Audyssey Manual Calibration “OCA’s REW + Audyssey Awesomeness”

[Hyacin]

M23R rules (I own one) but I am not sure the benefits justify the extra cost realtive to an Umik-2 unless you have really low punching subs.

I have a pair of PB-2000s on the roadmap

I thought about a UMIK-2 but you mentioned USB timing issues in one of your videos, so after digging a tiny bit into that, plus the range of the SVS subs, it was enough for me to justify it :-D

 

[Hyacin]

No experience with shakers. There's an Audyssey One FB group though and you might find relevant info there. I have never heard A1 producing incompatible calibration files.

Just a case of garbage in garbage out. I need to figure out why the damn MultiEQ app is producing said garbage :-/

 

[DeLub]

Hi @OCA,

Just tried Evo v2.0. As always great work!

Couple of remarks if I may:

• Your script applies smoothing when someone wants to correct a large range. I get that when the upper limit is 20kHz, but I think you have set it a bit conservatively. I would argue to apply smoothing only when the range is above 500Hz (I experimented with this on my setup an subjectively found that I want to correct until 400Hz without smoothing)
• I think you try to mimic the levelling the same way Audyssey does this; correct? Audyssey seems to level the channels between 500 and 2000Hz. So that are 2 octaves instead of 4 used in the script.
  
  
  
  link
  
  I'm not sure but I guess this can be changed here:
  

  const volShift = await postNext('Align SPL', [i], {frequencyHz: "1000", spanOctaves: "4", targetdB: 75.0});

  But perhaps this should also be changed here, in updateADY()? (Not sure what is happening here in the code):
  

      for (k = 0; k <= 1531; k++) {
        const fltMagnitude = data.getFloat32(k * 4);
        fltArray.push([freqArray[k], fltMagnitude]);
        if (k >= 1000 && k <= 1191) {spSum += (fltMagnitude)};
      }

• I understand that the XO algorithm changed significantly. My speakers are correctly wired and previous versions of Audyssey One never mentioned anything. But this version suggested better performance when inverting polarity of my sub and pair of surrounds. I did that and a short test sounds good, but can you explain why that is, please?

 

[OCA]

I have a pair of PB-2000s on the roadmap

I thought about a UMIK-2 but you mentioned USB timing issues in one of your videos, so after digging a tiny bit into that, plus the range of the SVS subs, it was enough for me to justify it :-D

I have a Umik-1, Umik-2, a Behringer and I have not used them once since I got the M23R. You are aware that you will need a descent audio interface with it, too right?

 

[OCA]

I want to correct until 400Hz without smoothing

With or without smoothing, it will look good but not sound better but actually worse. You can change it as you with just editing this line:

endFrequency > 250 ? smoothing = "Var" : smoothing = "None";

If you want to do some half descent filtering above 200Hz, you need to apply FDW to the response first, this will smooth it, too but do much more behind the doors.

Anti auto-leveling works correctly and compensates for what you have referred to above (for speakers, subs are different).

 

[PHD]

v2.2 was released but I couldn't find a download link yet. I'm guessing it's because the script can now be run online:

A1 Evo - OCA 2024

 

[icebaer72]

Simple answer is standardization with Audyssey calibration files. Although Evo would benefit a lot from high resoltuion measurements, it would be a major mess to standardize all sorts of manual REW measurements in a way they could all be correctly processed and turned into an Audyssey calibration file.

One way to get standardized measurements could be to generate measurements from a script via REW API.
As a first step all required speakers could be selected. After that, measurements for each speaker could be generated with a break to move mic to a different position. Just like Audyssey is doing it.
Unfortunately no height speakers could be measured like that.

 

[OCA]

v2.2 was released but I couldn't find a download link yet. I'm guessing it's because the script can now be run online:

A1 Evo - OCA 2024

This is not the latest version though. I haven't found the opportunity to update it yet

The latest version is always in the YT video description link.

 

[PHD]

This is not the latest version though. I haven't found the opportunity to update it yet

The latest version is always in the YT video description link.

Isn't this the latest?

@ocaudiophile - Google Drive

drive.google.com

By the way, I'd suggest that you make the Evo fully online, including running your own REW API on your server. This way users could just upload the *.ady files, select a target curve and some settings (XO, filter range, etc) using a drop-down menu, and get a calibrated ady file they can download.


Why would you do that? Because this way you could run Google ads and get some revenue, because you deserve to get compensated for your hard work.

 

[OCA]

I have updated the web page with the latest version:

A1 Evo - OCA 2024

v2.3, should be written near the logo also in the begining of the logs.

 

[ooheadsoo]

2.3 is giving me drastically different results from 1.3 with the surround levels much lower than they were previously set. This makes A/B comparison difficult to do, as the fronts are about the same volume but the surrounds are 6-9+db lower than they used to be. The speaker crossovers also have generally gone up almost an octave. Very interested to listen more when there's time.

@OCA if we have measurements for our speakers from, say, https://www.spinorama.org/, is there any way we can incorporate those eq filters into the audyssey files?

 

[PHD]

@joentell @OCA

MMM approach and a new calibration app (magic beans)

I wonder what you guys think of this calibration approach? The app is not yet available for purchase but I was told that it should become available in about a week...

www.audiosciencereview.com

Just a thought, I think it would be interesting if OCA's Evo is combined with Joentell's Magic Beans. I mean use Magic Beans to create your near-filed optimized target curve and then use Evo to apply it.


@joentell what's the current status of Magic Beans? Is it close to being released to the public?

Combining the two could bring Audyssey to a near Dirac Live level...

 

[Hyacin]

I have a Umik-1, Umik-2, a Behringer and I have not used them once since I got the M23R. You are aware that you will need a descent audio interface with it, too right?

Yep! Ordered a Scarlett Solo 4th gen I read should go quite well with it

 

[Hyacin]

One way to get standardized measurements could be to generate measurements from a script via REW API.
As a first step all required speakers could be selected. After that, measurements for each speaker could be generated with a break to move mic to a different position. Just like Audyssey is doing it.
Unfortunately no height speakers could be measured like that.

You can measure in 'playback mode' via the API, so you could still do the height speakers ...

The real sticky part would be that triggering measurements via API at all require a Pro upgrade.

I'm so lazy and sick of the repetition of set name, start measurement, play file, rinse/repeat 13 times that I actually did buy Pro to automate this (and support REW of course), but my code will be useless to anyone who doesn't also have a Pro key.

And I'm doing it with the aim of doing a full Supreme cal - it doesn't help with the A1 integration you were discussing.

 

[OCA]

You can measure in 'playback mode' via the API, so you could still do the height speakers ...

The real sticky part would be that triggering measurements via API at all require a Pro upgrade.

I'm so lazy and sick of the repetition of set name, start measurement, play file, rinse/repeat 13 times that I actually did buy Pro to automate this (and support REW of course), but my code will be useless to anyone who doesn't also have a Pro key.

And I'm doing it with the aim of doing a full Supreme cal - it doesn't help with the A1 integration you were discussing.

You don't need to use API to measure Atmos channels. This old video shows how:

but use the "lossless" Atmos sweeps I created later from this video:

 

[OCA]

Yep! Ordered a Scarlett Solo 4th gen I read should go quite well with it

So long as it has 2 inputs and ideally digital loopback function, you are uncompromised for all sorts of use cases.

 

[Hyacin]

You don't need to use API to measure Atmos channels. This old video shows how:

He was talking about doing it all automatically like how the original Audyssey measurements are done, but saying that you then couldn't do highs if you were doing it from REW.

 

[Hyacin]

and ideally digital loopback function

Oooohhhh, I thought this was standard. I'd better do some digging before it ships to make sure it has it then!! ty

 

[joentell]

@joentell @OCA

MMM approach and a new calibration app (magic beans)

I wonder what you guys think of this calibration approach? The app is not yet available for purchase but I was told that it should become available in about a week...

www.audiosciencereview.com

Just a thought, I think it would be interesting if OCA's Evo is combined with Joentell's Magic Beans. I mean use Magic Beans to create your near-filed optimized target curve and then use Evo to apply it.


@joentell what's the current status of Magic Beans? Is it close to being released to the public?

Combining the two could bring Audyssey to a near Dirac Live level...

It's been out for months

 

[OCA]

2.3 is giving me drastically different results from 1.3 with the surround levels much lower than they were previously set. This makes A/B comparison difficult to do, as the fronts are about the same volume but the surrounds are 6-9+db lower than they used to be. The speaker crossovers also have generally gone up almost an octave. Very interested to listen more when there's time.

@OCA if we have measurements for our speakers from, say, https://www.spinorama.org/, is there any way we can incorporate those eq filters into the audyssey files?

You can export any REW filter into an Audyssey ady file by using the old scripts in Audyssey ART video. It's not a fully automatic process but wouldn't take long either.