Home Theater Forum and Systems banner

1 - 20 of 20 Posts

·
Registered
Joined
·
13 Posts
Discussion Starter · #1 ·
Hi Guys,

First time user today of REW so please forgive my noobish question. I received my UMC minidsp USB mic today and ran my first two measurements. Firstly, on my sub from 10hz-200hz, then a full range sweep on my mains.

Now what I am confused about (again please forgive me for stupid question) is where do I need to set target level? Do I run my test tones to a level of say 75db, therefore that will become what I target for when I want to apply EQ filters? Sorry I just don't get what target level should be.

I've attached two pics first the response of my subwoofer then then of my main speakers.

Any feedback on the graph and how/what I should look for to set target level would be appreciated.

sub 1.jpg

main speaker.jpg
 

·
Banned
Joined
·
4,838 Posts
It is a very good question. A couple of things to remember when EQing speakers

Boosting the narrower and deeper valleys and dips very much is not recommended. They are usually due to phase cancellations from room modes, which can only be corrected acoustically. If you try to boost them it is like pouring energy into a black hole. You will not make a significant difference, and will probably be causing problems elsewhere in the room and frequency response curve, plus pushing the drivers to do the impossible, causing distortion, etc. But you can cut the peaks quite safely.

Also, and some may disagree with this point, but with speaker EQ, LESS IS MORE. You can EQ a curve to flat and possibly mess up the soundstage and imaging (SS&I) and end up with a disappointing sound. Or apply minimal EQ to cut the worst peaks and add some gentle boost to broader dips and end up with a much nicer sound even though the response curve is not flat.

I like to set the speakers for best SS&I first, then apply minimal EQ to L & R equally, preserving the SS&I characteristics.

Knowing all this affects the setting of the target curve and other REW filter parameters. First, apply 3rd octave smoothing so you do not end up with an over-abundance of filters. Then:
  • Target Level = 67 dB
  • Individual Max Boost = 3 dB
  • Overall Max Boost = 3 dB
  • Flatness Target = 3 dB
See what that gives you.
 

·
Registered
Joined
·
13 Posts
Discussion Starter · #3 ·
Thanks for your feedback, hear you loud and clear I was only going to cut rather then boost any frequencies, when you say target 67db do you mean my full range measurement or my sub as well?
 

·
Registered
Joined
·
13 Posts
Discussion Starter · #5 ·
Thanks raver, what I'm still not sure of is how do you determine the target level? Reason I'm confused if I take 67db for my full range graph that looks reasonable but for the sub the entire frequency response is above the target line which means every point will be cut? Sorry it seems like stupid question
 

·
Banned
Joined
·
4,838 Posts
Perhaps I misunderstood. I was looking at the full range plot which included the subwoofer. Will you have a separate EQ device for the subwoofer or will there be a single device which EQs the entire range including the main speaker and the subwoofer together? For a single full range EQ 67 dB looks like the right number. For separate EQ of the subwoofer frequencies, 75 dB looks better.
 

·
Registered
Joined
·
13 Posts
Discussion Starter · #7 ·
Perhaps I misunderstood. I was looking at the full range plot which included the subwoofer. Will you have a separate EQ device for the subwoofer or will there be a single device which EQs the entire range including the main speaker and the subwoofer together? For a single full range EQ 67 dB looks like the right number. For separate EQ of the subwoofer frequencies, 75 dB looks better.

Great thanks for clarification mate, yeh it would be separate EQ, and your advise is kinda where I thought, late 60's for the mains and 70's for the sub.

Also do you mind having a look at my spec graph, If Im reading this right, it seems like a reasonable good decay result. As the yellow (which I am assuming relates to the higher DB readings) is within 100ms. Whilst there is reverb time out to 400ms it seems to be at low db levels, any thoughts on how it looks would be great.

spectrograph.jpg
 

·
Banned
Joined
·
4,838 Posts
Looks like T30 is about .3 to .4 seconds. So T60 is probably under .5 to .6 seconds, which is certainly as long as I would want it to be. It is very even, which is good, and dropping slightly at high frequencies, also good.

For a listening room, this is probably very nice, depending on your personal taste. I like T60 in the .3 to .4 range myself, but the norm is higher than that.

For a studio control or mixing room, an ideal T60 would be .4 or under.

It is very even, kinda hate to suggest that you mess with it. Also curious what the impulse diagram looks like at the listening position, that will tell a lot about soundstage and imaging.
 

·
Registered
Joined
·
13 Posts
Discussion Starter · #9 ·
Hey mate yeh I thought it looked pretty even as well, only been playing for 24 hrs but I'll post the impulse graph for you to have a look at :)
 

·
Registered
Joined
·
13 Posts
Discussion Starter · #10 ·
Looks like T30 is about .3 to .4 seconds. So T60 is probably under .5 to .6 seconds, which is certainly as long as I would want it to be. It is very even, which is good, and dropping slightly at high frequencies, also good.

For a listening room, this is probably very nice, depending on your personal taste. I like T60 in the .3 to .4 range myself, but the norm is higher than that.

For a studio control or mixing room, an ideal T60 would be .4 or under.

It is very even, kinda hate to suggest that you mess with it. Also curious what the impulse diagram looks like at the listening position, that will tell a lot about soundstage and imaging.

Hey Raver, as requested please see attached, to be honest I have no idea what to look for on this one???
Any information on what to look out for would be appreciated :)

Mic was approx 3m in Listening Position

impulse db.jpg

impulse %.jpg
 

·
Banned
Joined
·
4,838 Posts
If you could post the .mdat file, that would be handy. What I like to look at is the lower version of the plot (%FS) from -0.001 to 0.020 , basically the first 20 mS after the initial impulse, with just the left and right mains (or left+sub and right+sub), just the 2 plots showing together on the Overlays panel and separated (with the Separate The Traces button) so we can see how the early reflections from the two coordinate.

Thanks.
 

·
Registered
Joined
·
13 Posts
Discussion Starter · #12 ·
If you could post the .mdat file, that would be handy. What I like to look at is the lower version of the plot (%FS) from -0.001 to 0.020 , basically the first 20 mS after the initial impulse, with just the left and right mains (or left+sub and right+sub), just the 2 plots showing together on the Overlays panel and separated (with the Separate The Traces button) so we can see how the early reflections from the two coordinate. Thanks.

Are ok so you want a separate measurement for left and right main speaker?
 

·
Registered
Joined
·
13 Posts
Discussion Starter · #16 ·
If you could post the .mdat file, that would be handy. What I like to look at is the lower version of the plot (%FS) from -0.001 to 0.020 , basically the first 20 mS after the initial impulse, with just the left and right mains (or left+sub and right+sub), just the 2 plots showing together on the Overlays panel and separated (with the Separate The Traces button) so we can see how the early reflections from the two coordinate.

Thanks.
Hey Raver, just got a chance to do a separate measurement, attached is a pic summed left and right and also the config files, thanks very much for your help very kind of you ;)

View attachment Speaker Left Impulse.mdat

View attachment Speaker Right Impulse.mdat

combined l + r impulse.jpg
 

·
Banned
Joined
·
4,838 Posts
It looks like
  • Your room is not exactly left-right symmetrical; if it was, the sequences of reflections would be the same, or close to it, for the two plots. Probably not much you can do about that.
  • Your speakers and LP are not set up with left-right symmetry in the room - there IS something you can do about that.
Early reflections (first 20 mS) have a major influence on the quality of soundstage and imaging you will experience. Where there are strong reflections - you see "sets" of reflections at 5.3, 9.8, and 13.5 mS on the left plot, and at 4.3, 8.4, and 11.9 mS on the right plot - it is important that they arrive at the LP coordinated, at the same time. That way imaging is precise and the soundstage is strong, clear, natural, and open.

With your right early reflections arriving a full mS and more earlier than the left, and spaced out differently, the imaging will be shifted right and will be soft and difficult to localize. You might re-position speakers &/or LP to even out the timing of these early reflections. You might be surprised at the difference it makes. You would ideally like them timed to arrive withing 0.1 mS of each other. That sounds really picky, but that is the precision that it takes. Trust me, the benefits will be well worth the trouble. The result may include better matching of the L and R frequency response curves, tool.
 

·
Registered
Joined
·
13 Posts
Discussion Starter · #18 ·
It looks like [*]Your room is not exactly left-right symmetrical; if it was, the sequences of reflections would be the same, or close to it, for the two plots. Probably not much you can do about that. [*]Your speakers and LP are not set up with left-right symmetry in the room - there IS something you can do about that. Early reflections (first 20 mS) have a major influence on the quality of soundstage and imaging you will experience. Where there are strong reflections - you see "sets" of reflections at 5.3, 9.8, and 13.5 mS on the left plot, and at 4.3, 8.4, and 11.9 mS on the right plot - it is important that they arrive at the LP coordinated, at the same time. That way imaging is precise and the soundstage is strong, clear, natural, and open. With your right early reflections arriving a full mS and more earlier than the left, and spaced out differently, the imaging will be shifted right and will be soft and difficult to localize. You might re-position speakers &/or LP to even out the timing of these early reflections. You might be surprised at the difference it makes. You would ideally like them timed to arrive withing 0.1 mS of each other. That sounds really picky, but that is the precision that it takes. Trust me, the benefits will be well worth the trouble. The result may include better matching of the L and R frequency response curves, tool.
Hey Craver that's for taking the time to have a look at results appreciated so if I am calculating it correctly it looks like the left speaker reflections are between 1.7m and 4.5m, eg its reflecting off surfaces at roughly those points , would you suggest trying to find those reflection points (on both sides and defuse if possible or is the more practical step to try and bring forward my left speaker slightly to align the time of arrival. You are right placement of speakers is somewhat compromised at moment due to layout of room!


Thanks mate really appreciated your advise and input
 

·
Banned
Joined
·
4,838 Posts
I would try it first by moving speakers. Those early reflections are not a bad thing if they all line up. L/R symmetrical setup, down to the quarter inch (seriously!), is rule #1 for getting good soundstage & imaging, although not always possible. Physical measurements get you close, but no substitute for fine tuning with L/R impulse plots. Then the early reflections that can not be paired and synchronized can then be dealt with via dissipation or absorption.

Do not hesitate to check with our experts in the Home Audio Acoustics forum, they are the real experts on the topic.
 

·
Registered
Joined
·
13 Posts
Discussion Starter · #20 ·
I would try it first by moving speakers. Those early reflections are not a bad thing if they all line up. L/R symmetrical setup, down to the quarter inch (seriously!), is rule #1 for getting good soundstage & imaging, although not always possible. Physical measurements get you close, but no substitute for fine tuning with L/R impulse plots. Then the early reflections that can not be paired and synchronized can then be dealt with via dissipation or absorption. Do not hesitate to check with our experts in the Home Audio Acoustics forum, they are the real experts on the topic.

Thanks for all your advice:)
 
1 - 20 of 20 Posts
Top