• 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!

Room Correction Software Measurements

Jeromeof

Addicted to Fun and Learning
Forum Donor
Joined
Jul 25, 2021
Messages
618
Likes
967
Location
Ireland
I hope most people agree that room correction is probably the most important aspect of getting the best sound in your listening space. And please correct me if I missed some older threads with details but I would love to know definitively / objectively if, for example, Dirac is better than Audyssey and Audyssey is better than YPAO etc. And specifically the criteria to judge why one better than the other.

I am probably going to invest some money this year is 'fixing' my converted garage, maybe some acoustic treatments, maybe minidsp with Dirac, maybe some 'home grown' configuration with CamillaDSP on a RaspberryPI, or maybe some other solution but before I do I would like to understand how to evaluate what is best. It just seems very subjective, once I bring this topic someone will quickly tell me Dirac is the best or that I need some specific room treatment all without IMO any satisfactory evidence other than the "trust me I know what I am talking about". And I mostly understand the science behind accoustic treatments and DSP / FIR etc based 'software' corrections.

Anyway, if there are studies or measurements comparisons between the different available options I would love to read about them.
 
OP
Jeromeof

Jeromeof

Addicted to Fun and Learning
Forum Donor
Joined
Jul 25, 2021
Messages
618
Likes
967
Location
Ireland
An additional factor to consider (aside from the software) is the technical capability of the EQ hardware. In the documentation for REW there is an attempt to tabulate and compare the capabilities of various compatible hardware eq devices...

https://www.roomeqwizard.com/help/help_en-GB/html/equaliser.html#top
Thanks I had seen that list numerous times in REW but mostly assumed the were older DSP hardware as “generic” is almost always chosen in tutorials for generating some eq settings
 

-Matt-

Addicted to Fun and Learning
Joined
Nov 21, 2021
Messages
675
Likes
551
Yes, some of the hardware listed is older. Perhaps it would be good if ASR could host a wiki type table comparing the capabilities of various (more recent) EQ offerings? Could also add columns for phase eq capabilities too, for example.

I'd guess generic is used in tutorials as it doesn't face any limitations imposed by specific hardware and can be used as a test case even if no hardware is present.
 

-Matt-

Addicted to Fun and Learning
Joined
Nov 21, 2021
Messages
675
Likes
551
On the measurement of EQ performance, it is an interesting idea...

I wonder if something like the following would be possible...

Can an audio analyser be configured to output a frequency sweep where the amplitude varies in a precisely controlled and repeatable manner? If so, this could be used as a test signal. EQ software and hardware could then be quantifiably tested by its ability to return the signal to a perfectly linear one. (The magnitude of remaining errors being inversely proportional to the score). Test signals with differing magnitude of errors and complexity could be defined. (Of course it would be best to make a standardised set of test signals).
 
Last edited:

-Matt-

Addicted to Fun and Learning
Joined
Nov 21, 2021
Messages
675
Likes
551
To be clear, in the above I'm proposing all electrical testing (no speakers or microphones involved). Of course that means you'd always need a way to import the test signal into the eq software, as though it were a microphone measurement. (Which might be difficult or impossible for some fully automated room eq devices such as AVRs for example. Perhaps in those cases a suitably attenuated test signal sent direct from the audio analyser into the microphone port would work).

Of course, this also depends upon someone with the necessary hardware thinking that such a test could be feasible and worthwhile.
 
Last edited:

-Matt-

Addicted to Fun and Learning
Joined
Nov 21, 2021
Messages
675
Likes
551
Clarification of the suggested method:

To allow the EQ device to generate filters for the given test signal (if direct digital input of the test signal as a microphone measurement is not possible)...

Analyser (test signal out) -> Attenuation if needed -> EQ system measurement input -> Auto generate EQ filters


Using those filters...

Analyser (test signal out) -> EQ input -> EQ filters -> EQ output -> Analyser input (measure residual errors)


Example results...
Test 1 (easy) - Mean of residual errors 0.5dB.
Test 2 (medium) - Mean of residual errors 1.5dB.
Test 3 (extreme) - Mean of residual errors 6dB.
(Note, poor score on test 3 because of insufficient range of correction. For errors up to 12dB in magnitude the eq performed well, although it struggled with the narrow features near 2kHz in test 3).
 
Last edited:

-Matt-

Addicted to Fun and Learning
Joined
Nov 21, 2021
Messages
675
Likes
551
I guess we both missed this thread where something similar to what I suggested above is done (with the filters for LXmini speakers used as the test signal). However, in that case the focus seems to be on how the noise floor (for a music signal) is increased when filters are applied (tested via multitone and single frequency FFT measurements).

The focus of the testing suggested above would be to see how accurately the filters (and the filter generation algorithms) can correct various distortions to the input signal frequency response (and maybe phase response too). Still a potentially worthwile testing program I'd think.
 
Last edited:
OP
Jeromeof

Jeromeof

Addicted to Fun and Learning
Forum Donor
Joined
Jul 25, 2021
Messages
618
Likes
967
Location
Ireland
I guess we both missed this thread where something similar to what I suggested above is done (with the filters for LXmini speakers used as the test signal). However, in that case the focus seems to be on how the noise floor (for a music signal) is increased when filters are applied (tested via multitone and single frequency FFT measurements).

The focus of the testing suggested above would be to see how accurately the filters (and the filter generation algorithms) can correct various distortions to the input signal frequency response (and maybe phase response too). Still a potentially worthwile testing program I'd think.
I had seen that thread - it is very interesting and have setup CamillaDSP on a RPI (and I even bought a separate toslink to USB interface so I can have a mostly full digital path in and out of the RPI for DSP (and I feel that will minimise the possiblity of noise).

Your proposal is very interesting and I would love to see Dirac and other commercial systems compared in this way so that we could know if its worth going down the commercial "slick room correction" route or do everything by hand or something in-between.

This guy on YouTube (with probably the most accurate name I can imagine if you watch his videos - he is definitely obsessively compulsive ) has some amazing details tutorials about basically hacking Audyssey with very fine tuned DSP generated by REW.

My question is mostly its hard to compare the benefits of the various approaches (though CamilaDSP obviously has advantages as far as noise floor)
 

Speedskater

Major Contributor
Joined
Mar 5, 2016
Messages
1,610
Likes
1,333
Location
Cleveland, Ohio USA
I hope most people agree that room correction is probably the most important aspect of getting the best sound in your listening space.
Why would you say that?
Over the years, many experts have written about the challenges of whole room correction.
* * * * * * *
OK, the newest correction systems getting are a lot smarter.
 

sarumbear

Master Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
7,604
Likes
7,314
Location
UK
I hope most people agree that room correction is probably the most important aspect of getting the best sound in your listening space.
It is the easiest and cheapest way for sure but certainly not the best way. Treating your listening place for good acoustic properties cannot be beaten.
 
OP
Jeromeof

Jeromeof

Addicted to Fun and Learning
Forum Donor
Joined
Jul 25, 2021
Messages
618
Likes
967
Location
Ireland
It is the easiest and cheapest way for sure but certainly not the best way. Treating your listening place for good acoustic properties cannot be beaten.
Yes I know - but my wife has a phobia about strange "patterns" on walls.:):)

I actually have a few 'listening' areas around the house and mostly the rooms with lots of "stuff" sound great. The problem is my new shiny modern bright "minimalist" working from home area doesn't have stuff (and my wife really wants to keep it "clean") hence why I am looking at room correction software.

But I imagine lots of space are like this and lots of spaces cannot be treated for various reasons - so the next best option has to be software and that is really my question - what is objectively (not subjectively) the best room correction software / hardware solution.

E.g. I am probably going to get a MiniDSP in the next few weeks (to compare with my RPI / CamillaDSP solution) but then was wondering do I bother with Dirac (or can I do the equivalent with REW). It just seems like this area is lacking in the detail we get about DAC / Amps / Speakers etc.I.e. some metrics to allow comparisons between solutions.
 

Tom C

Major Contributor
Joined
Jun 16, 2019
Messages
1,501
Likes
1,370
Location
Wisconsin, USA
Not that it’s important in any way, but just thought I’d mention that no, I don’t believe room correction is the most important aspect of getting the sound right in your listening space. I would say that having well designed speakers is the most important aspect. Start with that, and the need for room correction should be reduced.
 

Daverz

Major Contributor
Joined
Mar 17, 2019
Messages
1,294
Likes
1,451
The problem is my new shiny modern bright "minimalist" working from home area doesn't have stuff

I don't think correction with DSP can do much for specular reflections from bare walls. I would say that you only need to treat the reflection points on side walls and the wall behind the listening position. You don't need to turn the room into a recording studio. That said, perhaps a system in the minimally furnished room will sound fine.
 

Hapo

Senior Member
Joined
Jan 11, 2022
Messages
464
Likes
320
R.b265956c610391671943a407e10ac822
 
Top Bottom