Room correction/DSP

[Camverton]

I don't ever measure both speakers at the same time or we could test your hypothesis. From which source did you get the idea that the two speakers should be measured at the same time?

When measuring my speaker’s in room response I usually measure left + right together, and then left and right separately. The combined response is useful for assessing lower frequencies (as mentioned somewhere in the REW info), particularly decay, but isn’t good for higher frequencies when looking at individual speaker measurements is more accurate. Measurements of individual speakers are also useful in the lower frequencies but I find it useful to also see how the two speakers are interacting with each other.

 

[flak monkey]

@tuga @newlash09 I see benefits to both approaches. But understanding the fundamentals/theory also can allow you to make better informed decisions. I enjoy reading the theory as it gives me a better fundamental understanding and let's me see where I'm going wrong. However, sound is very subjective, so while I like using measurements as a basis, I still like to listen and adjust things to preference.

Turn out that Roon maybe does have individual channel EQ, I need to play with that today. Will share my "half baked" results later...

 

[MartinC]

However, sound is very subjective, so while I like using measurements as a basis, I still like to listen and adjust things to preference.

This is absolutely the right thing to do  . 

 

[Andrew_C]

@tuga @newlash09 I see benefits to both approaches. But understanding the fundamentals/theory also can allow you to make better informed decisions. I enjoy reading the theory as it gives me a better fundamental understanding and let's me see where I'm going wrong. However, sound is very subjective, so while I like using measurements as a basis, I still like to listen and adjust things to preference.

Turn out that Roon maybe does have individual channel EQ, I need to play with that today. Will share my "half baked" results later...

How do you do that in roon?

 

[newlash09]

This is absolutely the right thing to do  . 

I would fully agree sir..each on their own journey to learn along the way I guess.

 

[newlash09]

How do you do that in roon?

The lady was trying to mend posts. So please explain yourselves. Thanks 

 

[MartinC]

I don't ever measure both speakers at the same time or we could test your hypothesis.

If you use the same time reference for both, you can measure each speaker response separately and then calculate the sum in REW. I'm somewhat wondering what you are describing as a hypothesis though so let's break this down...

Do you understand and agree with my key point? This is, if you were able to EQ the amplitude response of each speaker at the listening position to be identically smooth (flat or whatever shape you want), the sum would not be flat if there were differences in the phase responses of the two. This is mathematical and therefore acoustical fact, not a some sort of hypothesis.

To take an extreme example just to maybe help explain the point - if you have an equal amplitude from each speaker but they end up being perfectly out of phase at the listening position, then they will sum to zero. Consider panning the signal level from all on the left, through equal on each channel, to all on the right. What would be heard at the listening position would be that the sound level will start at one level and gradually drop as the signal moves to the center, and then rise again as it's moved to the right to get back up to the starting level. For the low-bass frequencies I'm talking about this is purely a level change, not something that affects the direction the sound appears to be coming from.

You very likely don't normally look at the phase responses but for info. when the measured amplitude responses for the two speakers are different it is basically guaranteed that the phase responses will differ too. It's a consideration I've likely looked at more as it affects my choice of crossover frequency between main speakers and subwoofer. I first realised the issue when considering EQ for my own speakers but then switched to using a subwoofer which removed the complication  .

For clarity let me also stress that I have not dogmatically said something like 'everyone must EQ low-bass based on the sum of left and right speaker signals'. What I've done is explain that there is reason to consider both the individual and summed signals. The nature of low-bass signals in I suspect most recordings does though make me lean towards prioritising using the combined response. What this certainly means is that I won't pronounce as 'wrong' someone who does this and likes the result.

From which source did you get the idea that the two speakers should be measured at the same time?

I'm afraid what I've shared above has come from my own brain, based on application of my own knowledge and experience rather than something that I can point you to some particular source for.

 

[flak monkey]

How do you do that in roon?

Procedural EQ

 

[tuga]

When measuring my speaker’s in room response I usually measure left + right together, and then left and right separately. The combined response is useful for assessing lower frequencies (as mentioned somewhere in the REW info), particularly decay, but isn’t good for higher frequencies when looking at individual speaker measurements is more accurate. Measurements of individual speakers are also useful in the lower frequencies but I find it useful to also see how the two speakers are interacting with each other.

That makes sense (regarding decay). Interaction information I don’t find particularly useful but it could be if you are using a sub (or a pair of subs).

 

[Camverton]

That makes sense (regarding decay). Interaction information I don’t find particularly useful but it could be if you are using a sub (or a pair of subs).

When listening we hear the two speakers playing together and so it is useful to measure them playing together. Summing the left and right measurements doesn’t necessarily give the same result, although it might be similar in a very symmetrical room layout. 

For higher frequencies, separate speaker measurements are best, but for lower frequencies I find separate measurements useful to see what is happening but a measurement with both speakers playing useful to see what we are actually hearing. 

 

[StingRay]

When listening we hear the two speakers playing together and so it is useful to measure them playing together. Summing the left and right measurements doesn’t necessarily give the same result, although it might be similar in a very symmetrical room layout. 

For higher frequencies, separate speaker measurements are best, but for lower frequencies I find separate measurements useful to see what is happening but a measurement with both speakers playing useful to see what we are actually hearing. 

I don't get listening to one speaker which I have read in some reviews, are they inputing in mono or just playing one channel. With 1 speaker you won't get  the same soundstage, totally different and some music won't make sense, I don't see the point.

 

[rdale]

So you feel that your comments are less agreesive and ad hominem by adding a dumb

Good for you.

Yes I use graphs to prove and also to illustrate a point. I also find it strange that people would rather read reviews instead of learning about audio. Strange world...

The primary purpose of this part of the forum is to share subjective experiences of equipment and accessories we've tried in our own HiFi or what we have heard in shop demos, bakeoffs and shows. In that context it makes perfect sense to discuss relevant HiFi reviews too. What we haven't signed up for is 'HiFi as a science experiment' and there are other forums which might be more targeted at that.

 

[MartinC]

When listening we hear the two speakers playing together and so it is useful to measure them playing together. Summing the left and right measurements doesn’t necessarily give the same result, although it might be similar in a very symmetrical room layout. 

The result will be the same no matter what the layout, provided the same and accurate timing reference is used for both measurements, and the sum is the (mathematically) complex sum rather than simply adding the amplitude responses. 

In practice for most it's probably simplest to just measure the two playing together though.

 

[Camverton]

I don't get listening to one speaker which I have read in some reviews, are they inputing in mono or just playing one channel. With 1 speaker you won't get  the same soundstage, totally different and some music won't make sense, I don't see the point.

I think that’s the distinction between measuring to find out what is going on and listening to the music as the final result; the former done to improve the latter - hopefully.

As you will gather from this thread there are different approaches to doing this measurement malarkey. When it comes to using REW I err towards the views of the designer of the software, but in the last analysis REW is a tool and as with many tools there are different, perfectly valid, ways of using it.

 

[flak monkey]

Some results from messing around at lunchtime. I won't pretend these are perfect, but I thought I would play with EQ of individual speakers or both at the same time. 

Left speaker, pre and post EQ (blue pre, green post)



Right speaker pre and post EQ (purple pre, red post)



Individual measurements of left (green), right (red) and combined (blue) post EQ



As far as the difference between EQing both together or separately, here's a direct comparison - Red is the result of EQing both speakers together, no individual measurements made. The blue is the result of EQing both speakers independently and then summing the result.

 

[tuga]

When listening we hear the two speakers playing together and so it is useful to measure them playing together. Summing the left and right measurements doesn’t necessarily give the same result, although it might be similar in a very symmetrical room layout. 

For higher frequencies, separate speaker measurements are best, but for lower frequencies I find separate measurements useful to see what is happening but a measurement with both speakers playing useful to see what we are actually hearing. 

I don't agree.

Unless the speakers are positioned in such a way that their response below 300Hz is identical, applying the same EQ correction to two speakers with a different response will neither make the individual nor the combined sound flatter at the listening spot. See below at 64Hz, 87Hz, 100Hz...

 

[tuga]

If you use the same time reference for both, you can measure each speaker response separately and then calculate the sum in REW. I'm somewhat wondering what you are describing as a hypothesis though so let's break this down...

Do you understand and agree with my key point? This is, if you were able to EQ the amplitude response of each speaker at the listening position to be identically smooth (flat or whatever shape you want), the sum would not be flat if there were differences in the phase responses of the two. This is mathematical and therefore acoustical fact, not a some sort of hypothesis.

To take an extreme example just to maybe help explain the point - if you have an equal amplitude from each speaker but they end up being perfectly out of phase at the listening position, then they will sum to zero. Consider panning the signal level from all on the left, through equal on each channel, to all on the right. What would be heard at the listening position would be that the sound level will start at one level and gradually drop as the signal moves to the center, and then rise again as it's moved to the right to get back up to the starting level. For the low-bass frequencies I'm talking about this is purely a level change, not something that affects the direction the sound appears to be coming from.

You very likely don't normally look at the phase responses but for info. when the measured amplitude responses for the two speakers are different it is basically guaranteed that the phase responses will differ too. It's a consideration I've likely looked at more as it affects my choice of crossover frequency between main speakers and subwoofer. I first realised the issue when considering EQ for my own speakers but then switched to using a subwoofer which removed the complication  .

For clarity let me also stress that I have not dogmatically said something like 'everyone must EQ low-bass based on the sum of left and right speaker signals'. What I've done is explain that there is reason to consider both the individual and summed signals. The nature of low-bass signals in I suspect most recordings does though make me lean towards prioritising using the combined response. What this certainly means is that I won't pronounce as 'wrong' someone who does this and likes the result.

I'm afraid what I've shared above has come from my own brain, based on application of my own knowledge and experience rather than something that I can point you to some particular source for.

It is never wrong if someone likes the result. It just does not make for a more accurate transduction of the recording/sugnal at the listening spot.

When room boundaries affect frequency response do they not also affect phase?

 

[tuga]

Some results from messing around at lunchtime. I won't pretend these are perfect, but I thought I would play with EQ of individual speakers or both at the same time. 

Left speaker, pre and post EQ (blue pre, green post)



Right speaker pre and post EQ (purple pre, red post)



Individual measurements of left (green), right (red) and combined (blue) post EQ



As far as the difference between EQing both together or separately, here's a direct comparison - Red is the result of EQing both speakers together, no individual measurements made. The blue is the result of EQing both speakers independently and then summing the result.

Would you mind exporting the plots with the vertical (Y) scale reduced to 50dB and apply a bit less smoothing (perhaps 1/24 octave)? As it stands the resolution is very limited.

 

[rdale]

I don't agree.

Unless the speakers are positioned in such a way that their response below 300Hz is identical, applying the same EQ correction to two speakers with a different response will neither make the individual nor the combined sound flatter at the listening spot. See below at 64Hz, 87Hz, 100Hz...

If two speakers are pressurizing the room with bass, the bass of the two will interfere with each other whether or not they both have the same EQ. So you cannot set the EQ anymore correctly by simply measuring one speaker or one sub at a time because it wouldn’t account for that interference.

 

[flak monkey]

Would you mind exporting the plots with the vertical (Y) scale reduced to 50dB and apply a bit less smoothing (perhaps 1/24 octave)? As it stands the resolution is very limited.

Sure - same again but with more wiggles to show just how crap my room really is!

Left before (blue) and after (green)

Right before (purple) and after (red)



Individually EQ - left (red), right (green), both (blue)



And comparing the overall response of EQ of both together (red) or individually (blue)