Audyssey One - Nexus (Automated hi-res Denon/Marantz calibration with REW/Umik-1/2)

[anphex]

Can I ask what AVR you are using and how many subs?

Denon X3700 and two subs.

 

[Andysu]

can nexus handle baby boom channels 2 and 4

 

[anphex]

Alright, I did a little more testing.
It all falls apart when I disable my E-APO Loudness Correction. But NOT because of Nexus, but because of the limitations of Audyssey.

When I rely solely on Denon Dynamic EQ, even already paired with the 10dB Harman Bass Boost - it's as if there is no bass.
I tried exporting different calibrations, even setting different volume offsets when Nexus asks for them before optimization but it's all the same.
Nexus can't fix the very stiff and unflexible Dynamic EQ.

Dynamic EQ will be disabled or at least unnoticable past -10dB Denon Volume Level. This is why I am pretty certain that the issue lies in:
a) Nexus lowers the volume significantly for the optimizations to work
b) my amp is set to 19dB and I am 90 % sure Audyssey expects an amp with about 30dB gain (denon has 32dB, I think)

These two factors mostly cause a bypass for Dynamic EQ, which is essential for having that warm, bass rich sound even at normal listening levels.
To be clear, not even the stock Audyssey Dynamic EQ works with a regular calibration. So I guess the crucial part is the gain.

I think this is pretty important since most never amps only achieve their +100dB Sinad performance with gain in the range 10-20dB, which already troubles Audyssey a lot.

BUT one very important disclaimer. WHEN you use tools like E-APO loudness correction instead of Dynamic EQ, even when you have a troublesome setup (weak gain), Nexus is GOAT. But it limits you to PC-only. When using consoles (PS5, XBOX) and maybe a streamer, the above mentioned problems arise.

Edit: I will try one more thing and see if this works. It will make a custom "flat" house curve (Nexus enforces a EQ curve even if you just want it flat, not sure why) and see if it avoid lowering the volume and see if MultEQ works then.
Edit 2: Nope, no improvement. But it shows that applying a EQ Curve seems to make no real difference in case it.
Edit 3: Fork it, I am ordering a second NCX500 stereo amp from Audiophonics and setting both amps to 29dB gain.

 

[Timmy]

[A1 Evo] Should give it a try then.

Went through this hard-to-stand video (youtu.be/lmZ5yV1-wMI?feature=shared) and somehow ended up with two new .ady-files,

multeq1_20241020_1256_A1EvoMaestroMJmaster.ady
multeq1_20241020_1256_A1EvoMaestroMJdEQ0dB.ady

Loaded them up to my AVR, and the result in both cases is that the volume is even lower than with the Audyssey MultEQ Editor app calibration

 

[anphex]

Went through this hard-to-stand video (youtu.be/lmZ5yV1-wMI?feature=shared) and somehow ended up with two new .ady-files,

multeq1_20241020_1256_A1EvoMaestroMJmaster.ady
multeq1_20241020_1256_A1EvoMaestroMJdEQ0dB.ady

Loaded them up to my AVR, and the result in both cases is that the volume is even lower than with the Audyssey MultEQ Editor app calibration

Can confirm, there really is some weird case of headroom.

Edit: I will do the whole measurement again when I have the second NCX500 amp here and see if it all boils down to amp gain mismatch.

 

[Timmy]

Can confirm, there really is some weird case of headroom.

That way, the whole thing is really just a frustrating waste of time.

I re-did my first run, but now I ticked the »Force small fronts« check-box. Guess what the result of that is ... Even lower volume than the low-volume result this afternoon ...

 

[OCA]

That way, the whole thing is really just a frustrating waste of time.

I re-did my first run, but now I ticked the »Force small fronts« check-box. Guess what the result of that is ... Even lower volume than the low-volume result this afternoon ...

Timmy, what do you do for a living?

 

[Timmy]

Did another run of the script ...

Made a copy&paste of the »Progress log« (why isn’t that log saved per default, btw?), to have a close look at the »warnings«:

[WARNING!] 'Allow' multiple file downloads when prompted to prevent the browser from blocking saving calibrated files at the end of the optimization.

[WARNING!] Please keep REW on 'SPL & Phase' tab, close any child windows and stay on this web page until optimization is completed!

[WARNING!] SW1: 3.62 meters, please SWITCH this subwoofer's POLARITY!

[WARNING!] Changing the above crossover frequencies manually in the receiver is not recommended, adjust them with Evo customization options instead!

[WARNING!] The calibration file with '_DEQ0dB.ady' extension is for LOW VOLUME LISTENING, use the 'master' calibration file at all other times.

Not quite sure if they are »warnings« actually. Fortunately, they appear in orange only, not in red

Not sure also, if those warnings can somehow »justify« the bad calibration results regarding volume level. With those calibration loaded I have to turn the volume knob to its 50% position nearly, to have a decent volume. That just can’t be it ...

 

[anphex]

@OCA

I had hoped you were in this board. Nice. I would like to take the opportunity to ask a few questions:

- When applying a target curve, is the final output result normalized to the peaks? Meaning, if I chose Harman 6dB, will the volume calibration be set to the 6dB boost as a cap, meaning the mids and highs will be quieter by 6 dB? I also have the feeling that the target curve is applied in REW, but then somehow not really applied in Audyssey. Though I have to make a control measurement in the next days to verify my claim.

- Does this tool take into account distortion? For example, I know my front speakers could handle full range, but everything below 40Hz has rather high amount of measurable distortion. My subs for example would handle this range below 60 Hz more gracefully. Is there or are there plans to implement a forced crossover frequency OR "optimize for minimal distortion"?



Small suggestions:
- Different colors for warnings and instructions - maybe color instructions purple so it differs from the warnings.
- Make a cleanup of REW after doing optimization so you can quickly start a new optimization.
- Maybe even provide a one-click solution for cross-correlate and vector align.



In the end, I really like the result so far. We just need to get the DynamicEQ right. But I'll admit, I have a strong hunch that Audyssey expects high gain amps +30dB. Anything lower and the bass boost offset is too weak.

 

[OCA]

I can't even understand if you are talking about Maestro or Nexus from the remarks but Maestro keeps the measurement volume intact which is 75dB determined by Audyssey's own sweeps. Nexus doesn't care about the measurement volume or the final volume and neither should you. It's a room correction tool, volume is aligned by a knob in the AVR. Target volume where everything is equalized at is calculated with a very sophisticated algorithm which puts minimum load on the speakers and the sub and guarantees the most immersive Atmos bubble.

Crossovers and everything else is fully customizable in both Evo versions (although not recommended) and there are detailed explanations in the logs for both on how.

Speakers don't distort when fed with full range signals unless you are using cheap sub + satellite kits, they just don't produce low frequencies loud enough. Audyssey's own sweeps start from 0Hz.

Warnings are in orange and points to actions that should be done by the user ie setting LFE + Main or bass extraction lpf in the AVR menu. Everything else will be done by Evo. Blue prompts are important which an enthusiast might wanna read (but not necessary) and white prompts are just info.

Final results stay in REW so you can check them. Removing them is a one click button in REW if you need to ("Remove All").

Cross correlation may get impulse peaks wrong especially with subwoofer measurements or impulses with strong reflections that's why I kept it manual in Nexus, in Maestro all this is automated.

 

[Andysu]

 

[OCA]

View attachment 400497

Actually A1 Kenobi was considered as a name for Nexus in a forum far away

 

[Timmy]

Does this tool take into account distortion?

As you mention distortion:

My speaker setup contains two »Heights«, sort of tiny Cantons (CD10) – and with the latest Evo calibration they seem to emit distortion only …

 

[Andysu]

this is first generation got wait for nexus 6th Eq

 

[OCA]

Again, is this Nexus or Maestro? In Nexus you can adjust your measurement level (AV volume), mic gain, sound card volume among others to remove distortion from your measurements unless of course there's something wrong with the speaker. Is the distortion linear or non-linear? What was the headroom REW showed during their measurements? Was it not in the green? If you are talking about resonances during measurement that's normal with tiny speakers with sweeps, doesn't mean it will be carried to the calibration after they're crossed over with the sub. I have one of the most expensive ceiling speakers from Focal and they resonate like crazy during the lower bass frequencies of the log sweep during measurement.

 

[anphex]

Again, is this Nexus or Maestro? In Nexus you can adjust your measurement level (AV volume), mic gain, sound card volume among others to remove distortion from your measurements unless of course there's something wrong with the speaker. Is the distortion linear or non-linear? What was the headroom REW showed during their measurements? Was it not in the green? If you are talking about resonances during measurement that's normal with tiny speakers with sweeps, doesn't mean it will be carried to the calibration after they're crossed over with the sub. I have one of the most expensive ceiling speakers from Focal and they resonate like crazy during the lower bass frequencies of the log sweep during measurement.

If you meant me, I am only talking about Nexus since I didn't try Maestro yet and this thread is about Nexus only (I think?).
I don't mean the whole distortion spectrum, just the bass range where, for example, a front speaker could play down to 20 Hz but only by giving it bass boost and thus raising the distortion. But now you also have subwoofers who can handle the 120 Hz-20 Hz range with far less distortion. Nexus might prioritize dip fixing and crossover alignment instead of phasing out the distortion of the front speaker, maybe setting it to 40 Hz crossover. That's what it meant.

But as I was writing this it came to me that crossover- and dip-fix optimization might also imply you also minimize distortion anyway...
But as I already mentioned multiple times - I will retry Nexus once the second NCX500 amp is here. With my prior setup Audyssey and Nexus applied these levels:

Front: -9dB
Surrounds: +9dB
Subs: -12dB (even alerting me that it's still too loud even though in REW the sub was significantly below the fronts)

Maybe there is a tip on how to level match your subwoofers before measurement?

I will report back when my amp arrives during this week. Also, thank you for taking your time to make such a tool!

 

[OCA]

If you meant me, I am only talking about Nexus since I didn't try Maestro yet and this thread is about Nexus only (I think?).
I don't mean the whole distortion spectrum, just the bass range where, for example, a front speaker could play down to 20 Hz but only by giving it bass boost and thus raising the distortion. But now you also have subwoofers who can handle the 120 Hz-20 Hz range with far less distortion. Nexus might prioritize dip fixing and crossover alignment instead of phasing out the distortion of the front speaker, maybe setting it to 40 Hz crossover. That's what it meant.

But as I was writing this it came to me that crossover- and dip-fix optimization might also imply you also minimize distortion anyway...
But as I already mentioned multiple times - I will retry Nexus once the second NCX500 amp is here. With my prior setup Audyssey and Nexus applied these levels:

Front: -9dB
Surrounds: +9dB
Subs: -12dB (even alerting me that it's still too loud even though in REW the sub was significantly below the fronts)

Maybe there is a tip on how to level match your subwoofers before measurement?

I will report back when my amp arrives during this week. Also, thank you for taking your time to make such a tool!

Evo doesn't calculate full range option if the front mains steady state 12dB/octave roll off is above 40Hz. Nexus will do that for all speakers below 60Hz (I increased that in Nexus because people even with smaller front speakers demanded this mode). So, LFE + Main mode is only for front mains and only if they are proper 3-way floor standers. And it's like adding two more subs in your room when done right.

The applied levels show some sort of amp gain mismatch but nothing to worry about, as long as you don't get 12dB limit exceeded warning, the correction stays optimal.

WIth Nexus lossless Atmos sweeps, there's still confusion actually about which LFE sweep to use for the sub (that's why there are two SW sweeps with 10dB difference) because some amps seem to apply 10dB boost to LFE sweeps and some just don't if they are full range. I would use SWx.mlp sweep in the video download folder unless you see a reason not to with the level shown in REW during measurement. You can't go wrong twice though as you will hear a 10dB misalignment quite easily.

PS Sub sweeps will always give this warning when they are full range (0-24000Hz) as they don't produce much beyond 200-250Hz and REW detects higher room noise than the sub's output from 250Hz to 24kHz and pops that warning. You can ignore it. Actually, if you don't get it, you are probably measuring your sub too loud. Btw, don't ignore it if it pops after a speaker measurement.

 

[anphex]

Evo doesn't calculate full range option if the front mains steady state 12dB/octave roll off is above 40Hz. Nexus will do that for all speakers below 60Hz (I increased that in Nexus because people even with smaller front speakers demanded this mode). So, LFE + Main mode is only for front mains and only if they are proper 3-way floor standers. And it's like adding two more subs in your room when done right.

The applied levels show some sort of amp gain mismatch but nothing to worry about, as long as you don't get 12dB limit exceeded warning, the correction stays optimal.

WIth Nexus lossless Atmos sweeps, there's still confusion actually about which LFE sweep to use for the sub (that's why there are two SW sweeps with 10dB difference) because some amps seem to apply 10dB boost to LFE sweeps and some just don't if they are full range. I would use SWx.mlp sweep in the video download folder unless you see a reason not to with the level shown in REW during measurement. You can't go wrong twice though as you will hear a 10dB misalignment quite easily.

PS Sub sweeps will always give this warning when they are full range (0-24000Hz) as they don't produce much beyond 200-250Hz and REW detects higher room noise than the sub's output from 250Hz to 24kHz and pops that warning. You can ignore it. Actually, if you don't get it, you are probably measuring your sub too loud. Btw, don't ignore it if it pops after a speaker measurement.

Oh wow, I never thought of it this way. Why just use one low end provider when you can use all of them? It now clicked! Will try that soon with the new amps.

For sub measurement I used the standard SWx. Also last question: In the video you instruct to set REW measurements to 0.00dBfs and in the change log later you say -12.00dB, but doesn't the Atmos sweep file have a fixed volume?
And last question: My interface and Mic have rather high self noise and I need to increase measurement volume moderately over what I would use as normal listening volume. What is more important? Setting the AVR on the final listening volume during measurement or getting a good SNR? Of course I tried multiple measurements for each speakers and did the correlation and vector align to center position of each speaker measurement but the noise floor always stayed rather high, even with "capture noise floor".

 

[OCA]

0dbFS was found too loud by many so I made new ones with -12dBFS which is also REW default and since Mac owners have to use REW's internal sweeps for subwoofer (Macs can't play lossless Atmos sweeps and lossy LFE sweep has a 120Hz internal lpf), it seemed more convenient that way. It doesn't matter whether you turn the AVR volume by 12dB or replace -12dBFS sweep with 0dBFS ones or change external amp gain by 12dB. As long as you measure with REW not warning for negative headroom (turns red) and your speakers don't sound like they are way over their limit, you are fine. It's good practice to use "capture noise floor". High room noise is not a problem so long as REW captures it. These log sweeps are very powerful in eliminating noise from the signal. If REW did not give you a low signal to noise ratio warning, the measurement is good enough for calibration purposes. The importance of AV volume level is to be in line with your preferred target curve and equal loudness. In general the louder the measurement volume, the higher the signal to noise ratio but you have to keep within the efficient limits of your speakers drivers. Measured frequency responses should be more or less identical with just an SPL offset when you increase the volume. If the FR shape starts deviating, that's the volume level you should avoid measuring for that speaker because it means one of its drivers started to give up.

 

[emptygreat]

0dbFS was found too loud by many so I made new ones with -12dBFS which is also REW default and since Mac owners have to use REW's internal sweeps for subwoofer (Macs can't play lossless Atmos sweeps and lossy LFE sweep has a 120Hz internal lpf), it seemed more convenient that way. It doesn't matter whether you turn the AVR volume by 12dB or replace -12dBFS sweep with 0dBFS ones or change external amp gain by 12dB. As long as you measure with REW not warning for negative headroom (turns red) and your speakers don't sound like they are way over their limit, you are fine. It's good practice to use "capture noise floor". High room noise is not a problem so long as REW captures it. These log sweeps are very powerful in eliminating noise from the signal. If REW did not give you a low signal to noise ratio warning, the measurement is good enough for calibration purposes. The importance of AV volume level is to be in line with your preferred target curve and equal loudness. In general the louder the measurement volume, the higher the signal to noise ratio but you have to keep within the efficient limits of your speakers drivers. Measured frequency responses should be more or less identical with just an SPL offset when you increase the volume. If the FR shape starts deviating, that's the volume level you should avoid measuring for that speaker because it means one of its drivers started to give up.

Hi Serko! Thank you for your great work. I was previously running a version of Evo, and recently made my first attempt with Nexus on 1.8. Currently running version 2.4.
I did get some strange SW distances from Nexus 2.4, more precisely 5.37m if I leave all checkmarks unticked which sets Front speakers as Large and therefore LFE+Main, and 6.09 meters if I check "Force small front speakers" option. Subwoofer is physically placed 2.35 meters from MLP.
After reading this comment, I do wonder if this could be related to the sweep I've used for measurement, as I was able to play the LFE atmos mkv sweep through VLC on a Mac connected to the AVR via HDMI, using Denon X3800h (encoded) – which I believe is passthrough – as the audio device output.
Do you believe I should remeasure using REW's sweeps? I don't actually need atmos sweeps as my system is 5.1.
Thank you!