Frequency response measurements - Interpretations

[Feyire]

What follows are my frequency response measurements using the miniDSP UMIK-1‎ and REW on my Pioneer S-3EX speakers in my living room. I hope some of you may find the results of interest.

Additional information:

- All measurements are taken at my listening position on the couch, which is approx 3.2 metres from each speaker
- Targeting 75 dB SPL
- The horizontal separation between the speakers is approx. 265cm
- Both speakers are toed in towards the listening position
- Both speakers are 25cm from the back wall (5cm less than the minimum Pioneer recommend)
- The microphone is at a height of 1 meter from the floor on the couch (this is the height recommended by Pioneer to match up with the tweeter), this is within 10-15cm of my ears at that sitting position
- There is a large and wide glass coffee table between the horizontal center point of the two speakers and the listening position
- The couch is L shaped and the bottom part of the L shape is in very close proximity to the left speaker

When both left and right speakers are measured simultaneously (microphone facing the center point between the two speakers) as shown above, there is a significant null(?) at approx 55Hz.

I'd appreciate any suggestions how to deal with the 55Hz problem and I'd also like to hear any thoughts/comments on the measurements.

Unpleasant observation during my testing: My SMSL SU-8 DAC commits suicide when I do measurements on only 1 channel (specifically the right channel). When it does, it shuts off output mid measurement and I hear a horrible crackle noise. Only way to continue using the right channel after that is to "reset" the device, by playing a stereo signal again. I recall hearing about such a problem in the past, perhaps related to ESS chips not liking it when you do sine sweeps on only 1 channel?

Thanks for your time guys

 

[amirm]

Unpleasant observation during my testing: My SMSL SU-8 DAC commits suicide when I do measurements on only 1 channel (specifically the right channel). When it does, it shuts off output mid measurement and I hear a horrible crackle noise. Only way to continue using the right channel after that is to "reset" the device, by playing a stereo signal again. I recall hearing about such a problem in the past, perhaps related to ESS chips not liking it when you do sine sweeps on only 1 channel?

That's correct.

 

[patate91]

What follows are my frequency response measurements using the miniDSP UMIK-1‎ and REW on my Pioneer S-3EX speakers in my living room. I hope some of you may find the results of interest.

Additional information:

- All measurements are taken at my listening position on the couch, which is approx 3.2 metres from each speaker
- Targeting 75 dB SPL
- The horizontal separation between the speakers is approx. 265cm
- Both speakers are toed in towards the listening position
- Both speakers are 25cm from the back wall (5cm less than the minimum Pioneer recommend)
- The microphone is at a height of 1 meter from the floor on the couch (this is the height recommended by Pioneer to match up with the tweeter), this is within 10-15cm of my ears at that sitting position
- There is a large and wide glass coffee table between the horizontal center point of the two speakers and the listening position
- The couch is L shaped and the bottom part of the L shape is in very close proximity to the left speaker

View attachment 68543


View attachment 68544


View attachment 68545

When both left and right speakers are measured simultaneously (microphone facing the center point between the two speakers) as shown above, there is a significant null(?) at approx 55Hz.

I'd appreciate any suggestions how to deal with the 55Hz problem and I'd also like to hear any thoughts/comments on the measurements.

Unpleasant observation during my testing: My SMSL SU-8 DAC commits suicide when I do measurements on only 1 channel (specifically the right channel). When it does, it shuts off output mid measurement and I hear a horrible crackle noise. Only way to continue using the right channel after that is to "reset" the device, by playing a stereo signal again. I recall hearing about such a problem in the past, perhaps related to ESS chips not liking it when you do sine sweeps on only 1 channel?

Thanks for your time guys

To adress 55hz dip there's few options :

Change your listening spot.
Change your speakers location.
Bass traps in corners.

Note that room modes cannot be fully avoid . 1 and 2 option will only change where the dips are.

Option 3 could improv greatly the FR, but you need to invest in mineral wool and fiddle with speakers placement to get the best result.

Note that sharp dips are less noticeable when listening to music or movies. (You can use psyhoacoustic smoothing to get an idea). Change your scale too and zoom 10hz to 200hz.

 

[Feyire]

I was able to reduce the 55Hz dip by a consistent ~5dB SPL, by moving my couch's chaise lounge so it was more out of the way of the left speaker. The shown measurement is an average of 4 other measurements of both speakers together.

I'll start moving around more furniture tomorrow and continue to measure how the frequency response changes as a result.

 

[patate91]

I would suggest to experiment with the Moving Microphone Measurements too. It will save you a lot of time.

Pink noise periodic + RTA (infinite average). Circle the microphone around your head (where it should). You can compare with your single point measurement average to show you that it works pretty well.

 

[patate91]

With a wireless keyboard it can be done alone.

 

[Feyire]

I would suggest to experiment with the Moving Microphone Measurements too. It will save you a lot of time.

Pink noise periodic + RTA (infinite average). Circle the microphone around your head (where it should). You can compare with your single point measurement average to show you that it works pretty well.

Indeed, this is a great method.

Here are the results:

It is quite similar to the previous single point measurements.

 

[QMuse]

Indeed, this is a great method.

Here are the results:

View attachment 68651

It is quite similar to the previous single point measurements.

That dip at 55Hz can't be corrected, but it looks worse than it sounds as it is relatively narrow. Real issues are circled red and they can be solved by EQ-ing to the violet target line.

 

[Feyire]

That dip at 55Hz can't be corrected, but it looks worse than it sounds as it is relatively narrow. Real issues are circled red and they can be solved by EQ-ing to the violet target line.

Curiosity got the better of me and I moved my glass coffee table that was between the speakers and the couch position I'm listening at, out of the way, so there is a clear path from speakers to the couch now.

I then re-measured to see if anything would change at <400Hz, but yeah, nothing really changed at all:

 

[Balle Clorin]

What is the distance from you/mic to the wall behind you?
Fwiw the plot seem quite normal ...

 

[QMuse]

Curiosity got the better of me and I moved my glass coffee table that was between the speakers and the couch position I'm listening at, out of the way, so there is a clear path from speakers to the couch now.

I then re-measured to see if anything would change at <400Hz, but yeah, nothing really changed at all:

View attachment 68654

Well yes, LF is not really affected by coffee table.

Where are your speakers standing? Do you have stands or they are on some kind of shelf?

 

[Feyire]

What is the distance from you/mic to the wall behind you?
Fwiw the plot seem quite normal ...

This is my room.

- The two red squares are the L and R speaker respectively.
- The orange square is the listening position
- The black square is a ceiling support

The listening position is about 350cm from the back wall.

 

[Feyire]

Where are your speakers standing? Do you have stands or they are on some kind of shelf?

They are floor standing speakers. The floor is wooden and they are on speaker feet/spikes.

 

[QMuse]

They are floor standing speakers. The floor is wooden and they are on speaker feet/spikes.

Your FR issues are normal, this kind of things happen in pretty much every room and EQ is the only effective cure.

 

[Feyire]

Your FR issues are normal, this kind of things happen in pretty much every room and EQ is the only effective cure.

I'll take a look at your suggested curve and and play around some more with the measurements and eventually start serious EQing.

One thing I have noticed though, and this has also come from my experience with my headphone listening, is that I like a linearly rising slope after 10kHz, peaking to approx. +10dB at 20kHz. My subjective experience (forgive me amirm!) here is that I perceive a widening of the sound-stage and there is more "air" around vocals. I get the same experience with speakers, but to a lesser extent. This seems to go against the notion of all these curves I see people suggest, where the majority roll-off the higher frequencies to various extents .

 

[ernestcarl]

I like a linearly rising slope after 10kHz, peaking to approx. +10dB at 20kHz.

Sorry, what?

 

[Feyire]

Sorry, what?

Yeah, I'm strange. I can hear the difference in real time by turning the filters on/off during playback. I like it when they are on

 

[ernestcarl]

Yeah, I'm strange. I can hear the difference in real time by turning the filters on/off during playback. I like it when they are on

View attachment 68681

I sort of understand if it were around half the amount of boost for the HD650 headphones, but rising earlier, since I already apply some boost there...

I cannot hear anything above 16kHz so your EQ would have no impact above that range to me. Can you still hear above that range though?

 

[Feyire]

I sort of understand if it were around half the amount of boost for the HD650 headphones, but rising earlier, since I already apply some boost there...
I cannot hear anything above 16kHz so your EQ would have no impact above that range to me. Can you still hear above that range though?

I can't hear past 15.8kHz.

I also use that Morphit plugin for my HD650, however, I do boost the higher frequencies a bit more than their "Corrected" profile. That is not to say though, that I don't like the "Corrected" profile, it sounds great as is.

 

[ernestcarl]

Under 10kHz the boost in your filter isn't that much anyway -- likely the effect is not as detrimental as I thought.