Gearslutz

Gearslutz (https://www.gearslutz.com/board/)
-   Studio Building / Acoustics (https://www.gearslutz.com/board/studio-building-acoustics/)
-   -   Monitors dip at 1KHz (https://www.gearslutz.com/board/studio-building-acoustics/1236049-monitors-dip-1khz.html)

Thavma 28th October 2018 11:37 PM

Monitors dip at 1KHz
 
Hello guys,

I have a dip at 1KHz with Adam A5Xmonitors, I have measured them separately but there is no dip, they only have dip when they both play in the sweet spot.

Could anyone help me to identify the problem as I am not a professional in acoustics?

I do have a pretty well treated room

Here is the problem

https://image.ibb.co/gPXUgV/44798450...22302976-n.jpg

I have done realtime measurement in each monitor closely but they play nicely. The dip is when both are playing


I am making measurements with Sonarworks Reference microphone and software.

Thanks a lot guys

pentagon 29th October 2018 04:39 AM

You have destructive interference.
Easy way is to put an spl meter up at the listening position, play a 1k sine wave out of both speakers and adjust the two speaker positions so that the spl meter reads loudest (toe them out a little.)

Thavma 29th October 2018 01:50 PM

Quote:

Originally Posted by pentagon (Post 13598465)
You have destructive interference.
Easy way is to put an spl meter up at the listening position, play a 1k sine wave out of both speakers and adjust the two speaker positions so that the spl meter reads loudest (toe them out a little.)

Thank you!

You think its the speakers position problem or any reflections?

bert stoltenborg 29th October 2018 04:57 PM

Quote:

Originally Posted by Thavma (Post 13598997)
Thank you!

You think its the speakers position problem or any reflections?

A pic or drawing of the situation would help....kfhkh

ReDRuMx 29th October 2018 05:33 PM

If I understand correctly, when you play each speaker individually there is no dip? If so, than you most probably have no problem.

The dip is a result of the mic not being perfectly centered in between the speakers, so the signals from L and R speakers reach it with a small time difference - and that creates interference.

In any case, that measurement setup is not a perfect replication of "listening with two ears". I.e, you don't experience the dip while listening with two ears, with a (hopefully) head in between :)

MediaGary 29th October 2018 07:27 PM

Quote:

Originally Posted by Thavma (Post 13598106)
...I have a dip at 1KHz with Adam A5Xmonitors, I have measured them separately but there is no dip, they only have dip when they both play in the sweet spot.
....
Could anyone help me to identify the problem....

Arrrgghhh!

My instincts told me to run from this question, but in fact, I had a discussion about this very issue (different frequency) with my Adam A7X speakers in my well-treated studio. With that emotional baggage revealed, here's what I learned:

Your chart doesn't show *how much* of a dip is being experienced, but as things go, let's simply assume (ugh) that what you're seeing/measuring is destructive interference between two sources that are being measured at one point in a 3D space.

A 1kHz wave in the air is about 34cm (13.4 inches) long. Therefore two 1kHz waves that arrive offset by 1/2 that length (17cm, 6.7in) will have a +peak/-peak confluence and show a reduction in the net level as a result. That maximum difference is about 1/2 millisecond offset in arrival time. That's also about the distance and arrival time difference between your ears!

Just with that information, any arrival differences between the two speaker sources at 1kHz that are in increments of 0.5ms, 1.5ms, 2.5ms, etc., will theoretically subtract from one another. Increments of 1.0ms, 2.0ms, 3.0ms, etc., will theoretically constructively add as coherent sources and add their power together.

The "theoretically" weasel word is in there because in the context of music in a real-world room, we're experiencing reflections of complex waveshapes from the walls, ceiling, table/console, and floor, arriving at different angles, different times and varying strengths; in addition to the direct nearfield signals coming to our two ears, all to be managed and interpreted by the incredibly refined audio signal processing wetware in our brains.

A variety of lessons learned have been documented by moving a measurement microphone by 2 inches and seeing the rather drastic differences that are captured by that small physical offset. It serves as a cautionary tale for those who shift their office chair and make recursive EQ decisions... but I digress.

One of the many bottom lines of what I'm offering to you is to simply ensure that you've done the due diligence of managing the most egregious reflections in your listening position, and then dial in your perceptive compensations to ensure that your mix decisions translate to the outside world of earbuds, cars, and the Grammy Awards' satellite feed.

bert stoltenborg 29th October 2018 08:29 PM

Wouldn't an out of phase situation tend to create a combfilter, not a discrete dip?

ReDRuMx 29th October 2018 08:37 PM

1 Attachment(s)
Quote:

Originally Posted by bert stoltenborg (Post 13599688)
Wouldn't an out of phase situation tend to create a combfilter, not a discrete dip?

Impossible to see from the graph OP attached if there is indeed a comb filter happening or not...

norfolk martin 29th October 2018 08:59 PM

Quote:

Originally Posted by MediaGary (Post 13599580)
Arrrgghhh!


A 1kHz wave in the air is about 34cm (13.4 inches) long. Therefore two 1kHz waves that arrive offset by 1/2 that length (17cm, 6.7in) will have a +peak/-peak confluence and show a reduction in the net level as a result. That maximum difference is about 1/2 millisecond offset in arrival time. That's also about the distance and arrival time difference between your ears!

Just with that information, any arrival differences between the two speaker sources at 1kHz that are in increments of 0.5ms, 1.5ms, 2.5ms, etc., will theoretically subtract from one another. Increments of 1.0ms, 2.0ms, 3.0ms, etc., will theoretically constructively add as coherent sources and add their power together.

Nice analysis. As you point out later, the 1/2 wavelength difference between one path and the other won't provide full cancellation at 1KHz because of all the alternate paths, but it can be significant.

IMO, if each monitor is reasonably flat at the specified distance, trying to measure them together is asking for cancellation/reinforcement anomalies.

bert stoltenborg 29th October 2018 09:24 PM

Quote:

Originally Posted by norfolk martin (Post 13599783)
Nice analysis. As you point out later, the 1/2 wavelength difference between one path and the other won't provide full cancellation at 1KHz because of all the alternate paths, but it can be significant.

IMO, if each monitor is reasonably flat at the specified distance, trying to measure them together is asking for cancellation/reinforcement anomalies.

Not in the low frequencies. But to inspect the mid/high freqs I would measure them individualy and overlay the measurements to see if there are anomalies.

bert stoltenborg 29th October 2018 09:25 PM

Quote:

Originally Posted by norfolk martin (Post 13599783)
Nice analysis. As you point out later, the 1/2 wavelength difference between one path and the other won't provide full cancellation at 1KHz because of all the alternate paths, but it can be significant.

IMO, if each monitor is reasonably flat at the specified distance, trying to measure them together is asking for cancellation/reinforcement anomalies.

Not in the low frequencies. But to inspect the mid/high freqs I would measure them individualy and overlay the measurements to see if there are anomalies.

ReDRuMx 29th October 2018 09:45 PM

Quote:

Originally Posted by MediaGary (Post 13599580)

A 1kHz wave in the air is about 34cm (13.4 inches) long. Therefore two 1kHz waves that arrive offset by 1/2 that length (17cm, 6.7in) will have a +peak/-peak confluence and show a reduction in the net level as a result. That maximum difference is about 1/2 millisecond offset in arrival time. That's also about the distance and arrival time difference between your ears!

Just with that information, any arrival differences between the two speaker sources at 1kHz that are in increments of 0.5ms, 1.5ms, 2.5ms, etc., will theoretically subtract from one another. Increments of 1.0ms, 2.0ms, 3.0ms, etc., will theoretically constructively add as coherent sources and add their power together.

The "theoretically" weasel word is in there because in the context of music in a real-world room, we're experiencing reflections of complex waveshapes from the walls, ceiling, table/console, and floor, arriving at different angles, different times and varying strengths; in addition to the direct nearfield signals coming to our two ears, all to be managed and interpreted by the incredibly refined audio signal processing wetware in our brains.

BTW, perhaps I'm misunderstanding you but...

1. If the distance between the ears is 15 cm, that doesn't mean that the sound from one speaker reaches one ear "15 cm earlier" than the other ear, because it is arriving at an angle. The delay will be shorter.

2. Same sound reaching one ear later than the other (say 0.5 ms) will not result in a perceived dip at the "expected" 1 kHz. Your brain is smarter than a microphone. Try it with headphones.

This is why it is usually suggested to measure each speaker individually, and not worry that much about their sum at one MONO location for high frequencies.

Thavma 29th October 2018 09:53 PM

Thanks guys, its not comb filtering, or is it?

here is the full graph

https://image.ibb.co/f1dp0q/graph.jpg

I will try various suggestions

Someone else told me to flip the phase ot one tweter since its bi-amped, lol?

ReDRuMx 29th October 2018 09:57 PM

I haven't used Sonarworks, but this to me looks like a dip on both the L and the R measurement.

Quote:

Originally Posted by Thavma (Post 13599881)

Someone else told me to flip the phase ot one tweter since its bi-amped, lol?

What, why? No.

bert stoltenborg 29th October 2018 10:14 PM

Quote:

Originally Posted by ReDRuMx (Post 13599885)
I haven't used Sonarworks, but this to me looks like a dip on both the L and the R measurement.



What, why? No.

Zjeeezz, is this the issue?

norfolk martin 30th October 2018 12:58 AM

Quote:

Originally Posted by ReDRuMx (Post 13599871)
BTW, perhaps I'm misunderstanding you but...

1. If the distance between the ears is 15 cm, that doesn't mean that the sound from one speaker reaches one ear "15 cm earlier" than the other ear, because it is arriving at an angle. The delay will be shorter.

2. Same sound reaching one ear later than the other (say 0.5 ms) will not result in a perceived dip at the "expected" 1 kHz. Your brain is smarter than a microphone. Try it with headphones.

This is why it is usually suggested to measure each speaker individually, and not worry that much about their sum at one MONO location for high frequencies.

jummpp But if you only had one ear, and walked across the stereo field...

dinococcus 30th October 2018 12:15 PM

Quote:

Originally Posted by Thavma (Post 13598106)
Hello guys,

I have a dip at 1KHz with Adam A5Xmonitors, I have measured them separately but there is no dip, they only have dip when they both play in the sweet spot.

hello,

You measure the two speakers in the same time and you have a deep ?

there is nothing wrong. The measurement with the two speakers in function are valid only for the bass. You generate destructive interferences with the two speakers in function.

The measurements of the frequency response must be take with caution.

When you sit in the place of the microphone, the measurement of the frequency response changes dramatically in the medium and high frequencies up to a longer wavelength of your body.

IMHO and With my experience, the interest of the frequency response becomes low and misleading for the mid and high frequencies.

MediaGary 30th October 2018 02:27 PM

Quote:

Originally Posted by ReDRuMx (Post 13599871)
BTW, perhaps I'm misunderstanding you but...

1. If the distance between the ears is 15 cm, that doesn't mean that the sound from one speaker reaches one ear "15 cm earlier" than the other ear, because it is arriving at an angle. The delay will be shorter.

2. Same sound reaching one ear later than the other (say 0.5 ms) will not result in a perceived dip at the "expected" 1 kHz. Your brain is smarter than a microphone. Try it with headphones.

This is why it is usually suggested to measure each speaker individually, and not worry that much about their sum at one MONO location for high frequencies.

Yup, I agree! I think we're on the same page.

I has hoping to achieve the most pithy assessment of what's going on in a stereo monitor situation; hence my trepidation about getting into this topic at all. I guess I can't cover all the implications in so few words. Here are some more words.

The speaker measurements in the OP's test are being done with a microphone at a single point in 3D space. I was clear about that in my append. That single mic has the effect of summing the contribution of the two speakers at a single point. Moving the single microphone just a little (like the 2-inches I mentioned) does in fact change the measured response. To me, it's daunting that such a small movement can create such differing results.

I alluded (not so clearly, sorry) to the "Two-Ear Problem". (It's not an official term, but TEP sounds kinda cool as an acronym.) With the TEP, we're catching two speaker sources at two different points in 3D space. It won't be the same 'dip' in each ear because the arrival of the indirect sound in each ear is different. The 3D position of the ears has a fixed offset between them, but each ear is having a different experience.

The wetware signal processing in our brain does a marvelous job of real-time calculation of relative strength, timing offsets, reflection interpretation, and EQ. I am always blown away by the joy of riding in my convertible through a wooded area. The fact that the 3D space of the bird calls occurring above, in front, in back, left/right, and at widely varying of distances can be readily ascertained in a moving vehicle is a fantastic testament to how good the system is.

The analysis gets to be a rather different problem at longer wavelengths as the speaker itself starts putting out a more widely dispersed signal, and the room interactions are both more dramatic and harder to absorb/control. I'm running for the hills before this discussion drifts in to the world of bass in a room. :)