Room correction - speakers moving

[MCH]

Hello all,
My problem is a bit silly but i guess i am not the only one...
After reading the forum for a while, my next project was to borrow an umik from ebay and perform room coŕrection in my livingroom.
I have floorstanding speakers and tiles floor.
Living with partner, kid, with friends of kid coming often, cleaning robots, cleaning humans... after doing some observations i think i need to live with the fact that from time to time the speakers are going to be moved, slightly or not so slightly...
And there are three things that I really want to avoid:
- having to repeat the measurements and correction every second week
- having to fix the speakers to the floor
- and most specially: being the grumpy guy always telling "don't touch my speakers!!"

So my question for you guys, before i decide if it is worth wasting my time on this, is, how sensitive are the room correction adjustments to small displacements of the speakers?
Would be enough to do some marks on the floor (i.e. where the four corners of the speakers are) so whenever they are moved i can just move them again to their correct place or are 2 mmm enough to make a significant change?
Thank you for your advices.

PS: i can live a happy life without room correction, just thought it could be worth if it is short work and long enjoyment, not the other way around

 

[Chrispy]

Marking the floor somehow sounds like a good plan to put back to your ideal spot if they get moved (or might help others keep them where you want?). I wouldn't worry about it too much in any case. How are you implementing room correction particularly after measurement with the Umik?

 

[MCH]

Marking the floor somehow sounds like a good plan to put back to your ideal spot if they get moved (or might help others keep them where you want?). I wouldn't worry about it too much in any case. How are you implementing room correction particularly after measurement with the Umik?

Thank you, that is what i wanted to hear
Well i still need to do some more research (as is evident) but my idea was to follow one of those tutorials using rew or similar and the applying the eq in moode/camilla

 

[Chrispy]

Thank you, that is what i wanted to hear
Well i still need to do some more research (as is evident) but my idea was to follow one of those tutorials using rew or similar and the applying the eq in moode/camilla

moode/camilia?

 

[MCH]

Yes, moodeaudio, that has camilla dsp implemented

 

[Chrispy]

Yes, moodeaudio, that has camilla dsp implemented

Thanks, searching moodeaudio got me further than moode/camilia did. Hadn't heard of this one before.

 

[MCH]

so, just to give it a try and before investing any significant money, the idea is to connect a raspberry pi with moodeaudio through USB to my DAC (already tested it and works perfect) and then implement the eq on camilla dsp that is already integrated in moodeaudio.
then i will just stream the music to the pi instead of my current streamer (that does not have eq) and that's it. Still need to learn how to "translate" the eq that result from the measurements/rew to camilla, but thats a different story...

(all this said by a complete newbie, take it with caution)

 

[Hipper]

I agree that small markers on your floor would be a good idea. If the speakers are slipping a bit some sort of non-slip pads may help.

I also find that my listening chair can move from me getting in and out.

If you consider that when you are listening you will be moving your head a bit, forward, backwards, side ways and also at different angles - head bent down or up - you can see that measuring by the millimetre isn't so important. Try moving your head and hear what a difference it can make, especially dipping your head forwards.

Moving the speakers - or measuring mic - could change the measurements by a small amount too. Trying to get repeat measurements even when the speakers are in the same position is not so easy either. Therefore as long as it still sounds good to you, and the speakers haven't moved a large amount, you should be OK.

 

[MCH]

I agree that small markers on your floor would be a good idea. If the speakers are slipping a bit some sort of non-slip pads may help.

I also find that my listening chair can move from me getting in and out.

If you consider that when you are listening you will be moving your head a bit, forward, backwards, side ways and also at different angles - head bent down or up - you can see that measuring by the millimetre isn't so important. Try moving your head and hear what a difference it can make, especially dipping your head forwards.

Moving the speakers - or measuring mic - could change the measurements by a small amount too. Trying to get repeat measurements even when the speakers are in the same position is not so easy either. Therefore as long as it still sounds good to you, and the speakers haven't moved a large amount, you should be OK.

was concerned that speakers pointing to slight different directions would result in different reflections messing everything else significantly, but understand from you guys this is not so relevant if the displacement is small, thanks!

 

[GalZohar]

Also I noticed that at least for bass frequencies (which I limit my correction to with audyssey app), small position/orientation changes are insignificant. Not sure about higher frequencies - Those might be more sensitive.

 

[Alice of Old Vincennes]

I agree that small markers on your floor would be a good idea. If the speakers are slipping a bit some sort of non-slip pads may help.

I also find that my listening chair can move from me getting in and out.

If you consider that when you are listening you will be moving your head a bit, forward, backwards, side ways and also at different angles - head bent down or up - you can see that measuring by the millimetre isn't so important. Try moving your head and hear what a difference it can make, especially dipping your head forwards.

Moving the speakers - or measuring mic - could change the measurements by a small amount too. Trying to get repeat measurements even when the speakers are in the same position is not so easy either. Therefore as long as it still sounds good to you, and the speakers haven't moved a large amount, you should be OK.

If speakers require listening in head vice, room correction is least of problem.

 

[phoenixdogfan]

if you had something like Dirac Live, there would be even less to worry about because it optimizes the speakers in the room over a wider listening area, such as the size of a couch. I chose that couch option, and did 17 measurements with the Umik. It took around 45 minutes, and did improve the sound of my LS 50 Meta- SVS SB 2000 system.

 

[kyle_neuron]

Also I noticed that at least for bass frequencies (which I limit my correction to with audyssey app), small position/orientation changes are insignificant. Not sure about higher frequencies - Those might be more sensitive.

This phenomenon is a function of wavelength. Higher frequencies have shorter wavelengths, therefore relatively smaller movements will move between regions of summation and cancellation due to phase offset.

This is one of the arguments against making lots of changes based on a single measurement position, or even just the listening position. You might make a very pretty, flat or target curve compliant trace in magnitude and phase, but if you even rotate your head a few degrees it no longer holds up in the higher frequencies.

That’s also why a general rule of thumb is that PEQ should use much broader filter widths as you move up in frequency. Narrow filters will only fix things at one spot.

9 measurement positions, spread across the listening area and vector averaged is a good minimum to target. If you can weight the measurements, then doing more positions farther from the listening area is also useful - you can then still give priority to the place you're going to be most of the time. A simple way to weight the data is to simply apply a level offset to the 'off axis far' positions before generating the average.

 

[abdo123]

if you had something like Dirac Live, there would be even less to worry about because it optimizes the speakers in the room over a wider listening area, such as the size of a couch. I chose that couch option, and did 17 measurements with the Umik. It took around 45 minutes, and did improve the sound of my LS 50 Meta- SVS SB 2000 system.

I'd recommend you to limit the number of measurements to where your head will be and limit the correction to the transition frequency.

Also congrats on a great system!

 

[GalZohar]

This phenomenon is a function of wavelength. Higher frequencies have shorter wavelengths, therefore relatively smaller movements will move between regions of summation and cancellation due to phase offset.

This is one of the arguments against making lots of changes based on a single measurement position, or even just the listening position. You might make a very pretty, flat or target curve compliant trace in magnitude and phase, but if you even rotate your head a few degrees it no longer holds up in the higher frequencies.

That’s also why a general rule of thumb is that PEQ should use much broader filter widths as you move up in frequency. Narrow filters will only fix things at one spot.

9 measurement positions, spread across the listening area and vector averaged is a good minimum to target. If you can weight the measurements, then doing more positions farther from the listening area is also useful - you can then still give priority to the place you're going to be most of the time. A simple way to weight the data is to simply apply a level offset to the 'off axis far' positions before generating the average.

My low budget Audyssey XT is known for doing the exact opposite for the high frequencies, which is even more of a reason to disable the correction for high frequencies when using it. I know XT32 is supposed to be better (not performing narrow corrections at high frequencies), but even then it seems like opinions are mixed about using it to correct high frequencies (and most lean towards limiting it to low frequencies).

Seems like if we only correct low frequencies, then the number and location of measurement positions is a lot less critical, as long as they're near the listening position, as the bass frequencies don't seem to vary too much with small mic movements.

 

[AudioX3]

notch filtering on highs is narrow for a reason, to remove ringing in that exact frequency, so it depends on why you are PEQing as to if you adjust wide or narrow (Q).

 

[abdo123]

If you limit the correction to 250Hz then it shouldn’t matter if they move a bit. (Which is something you should be doing already regardless).