REW Measurements, any thoughts?

[strea]

Not sure if this is the place to post this, please move if necessary.

Genelec 7050c + 8030c's driven by Topping E70 in 11ft x 11ft x 9ft room

Green is my 2nd correction.. I had moved my subwoofer into the corner and my speakers were directly on my desktop.
Red is my 3rd correction.. the subwoofer is still in the corner but I moved my speakers up to ear height utilizing K&M 26773 Monitor Stands. Speakers are still same distance from listening position.

All measurements were recorded from the same position for before and after (and many measurements were taken, these are the averages).

Any idea why the big 300hz dip all of the sudden once I moved the monitors up to ear level?

Goal here for me isn't necessarily to correct every little peak and dip, but more to learn about REW, monitors, rooms, and how to optimize. Additionally, I was unable to utilize any settings in REW to get it to boost 300hz. I tried many Individual Max Boost and Overall Max Boost settings, but 300hz wouldn't budge, not sure why.

Thank you for any info/education.

 

[ebslo]

Not sure if this is the place to post this, please move if necessary.

Genelec 7050c + 8030c's driven by Topping E70 in 11ft x 11ft x 9ft room

Green is my 2nd correction.. I had moved my subwoofer into the corner and my speakers were directly on my desktop.
Red is my 3rd correction.. the subwoofer is still in the corner but I moved my speakers up to ear height utilizing K&M 26773 Monitor Stands. Speakers are still same distance from listening position.

All measurements were recorded from the same position for before and after (and many measurements were taken, these are the averages).

Any idea why the big 300hz dip all of the sudden once I moved the monitors up to ear level?

Goal here for me isn't necessarily to correct every little peak and dip, but more to learn about REW, monitors, rooms, and how to optimize. Additionally, I was unable to utilize any settings in REW to get it to boost 300hz. I tried many Individual Max Boost and Overall Max Boost settings, but 300hz wouldn't budge, not sure why.

Thank you for any info/education.
View attachment 268362

It's a reflection off something with a path length about 22 inches longer than the direct path; can't say more than that. I think it's too low to be bounce off the desk. Looking at the spectrogram in wavelet mode may give you some insight into what is happening here, I expect you'll clearly see a cancellation.

As for REW, I would guess the required filter would have too high Q for the frequency it's at. Are both of the Q-related options checked in the EQ filter actions? If so and you really want to try to boost it, then you will have to manually add that filter; however, since it's likely a cancellation, that filter may not be effective.

edit: Also, I bet the smoothing is hiding how deep and narrow it really is.

 

[strea]

It's a reflection off something with a path length about 22 inches longer than the direct path; can't say more than that. I think it's too low to be bounce off the desk. Looking at the spectrogram in wavelet mode may give you some insight into what is happening here, I expect you'll clearly see a cancellation.

As for REW, I would guess the required filter would have too high Q for the frequency it's at. Are both of the Q-related options checked in the EQ filter actions? If so and you really want to try to boost it, then you will have to manually add that filter; however, since it's likely a cancellation, that filter may not be effective.

Thanks for your response ebslo. Could it be a reflection off my face? I was sitting behind the microphone approx 30 inches from the monitors (microphone was only a few inches in front of my face). I figured I shouldn't move so the adjustments include me like they would include other stationary objects. Is this improper thought process / methodology? I clicked spectrogram, but it's empty, must I record new measurements while having that window open?

Yes, both boxes were checked.

 

[strea]

It's a reflection off something with a path length about 22 inches longer than the direct path; can't say more than that. I think it's too low to be bounce off the desk. Looking at the spectrogram in wavelet mode may give you some insight into what is happening here, I expect you'll clearly see a cancellation.

As for REW, I would guess the required filter would have too high Q for the frequency it's at. Are both of the Q-related options checked in the EQ filter actions? If so and you really want to try to boost it, then you will have to manually add that filter; however, since it's likely a cancellation, that filter may not be effective.

edit: Also, I bet the smoothing is hiding how deep and narrow it really is.

Here's unsmoothed.

 

[Rednaxela]

Additionally, I was unable to utilize any settings in REW to get it to boost 300hz. I tried many Individual Max Boost and Overall Max Boost settings, but 300hz wouldn't budge, not sure why.

What does your target look like? Could you post a screenshot of your EQ window?

Also, you could always enter your desired boost manually.

 

[strea]

What does your target look like? Could you post a screenshot of your EQ window?

Also, you could always enter your desired boost manually.

Sure, here is the target (harmon curve download from Julian Krause video). The other line is my original frequency response without any correction.

 

[ebslo]

Thanks for your response ebslo. Could it be a reflection off my face? I was sitting behind the microphone approx 30 inches from the monitors (microphone was only a few inches in front of my face). I figured I shouldn't move so the adjustments include me like they would include other stationary objects. Is this improper thought process / methodology? I clicked spectrogram, but it's empty, must I record new measurements while having that window open?

Yes, both boxes were checked.

Hard to say. I would be surprised because, by your description, the path length difference between direct and face bounce would only be 2*(a few inches). But possibly face + body + back of chair? I don't know. In any case, I would say it's generally incorrect to include yourself in the measurement, but I don't think I could explain why concisely enough to write here. Hopefully others can weigh in on this.

As for the spectrogram, you can generate it for any measurement with an impulse response. For example, if the measurement was taken using the measurement window then REW will be able to generate a spectrogram; if the measurement was saved from the RTA (ie. for MMC method) then it cannot.

 

[ebslo]

Here's unsmoothed.View attachment 268363

Is this summed L+R? One thing to narrow it down, compare (ie. overlay) L and R. If both have this cancellation at the same frequency then it's probably something on the vertical plane. If they are asymmetric (ie. only one has dip, or dips are at substantially different frequencies) then it is likely something off-center in the horizontal plane.

 

[Somafunk]

Perhaps try and reduce the height of your speakers and have them angled up towards your ears then take a measurement, similarly have the speakers higher than your ears and angled down and take a measurement.

Compare all the measurements and see what happens to the dip then you’ll be a step closer to figuring it out at least

 

[strea]

Is this summed L+R? One thing to narrow it down, compare (ie. overlay) L and R. If both have this cancellation at the same frequency then it's probably something on the vertical plane. If they are asymmetric (ie. only one has dip, or dips are at substantially different frequencies) then it is likely something off-center in the horizontal plane.

Yes, my output was set to L+R. I'll give this a try tomorrow, thank you. And yes, my head/body/chair are all close, perhaps I'll try moving out of the way and pulling it all back lol.

 

[ppataki]

Sure, here is the target (harmon curve download from Julian Krause video). The other line is my original frequency response without any correction.
View attachment 268364

Try reducing the Target Level to 67dB and use 0 boosting (both Individual and Overall). Whatever dip you will have there after applying this method I would just leave it there and would not try to compensate for it - having said that you can try boosting it still but that might introduce excessive ringing there.

 

[Rednaxela]

Sure, here is the target (harmon curve download from Julian Krause video). The other line is my original frequency response without any correction.
View attachment 268364

Thank you.

The reason it leaves the 300Hz dip alone is that it is 9dB below your target, and your individual max boost setting is 6dB.

Hope this makes sense.

 

[strea]

Try reducing the Target Level to 67dB and use 0 boosting (both Individual and Overall). Whatever dip you will have there after applying this method I would just leave it there and would not try to compensate for it - having said that you can try boosting it still but that might introduce excessive ringing there.

Thanks ppataki, this is the result of that. Although, this warning box pops up in the bottom right. Seems that this produced a pretty flat and amazing respeonse for my room, what would be the downside of this?

 

[ppataki]

Seems that this produced a pretty flat and amazing respeonse

It will also sound like amazing, not just the looks
The downside is the efficiency loss - your previous target curve was at 77.3dB, the new one is 67db = you lost 10.3dB of efficiency. It is kind of like you replaced your speaker with another one with a 10dB lower sensitivity.
This is a non-issue if you don't listen at 'neighbors calling the police' levels and if you have adequate amp power to compensate for the loss.
But I would recommend just listening to it: if you like it you keep it, if not, there are still other ways you can try
(btw this is the method I use for all my systems and for me it works perfectly all the time, just check my projects in the DIY section of this forum)

 

[strea]

It will also sound like amazing, not just the looks
The downside is the efficiency loss - your previous target curve was at 77.3dB, the new one is 67db = you lost 10.3dB of efficiency. It is kind of like you replaced your speaker with another one with a 10dB lower sensitivity.
This is a non-issue if you don't listen at 'neighbors calling the police' levels and if you have adequate amp power to compensate for the loss.
But I would recommend just listening to it: if you like it you keep it, if not, there are still other ways you can try
(btw this is the method I use for all my systems and for me it works perfectly all the time, just check my projects in the DIY section of this forum)

So it's like changing the input sensitivity on my Genelec 8030c's from +6 to +16 essentially?

This may actually be very good for me considering I was planning to get an ADI-2 DAC, but supposedly it requires attenuators in order to run the 8030c's in it's optimal dynamic range (-18dBr to +2.5 dBr on the ADI-2) and typically that requires 10-30db of attenuation via something like a Shure a15as. Seems that this is almost like a digital method of attenuation that will be quite helpful for my situation and provide a great flat response, so a win win?

Thanks again.

 

[ppataki]

Thanks again.

Welcome!

his may actually be very good for me considering I was planning to get an ADI-2 DAC, but supposedly it requires attenuators in order to run the 8030c's in it's optimal dynamic range (-18dBr to +2.5 dBr on the ADI-2) and typically that requires 10-30db of attenuation via something like a Shure a15as. Seems that this is almost like a digital method of attenuation that will be quite helpful for my situation and provide a great flat response, so a win win?

Well, the input signal will not get impacted by this method, so the output signal from the ADI-2 will go to your Genelecs as-is
So I don't think it will be the same as with using a Shure a15as - but again now you have this method in your hands, just try it and use your ears as the final jury
(I would let others comment on this attenuation topic, I am sure there are people here with ample knowledge on this subject)

 

[strea]

Welcome!


Well, the input signal will not get impacted by this method, so the output signal from the ADI-2 will go to your Genelecs as-is
So I don't think it will be the same as with using a Shure a15as - but again now you have this method in your hands, just try it and use your ears as the final jury
(I would let others comment on this attenuation topic, I am sure there are people here with ample knowledge on this subject)

Understood. Although, when toggling the correction on/off in EQ APO, I can hear a very obvious volume change (10db or so perhaps as the correction predicted) so it may not be adjusting the sensitivity exactly, but it does seem like it's serving a similar purpose but working in a different way type of thing. Anyways, thanks a lot for your help, going to listen with this correction today and see how everything sounds.

 

[thewas]

Seems that this produced a pretty flat and amazing respeonse for my room, what would be the downside of this?

In my long experience such broadband with level reduction with negative peaking filters not only significantly reduces the dynamic range of your loudspeakers but also makes the sound dull as such a "chain" of peaking filters have also temporal effects.

Some general advice which works usually well in most typical cases:

1) Place your target curve somewhere in the middle of your response without correction so you have similar sized peaks over and dips under it (so in your above case approx. 10 dB higher than now)

2) Measure and correct left and right channels separately, ideally with the moving microphone method

3) Use various smoothing

3) Below transition frequency (usually 250-500 Hz) let REW auto EQ correct the peaks but not the dips

4) Above transition frequency (usually 250-500 Hz) find and use EQ for your loudspeaker based on anechoic data, https://www.spinorama.org/eqs.html (your 8030C don't
need significant correction there so you could even ignore this step)

 

[ebslo]

Thanks ppataki, this is the result of that. Although, this warning box pops up in the bottom right. Seems that this produced a pretty flat and amazing respeonse for my room, what would be the downside of this?

It's implicit in what @thewas already said, but I'll say it explicitly. It is "correcting" above the transition frequency based on in-room measurements instead of anechoic response. Many here can explain this better than me (and I'll probably get some of it wrong) but I'll give my best shot. Basically, perception of timbre and locality are based on "direct" sound (determined by first arrival), but your in-room measurement is the summation of the direct sound and all the reflections. Your Genelec 8030c's produce a very flat direct-sound frequency response based on anechoic or pseudo-anechoic measurements. Therefore, when you correct based on direct+reflected sound, you make the direct sound less flat, which, as the theory goes, reduces accuracy of perceived timbre and localization.

 

[JustJones]

Are you running your signal through the 7050c then on to the 8030c to use any of the 7050 bass management dip switches? Are you using any of the tone controls on the 8030's.?