bass frequencies.
 
Made the mistake of saving the settings to USB this morning instead of last night. It's been 40 minutes and it's still going?!!! Does the screensaver interrupt progress?

Before I saved settings, I increased the crossover to 150 Hz (there is no 160Hz as the next step up is 180Hz?!!) and there was marked improvement in the midbass region. I will try 180Hz next and I might be able to get away from needing to redo calibration if I'm happy.

Does the crossover need to be the "correct" setting before Dirac calibration?
You are using the flat target curve, so if you want more bass, try the popular tilted curve. It us easy to grab that blue handle on the left and drag it upward until you get about 6 dB increase at about 30 Hz. I think with that simple change, along with using 120 zHz crossover, (may be 130 Hz will work too) Hz, it will sound better to you overall.
 
Last edited:
Ahead of tomorrow morning's redo of Dirac, here are my curves. Let me know if any insight can be gleaned from these.

Left Front
View attachment 337470

Right Front
View attachment 337471

Center
View attachment 337472

Surround Right
View attachment 337473

Surround Left
View attachment 337474

Top Front Right
View attachment 337475

Top Front Left
View attachment 337476

Subwoofer
View attachment 337477

Note that rear surrounds and top front speakers are in-ceiling speakers (don't know the brand). My L/C/R is a Martin Logan SLM X3 and the sub is a DefTech Supercube 1.

As others have noted, most would not like a system measureing flat at the main listening position (it is unnaturally gright sounding/lacking bass). As @peng mentioned, grap the shelf filter tool for bass and drag it up to add bass (4-6db). You might also consider the shelf tool on the high frequencies too and maybe have it at -1db to -3db which would be a closer match to the roll off of the speakers. Adding back the missing bass should make a very big difference I would think.
 
@peng and @Dj7675, appreciate the suggestion! I was hoping to not have to tinker too much since I spent so much money. Despite my best efforts, it looks like I'll have to "get with the program", especially since I invested in the UMIK-1.

Just as a confirmation: what you are asking me to do will simply bring me back to what I appreciated w/near-free Audyssey, right? Still not understanding why Dirac exists!
 
@peng and @Dj7675, appreciate the suggestion! I was hoping to not have to tinker too much since I spent so much money. Despite my best efforts, it looks like I'll have to "get with the program", especially since I invested in the UMIK-1.

Just as a confirmation: what you are asking me to do will simply bring me back to what I appreciated w/near-free Audyssey, right? Still not understanding why Dirac exists!
In my experience, Dirac is more effective in getting smoother bass with much less effort. You can achieve the same or eveb better in some ways if you spend a lot if time tweaking with the app and Ratbuddyssey or use the $200 MultEQ X.

Dirac also does a better job in the high frequencies for me too. It is a 3 minutes job to implement the sort of "Harman curve" using the 6 dB shelf filter. Actually I thought that's done by default, or selectable by one simple click, but I guess it depends on the version. I am confident you will like the result.
 
Back
Top Bottom