My Rew(view)

[mimoza]

Hello and best regards from France,
first sorry for my english and poor sens of humour,
i need some inputs on my REW mesurements, a kind of REW(view) ,
my room is 3.5 deep by 4.1m wide, ceiling 2.5m,
back wall is made of 2cm high density foam + akupanels (i thought it would be enough for correct reverberation time, but my mesurements seems to be not as good as excpected)

my usage : 75% Home theater / 25% music
here is the file : https://files.fm/u/jy8ngh3vu

my setup:
AVR : marrantz cinema 50
FRONT : cabasse java mt4 (20 years old)
LFE : svs 2000 pro
TOP : focal icw6 300
SR : cabasse IO2

how does it sound to me ? Good, but can i acheive better?

nosubwoofer on those recordings i just talk about the stereo here (adding the sub is not a good improvement on my system)

here are some pictures:

here are some screen captures with var smoothing

R

L

LR

LR with dirac

i am not sure about interpreting those recordings, how does it look for you?
what do you think i can improve? under 1500Hz seems not very good
shall i go further with room correction? or is correct?

what do you think about L speaker response? is it due too the glass window?

merci par avance / thanks in advance !

 

[DJBonoBobo]

LR Dirac looks wrong because you measured both channels at the same time (-> cancellations between L and R). Better to measure L and R seperately and calculate an average with REW. Maybe show that again then.

Bass looks good, you really don't need a sub for stereo. Personally, i always find the Dirac standard target too bass heavy. You may dial it down a bit if you want, but that's up to taste mostly.

 

[Rednaxela]

Perhaps a couple of things that might help the conversation.


The 1/48 graphs are mainly noise in the context of your question. Var smoothing is fine.

Use the All SPL tab instead of SPL & Phase. It enables you, for instance, to show two plots in one graph.

There’s a Capture button in the upper left corner (camera icon). Please use it instead of screenshots. In there, select the 50 dB/dec aspect ratio, to make the plots you share easier to compare.


Hope this helps!

 

[mimoza]

LR Dirac looks wrong because you measured both channels at the same time (-> cancellations between L and R). Better to measure L and R seperately and calculate an average with REW. Maybe show that again then.

Bass looks good, you really don't need a sub for stereo. Personally, i always find the Dirac standard target too bass heavy. You may dial it down a bit if you want, but that's up to taste mostly.

thanks for reply, in real world i guess it would happen also, no ? isnt it more realistic to have the real both channel response also ? rather than an averaged response? if needed ill post it tomorrow
what do you think about the bass decay time?

 

[mimoza]

Perhaps a couple of things that might help the conversation.


The 1/48 graphs are mainly noise in the context of your question. Var smoothing is fine.

Use the All SPL tab instead of SPL & Phase. It enables you, for instance, to show two plots in one graph.

There’s a Capture button in the upper left corner (camera icon). Please use it instead of screenshots. In there, select the 50 dB/dec aspect ratio, to make the plots you share easier to compare.


Hope this helps!

thanks for your reply and the tips
below some more screenshots but what about decay time, is it ok?

 

[DJBonoBobo]

thanks for reply, in real world i guess it would happen also, no ? isnt it more realistic to have the real both channel response also ? rather than an averaged response?

No, because you have two ears it is different anyway.

what do you think about the bass decay time?

Looks fine, i think. If you are considering more absorbers, i would focus on the 150-300Hz area (thick and with a scatterplate or something to avoid more treble Absorption)

 

[mimoza]

decay from capture function

 

[mimoza]

No, because you have two ears it is different anyway.

true i have two
thanks for the bass tip

 

[Rednaxela]

LR Dirac looks wrong because you measured both channels at the same time (-> cancellations between L and R). Better to measure L and R seperately and calculate an average with REW.

Interesting take. Never heard of this technique.

 

[TheZebraKilledDarwin]

No, because you have two ears

But only one brain. We are not really hearing with the ears, but with the brain.

LR Dirac looks wrong because you measured both channels at the same time (-> cancellations between L and R). Better to measure L and R seperately and calculate an average with REW.

Acoustically a room is a linear system. That means recording the speakers separately and summing the measurements, should give the same result as measuring the acoustic sum of the speakers. Is there a reason, why this wouldn't apply here?

 

[mimoza]

here the test seperate channel, both and sperate with average response

dirac

direct:

 

[ZolaIII]

It's pretty much typical room response with 50 Hz boost and 75 Hz deep. For starters put a thick curtains on windows and pool the speakers a bit more out (as much as you can). I guess listening position is where kid whose sitting and reading and you did measurements from there. That's a bit of a problem because highs refractions are direct from the walls which is very close to your ears. As room configuration is wide I suggest diffusion panels behind you. Closed enclosure subwoofer's would help a lot by boosting the level up and then cutting it more so than the deep at 75 Hz get smaller and crossed higher (about 100 Hz) it would also help filling the mid and uper bass better. Try two pass REW PEQ's instead of Dirac (I do) and it gets better results. Waterfall plots don't look bad.

 

[TheZebraKilledDarwin]

When I saw the pics, honestly I didn't expect such nicely controlled acoustics. Not bad for that room and the extremely short axis.

I see two potential problems.

Does it hurt the ears, when Linda Blair in "The Exorcist" is shouting extremely loud? If yes, then I would check that 1140 range. With 1/3 (or var) smoothing you see, that it alone brings the range 1 dB over the 700-800 region. It should be less loud than the 700 range, or every hysterical women will always hurt your ears...

Another potential problem may be the very loud 4k (3.9k) range. Generally you do not want the 4k to be louder than the 3k.
If you notice intelligbility problems after bringing that range down where it belongs, for movies you do not want to improve it by boosting the treble range, but by paying attention to the upper midrange (in your case, you have a 1,4-1,5k dip).

 

[DJBonoBobo]

That means recording the speakers separately and summing the measurements, should give the same result as measuring the acoustic sum of the speakers. Is there a reason, why this wouldn't apply here?

It is simple: When i measure LR together or L and R seperately i get different results. The LR one looks wrong, the others right. So i don't use the LR, because it worries me and it does no good trying to "correct" this.
I don't know the theory behind this, just an empirical experience.

 

[TheZebraKilledDarwin]

I don't know the theory behind this, just an empirical experience.

Theory is good, but empirical knowledge is better. Interesting. I never listen in stereo, but have to check that out.

 

[Rednaxela]

Averaging L and R will give a result that won’t change when you switch one speaker’s polarity right? If so, how can it be a viable method?

 

[DJBonoBobo]

Averaging L and R will give a result that won’t change when you switch one speaker’s polarity right? If so, how can it be a viable method?

I am talking about room eq here. For this, i don't touch the higher frequencies anyway, but i need them to calculate a target. This works better with seperate measurements and averaging, if i want to combine L and R.
I assume the polarity to be correct, i don't check this with a measurement. I don't say it is a viable method for everything, just in the context here.

 

[TheZebraKilledDarwin]

Averaging L and R will give a result that won’t change when you switch one speaker’s polarity right? If so, how can it be a viable method?

Phase is omitted in REW, when averaging?

 

[DJBonoBobo]

Hello and best regards from France,
first sorry for my english and poor sens of humour,
i need some inputs on my REW mesurements, a kind of REW(view) ,
my room is 3.5 deep by 4.1m wide, ceiling 2.5m,
back wall is made of 2cm high density foam + akupanels (i thought it would be enough for correct reverberation time, but my mesurements seems to be not as good as excpected)

my usage : 75% Home theater / 25% music
here is the file : https://files.fm/u/jy8ngh3vu

my setup:
AVR : marrantz cinema 50
FRONT : cabasse java mt4 (20 years old)
LFE : svs 2000 pro
TOP : focal icw6 300
SR : cabasse IO2

how does it sound to me ? Good, but can i acheive better?

nosubwoofer on those recordings i just talk about the stereo here (adding the sub is not a good improvement on my system)

here are some pictures:

View attachment 287062

View attachment 287059

View attachment 287060

View attachment 287056
here are some screen captures with var smoothing

R

View attachment 287051


L
View attachment 287052


LR

View attachment 287053

LR with dirac

View attachment 287050

i am not sure about interpreting those recordings, how does it look for you?
what do you think i can improve? under 1500Hz seems not very good
shall i go further with room correction? or is correct?

what do you think about L speaker response? is it due too the glass window?

merci par avance / thanks in advance !

BTW, this is how your ETC looks like - left channel (Impulse -> Envelope).

You have quite a lot early reflections. The one -6 dB at around 2.5 ms is probably the reflection from the back wall (behind the LP). Others are probably the ceiling, maybe the floor and also the side walls.
There is no consensus how to deal with early reflections, but f you want to change something, you could try working on them.
Some say that you should aim for no peaks higher than -20 or -15 in the first 20ms (=20m). Others say, they can be beneficial. I can´t tell you what to do, but that´s one more thing you can look at.

 

[TheZebraKilledDarwin]

Theory is good, but empirical knowledge is better. Interesting. I never listen in stereo, but have to check that out.

You are correct. Just verified your observation. The REW averaged SPL graph indeed looks significantly different compared to the measured superposition.
This seems to happen because of omitting the phase response for averaging.