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

[middlemarch]

I wanted to close the loop as to my initial question. Now running version 1.4 of the Evo script and John M's version 5.40 Beta 33 of REW.

I can now report that the script runs almost flawlessly natively on my MacBook pro. No more issue with not saving the measurements zip file in Safari (always worked OK in Chrome) and the script does not hang anywhere. Instead it runs all the way through with no errors or warnings and best of all, no sub inversion! (a pain when using 4 subs via a miniDSP 2x4HD)

I say almost flawlessly because at the end, doing a right click in the log window doesn't generate the option to save that to a file. But I can just copy the log contents and save it to Notepad so no real biggy. Script run time is around 145 seconds running natively versus 530 seconds running emulated in Windows 11.

I'm now off to test three versions for my 4 sub system and one for my two sub stereo system.

Thank you Serko and John Mulcahey! Go off and have a beer or two, you deserve it

 

[middlemarch]

Hi

Does the 255Hz limit in v1.4 only apply to the A1 script optimization or does the script also defeat the original Audyssey optimization beyond 255Hz?

Can I use A1 for 20Hz to 255Hz optimization and then use MultiEQ app for the remaining freq gange? I do notice an improvement up to 500Hz with the Audyssey app.

Thanks

I don't think the Audyssey app can be used to modify the optimized .ady file (at least it won't for me). It's trivial to go into the Evo script however and change the 225 Hz to whatever you want. The constant is at line 225 in the script. I've changed one of my runs to 499Hz just as a test.

 

[PHD]

I don't think the Audyssey app can be used to modify the optimized .ady file (at least it won't for me). It's trivial to go into the Evo script however and change the 225 Hz to whatever you want. The constant is at line 225 in the script. I've changed one of my runs to 499Hz just as a test.

So the A1 is only good for lower-end optimization. I wish it could have preserved the original Aydyssey correction for frequencies above the cutoff. I'm not sure why OCA recommends not going above 255Hz unless his script tries to mimic Dirac ARC which won't work for midbass and above.

When I use the MultiEQ app, I notice better results when not limiting the curtain range below 500Hz. I usually leave it at 1kHz for best results.

 

[Sal1950]

When I use the MultiEQ app, I notice better results when not limiting the curtain range below 500Hz. I usually leave it at 1kHz for best results.

Opinions are great but not as good as a meter and some measurements.
How well Audyssey works above 500 or so is very dependent on the speakers, room, and ears.
So many claim things should be cut off above the Schroeder frequency, while others recommend a bit of a FR tweak above for this or that speaker?
Last time in, I did one of each and personally found the one doing some minor tweaks above, the one I liked most.
YMMV

 

[EasyC]

Version 1.6 online now at OCA's YT channel. It's an evolving thing...

 

[middlemarch]

I've been following along in real time. Every time I get a set of test files ready for a listening session a new version arrives and I do it all over again. Still getting some weird results with 1.6, but I did get some quality listening time on one version yesterday and sounded pretty good to me. Biggest problem is the constant back and forth on polarity inversion, or not, of my subs. Since all four are behind a minidsp, I have no way to invert the whole bunch without building a dedicated RCA cable, so I have to invert all four individually. I'm not sure that results in the same thing. Wish there was an invert button on the Input side of the minidsp, but oh well...

On the bright side, I am becoming VERY familiar with running the tools and am learning yet a bit more about REW, so there's that.

I'm betting on version 2.0 being the cat's meouw. Fingers crossed

 

[PHD]

I just cannot get it to work with my X4400H + 3 subs connected via MiniDSP 2X4HD and precalibrated using MSO. The script (v1.6) still gives a warning regarding max speaker-to-speaker distance (6 meters) being exceeded (sub distance is set to +6m with respect to the closest speaker distance)

I give up (I may try again with v2.1, or with the X4800H, I'm planning on getting, whichever comes first...)

 

[PHD]

I've been following along in real time. Every time I get a set of test files ready for a listening session a new version arrives and I do it all over again. Still getting some weird results with 1.6, but I did get some quality listening time on one version yesterday and sounded pretty good to me. Biggest problem is the constant back and forth on polarity inversion, or not, of my subs. Since all four are behind a minidsp, I have no way to invert the whole bunch without building a dedicated RCA cable, so I have to invert all four individually. I'm not sure that results in the same thing. Wish there was an invert button on the Input side of the minidsp, but oh well...

On the bright side, I am becoming VERY familiar with running the tools and am learning yet a bit more about REW, so there's that.

I'm betting on version 2.0 being the cat's meouw. Fingers crossed

How can use the script to detect subs connected to a MiniDSP? Do you have 2 subs with dual sub setup in the AVR? I have three subs introduced to the AVR as a single sub, precalibrated using MSO. The script like the AVR thinks I have a single sub but it tries to set its distance too far exceeding the 6m limit...

This setup works great w/o the EVO script (running the MulEQ app which sets the sub distance at about 5 meters further from the closest speaker...)

 

[middlemarch]

The script cannot detect individual subs. I have them all run mono as a single sub. Time aligned and EQ'd via REW. So the script sees them as a single sub. Same deal as yours. The reason the delays are greater than the real distance is because the miniDSP takes some amount of time to do it's processing. In addition we had to set delays relative to each of the subs in the miniDSP in order to time align them. All this delay adds up and translates to distance when measuring the bass. That all makes sense. Mine also runs great with the standard MulEQ app.

Serko is trying to do something quite a bit more sophisticated with his scripts. I think this adds more variability and results in sometimes weird results. But I think he's getting closer, it's still a work in progress, why my bet is on version 2.0. Be easier and quicker if he was able to test all the system variations himself, but we have to do that for him, so it's a big loop

 

[Hyacin]

I was just reviewing the Dolby whitepapers and guides as I'm updating my seating placement and moving some speakers to accommodate some wall treatments, and the following jumped out at me which is contrary to what you say in your Supreme Audyssey video @OCA -

Calibrate the subwoofer level to give the same level for redirected bass content from full-range speakers as those speakers produce in their pass bands.

(From https://www.dolby.com/siteassets/te...atmos-installation-guidelines-121318_r3.1.pdf, emphasis mine)

As I do notice wildly different levels between the full range test tones you provided and the built-in test tones which I imagine don't engage the sub at all, I guess I'll try calibrating for just the 'pass bands' this time. Amazingly though, even with such large differences in level between the two methods, I have a difficult time telling the difference after adjusting from one to the other. :shrug:

Figured it was worth pointing out though!

 

[OCA]

I was just reviewing the Dolby whitepapers and guides as I'm updating my seating placement and moving some speakers to accommodate some wall treatments, and the following jumped out at me which is contrary to what you say in your Supreme Audyssey video @OCA -



(From https://www.dolby.com/siteassets/te...atmos-installation-guidelines-121318_r3.1.pdf, emphasis mine)

As I do notice wildly different levels between the full range test tones you provided and the built-in test tones which I imagine don't engage the sub at all, I guess I'll try calibrating for just the 'pass bands' this time. Amazingly though, even with such large differences in level between the two methods, I have a difficult time telling the difference after adjusting from one to the other. :shrug:

Figured it was worth pointing out though!

Did you compensate for the 10dB LFE boost applied by the receiver for Dolby? All sweeps are extracted from a single sweep by Atmos encoder. I don't have control over individual channel volumes.

 

[Hyacin]

Did you compensate for the 10dB LFE boost applied by the receiver for Dolby? All sweeps are extracted from a single sweep by Atmos encoder. I don't have control over individual channel volumes.

I did yes!

I meant the difference between 'full range' LFE + channel SPL readings, vs. just from the channel. And I don't mean between different sample sources - just that the channel-to-channel levels vary widely between the two methods. FWL may be +3 dB higher than FWR for example in one of the two methods compared to the other. So it was interesting to see that Dolby recommends level matching based only on the pass bands vs. the full range.

I'll still use your full range video for sure, I'll just disable the subs via my MiniDSP next time I think

 

[OCA]

Did you try A1 Evo yet? Although it's based on Audyssey's low resolution measurements, the methods and techniques in it are light years ahead of supreme Audyssey and you can easily implement as it's open source and simple java script.

 

[Hyacin]

Did you try A1 Evo yet? Although it's based on Audyssey's low resolution measurements, the methods and techniques in it are light years ahead of supreme Audyssey and you can easily implement as it's open source and simple java script.

I'm generally heading in quite the opposite direction, having ordered an M23R just a couple days ago!

Having just installed a lot of room treatment I'm due for a re-cal, but with it arriving hopefully next week, I'm reluctant to do a big cal with my UMIK-1 - so I may actually try A1 Evo tonight!! It'll give me a quick and dirty at least with the treatments that should sound tons better than my old cal, and then I can hold off until I get my M23R before doing a big Supreme cal!

 

[PHD]

Did you try A1 Evo yet? Although it's based on Audyssey's low resolution measurements, the methods and techniques in it are light years ahead of supreme Audyssey and you can easily implement as it's open source and simple java script.

Thank you for your excellent work. I don't know if you have been asked this before, but I wonder why can't Umik-1/2 measurements be taken using REW for the A1 script to do its thing? Wouldn't such an approach be equivalent to manual calibration, only automated by the A script?

Thanks

 

[Hyacin]

Thank you for your excellent work. I don't know if you have been asked this before, but I wonder why can't Umik-1/2 measurements be taken using REW for the A1 script to do its thing? Wouldn't such an approach be equivalent to manual calibration, only automated by the A script?

Thanks

Yeah that would be super cool.

 

[Hyacin]

Having just installed a lot of room treatment I'm due for a re-cal, but with it arriving hopefully next week, I'm reluctant to do a big cal with my UMIK-1 - so I may actually try A1 Evo tonight!! It'll give me a quick and dirty at least with the treatments that should sound tons better than my old cal, and then I can hold off until I get my M23R before doing a big Supreme cal!

Well, I tried it ... well, I tried a basic 3pt cal to start and wanted to just throw that on the receiver to quickly see how it sounded. Kept getting this "Could not transfer all data to the AV receiver" error.

Thought I'd try running it through A1Evo to see if that sorted out - kept getting the same error :-/

So I opened the file up in the jsoneditor and it's showing one too few channels ... just an SW2, no SW1 ?!

I don't _think_ this is my first Audyssey run since getting the MiniDSP, but it is certainly my first Audyssey run since connecting my bass shakers to the MiniDSP and routing both input 1 and input 2 to them - though they were off during my measurement run.

Sooooo I thought maybe it was just a one off and did another set of measurements ... and again, it's only producing a file with 14 detected channels instead of my usual 15 :-/ :-/

I'm guessing something I've done with the MiniDSP config is screwing something up, but it's getting too late here to be running loud sweeps, so I'll have to try to figure it out tomorrow :'(

I currently have my two sub outs from my 6700H going to the two inputs on the MiniDSP, then sub 1 taking input 1, sub 2 taking input 2, and the shakers (bridged) taking both input 1 and input 2.

I was thinking while I was testing though, maybe it would be a good idea to route the sub1 in signal to both of the actual subs, and dedicate the sub2 in signal to the shakers ... though I'm not sure if that would be of any specific use or if Audyssey would even be able to pick up their vibrations or attempt to time align them or anything useful at all.

More to unravel and figure out tomorrow! Watching A1Evo run though was fun, it looks like it dives in a lot deeper than the previous version did, especially on crossovers and such!

 

[OCA]

Thank you for your excellent work. I don't know if you have been asked this before, but I wonder why can't Umik-1/2 measurements be taken using REW for the A1 script to do its thing? Wouldn't such an approach be equivalent to manual calibration, only automated by the A script?

Thanks

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.

 

[OCA]

having ordered an M23R just a couple days ago

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.

 

[OCA]

Well, I tried it ... well, I tried a basic 3pt cal to start and wanted to just throw that on the receiver to quickly see how it sounded. Kept getting this "Could not transfer all data to the AV receiver" error.

Thought I'd try running it through A1Evo to see if that sorted out - kept getting the same error :-/

So I opened the file up in the jsoneditor and it's showing one too few channels ... just an SW2, no SW1 ?!

I don't _think_ this is my first Audyssey run since getting the MiniDSP, but it is certainly my first Audyssey run since connecting my bass shakers to the MiniDSP and routing both input 1 and input 2 to them - though they were off during my measurement run.

Sooooo I thought maybe it was just a one off and did another set of measurements ... and again, it's only producing a file with 14 detected channels instead of my usual 15 :-/ :-/

I'm guessing something I've done with the MiniDSP config is screwing something up, but it's getting too late here to be running loud sweeps, so I'll have to try to figure it out tomorrow :'(

I currently have my two sub outs from my 6700H going to the two inputs on the MiniDSP, then sub 1 taking input 1, sub 2 taking input 2, and the shakers (bridged) taking both input 1 and input 2.

I was thinking while I was testing though, maybe it would be a good idea to route the sub1 in signal to both of the actual subs, and dedicate the sub2 in signal to the shakers ... though I'm not sure if that would be of any specific use or if Audyssey would even be able to pick up their vibrations or attempt to time align them or anything useful at all.

More to unravel and figure out tomorrow! Watching A1Evo run though was fun, it looks like it dives in a lot deeper than the previous version did, especially on crossovers and such!

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.