Emotiva RMC-1 AV Processor Review

[jhaider]

Of course, without a robot arm their with be variations by slightly different microphone positions. With enough, smoothing though, it will be great

- Rich

I have no idea what you're trying to express.

What does a "robot arm" have to do with the relatively simple task of sampling the response at several points?

Obviously single point measurements are not appropriate to measure in room response. That's why we take spatial averages. It's called the statistical region for a reason.

 

[Francis Vaughan]

The trouble with spatial averages is that they can be just as misleading as a point measurement. Just in a different way. There is no simple metric that tells you all you need to know. If you are testing the precise effects of parameter changes you need to have exact reproduction of the sampling locations. Minimally, resetting parameters to an earlier set of values must yield the same measurements. Otherwise you are never going to understand what is going on.

 

[RichB]

I have no idea what you're trying to express.

What does a "robot arm" have to do with the relatively simple task of sampling the response at several points?

Obviously single point measurements are not appropriate to measure in room response. That's why we take spatial averages. It's called the statistical region for a reason.

I am trying to say what Francis said (much better than i) :

The trouble with spatial averages is that they can be just as misleading as a point measurement. Just in a different way. There is no simple metric that tells you all you need to know. If you are testing the precise effects of parameter changes you need to have exact reproduction of the sampling locations. Minimally, resetting parameters to an earlier set of values must yield the same measurements. Otherwise you are never going to understand what is going on.

- Rich

 

[jhaider]

The trouble with spatial averages is that they can be just as misleading as a point measurement. Just in a different way. There is no simple metric that tells you all you need to know. If you are testing the precise effects of parameter changes you need to have exact reproduction of the sampling locations. Minimally, resetting parameters to an earlier set of values must yield the same measurements. Otherwise you are never going to understand what is going on.

If you need an "exact reproduction of the sampling locations," then what you're doing is irrelevant to actual listening by human beings with two ears and a brain.

 

[RayDunzl]

Re: Measurements before and after Digital Room Correction

Someone was asking about distortion (much) earlier and received different answers.


Experiment:

Setup:

PC REW -> USB -> Topping D10 -> Optical -> Audio authority 1177A switch -> Coaxial -> Focusrite Clarett 4Pre USB -> Focusrite Internal Matrix -> Focusrite Monitor Output 1 -> Focusrite Internal Matrix -> USB -> PC REW

First stream passes through all the gear above, but no filters are applied.
Second stream has some old AcourateDRC filter applied at the miniDSP OpenDRC-DI, a single point measurement was used to create the filter, for MartinLogan reQuest speakers in this room, with an "aggressive" full range filter. Both FIR and IIR filters are part of the correction.

Measurement is of the left channel digital stream as received by REW.

-6db is applied to both streams at the miniDSP OpenDRC-DI.


Conclusion:

To say there is "no distortion" added to the digital signal with digital room correction is false in this case

However, I would consider the increase in distortion at the digital signal level as measured here to inconsequential.



Frequency Response:

Green - no "correction"
Violet - "corrected"

Phase adjustments

Distortion as THD

Harmonics:

No correction:

(noise floor is the upper brown trace)

With correction:

 

[RayDunzl]

I do notice this in the digital output:

Without "correction"

With correction filter in the path, a rise in the low frequency noise, and a little in the high frequencies, when a signal is applied.

Again, rather inconsequential.

 

[Francis Vaughan]

If you need an "exact reproduction of the sampling locations," then what you're doing is irrelevant to actual listening by human beings with two ears and a brain.

That is still missing the point. If you are using a black box to correct your room this is true. If you are trying to understand how the correction is working or design corrections yourself you must have reproducibility. We know that small changes in position can and do cause significant changes in results. If you trust an automated system to do the work you assume it does the best job it can, each time you assume the variation in results are simply within the bounds of correct operation. But if you wish to delve deeper you cannot have random variations in results from run to run. You will be searching blindly for understanding.

 

[jhaider]

If you are trying to understand how the correction is working or design corrections yourself you must have reproducibility.

Again, if the equalization system requires exacting duplication of microphone placement to measure what it's doing, then it is an inherently unstable system and simply not worth our time discussing.

Have you ever attempted to measure what a room correction system is doing using valid spatial sampling methods? I have done so, multiple times. You can see some of the results in my AVR, SSP, and integrated amp reviews below. I use the method outlined in the Geddes and Blind paper; the "moving mic method" seems reasonable as well.

My conclusion based on real world practice is, this exact microphone placement stuff is pointless and stupid.

 

[RayDunzl]

With correction filter in the path, a rise in the low frequency noise, and a little in the high frequencies, when a signal is applied.

I suspect that "noise", whatever it comes from, is being interpreted as harmonic levels in the THD measures.

 

[QMuse]

That’s not necessary. If the system is any good a spatial average over the same general area will be very stable.

I agree. I usually do spatial averaged measurement with moving microphone method with REW RTA mode as it takes much less time than to do multiple sweeps and indeed the results I'm getting practically don't differ.

 

[QMuse]

That is still missing the point. If you are using a black box to correct your room this is true. If you are trying to understand how the correction is working or design corrections yourself you must have reproducibility.

I'm not using black box to build my filters and I never had issues with reproducibilty.

Let's start with this: my listening sofa is at the fixed playce in the room where I listem my music. Let's agree to define the LP as a 3D space which covers the 2D footprint area of the sofa and for the height dimension it spans from 50 to 100cm above the floor. I take spatial measurement of uncorrected response within that area using 60+ RTA samples with moving microphone method to cover the entire LP. If I repeat that measurement tomorrow I will get the same result. If I repeat that measurement using multiple sweeps taken at various point at LP and average them I will get the same result, although it will take much much longer to take seuch measurement. Then I build correction filters mannually suing that uncorrected measurement, load the filters to convolution engine and repeat the measurement. I get corrected response that looks very similar to the target curve I was aiming for when designing filters. If I repat that measurement tommorrow I get the same result. So yes, I'd say I do have reproducibility.

 

[QMuse]

That hasn't been my experience.

Maybe there is a misunderstanding here. What exactly do you mean when you say spatial average measurement is not stable in your experience?

 

[QMuse]

You haven’t looked that hard!

Ah, my bad - your detailed room EQ reviews at Home Theater HiFi somehow skipped my search.

 

[Dimifoot]

If you want some illustrative room correction measurements follow the link in my signature.

I read your reviews of the Bryston and the 7702 (Dirac-Audyssey).

This is what I head in mind when I suggested that RoomEq and bass management implementation should be included in processor reviews, since these are very important parameters of how such a device sounds in the real world.

Of course there is room for improvement, but you have done a great job!

 

[Costas EAR]

To say there is "no distortion" added to the digital signal with digital room correction is false in this case

Thank you for sharing, i don't think I have ever seen measurements like these.

Your findings are truly important.

I wonder if peq or dirac on a processor like emotiva have similar findings in distortion measurements.

 

[Costas EAR]

. I use the method outlined in the Geddes and Blind paper; the "moving mic method" seems reasonable as well

Yes it is. It is also the quickest method for average measurement of the listening position, without the need of multiple microphone positions or any accuracy in microphone placement.

Your reviews are awesome.

 

[RayDunzl]

I wonder if peq or dirac on a processor like emotiva have similar findings in distortion measurements.

That question should be posed to a mathematician who understands the manipulative algorithms involved.

Unfortunately, that is above my pay scale.

I think there are several members of this site that are qualified.

Start a thread...

The creator of rePhase is here, @pos , I think, among others.

 

[QMuse]

I do notice this in the digital output:

Without "correction"

View attachment 52469



With correction filter in the path, a rise in the low frequency noise, and a little in the high frequencies, when a signal is applied.

View attachment 52471

Again, rather inconsequential.

And potentially only specific to your platform.

 

[markus]

We have wondered way way of topic , there's threads for talking about room correction.

Your all free to start new ones of course.

Anything other than emotiva and these measurements and I'd ask you to take it to another thread.

Cheers .

I've created a new thread "Testing room correction systems".

 

[jhaider]

I read your reviews of the Bryston and the 7702 (Dirac-Audyssey).

This is what I head in mind when I suggested that RoomEq and bass management implementation should be included in processor reviews, since these are very important parameters of how such a device sounds in the real world.

Of course there is room for improvement, but you have done a great job!

Thank you. I really appreciate it. My point about it taking forever to generate such measurements is in a case like Amir's, where there is rapid turnover. Teardown and setup, followed by and teardown and re-setup after measurement, is a huge time sink. When you have an AVR or SSP in your system for 3 months (customary for an in depth review), it's more a question of competence and will than anything else.