Topping PA5 & E50 + Bowers Wilkins 805 & Complex room (+RePhase, A07, Rotel, Up/Front, Toe-in)

[tiago.trindade]

Hello everyone,

Finally got my UMIK and started to use it with REW and CamillaDSP to get some convolution filters to correct my speakers/room inefficiencies.
The results have amazed me (in a positive way!) and I have some questions for you (and also some conclusion to share):

A few introduction notes:

• This room is our living room (~35 square meters), but it is an open space, shared with the kitchen), and there are some individual sofas in between the speakers and the listening position (I shared it with my wife and 3 kids, so fewer options to "change the way things are located")
  
  
  
• The speakers are a little bit high (on top of some furniture) - the tweeter is at 125 cm from the floor
• Each speaker is at the corner of the living room space (370 cm between each other) - with additional resonances.
• The listening position (sofa) is not equilateral triangle regarding the speakers (400 cm from the middle of speakers line until the listening position)
• These 805N speakers have a "flat" anechoic response (previous B&W philosophy) as measured by @John Atkinson at Stereophile https://www.stereophile.com/content/bw-nautilus-805-loudspeaker-measurements-part-2
  
  
  
  
  
  
  
• All measurements were done in the same position (single sweep) - i.e. did not move the UMIK - and using the calibration files (as recommended here )
• The DAC+Amplifier should be pretty flat and not influence the response of these results (I also tested it, with the same setup, the Aiyima A07 and the Rotel RA-1062 I also own)

The main problem I have (and would like your help with ) is the fact that I feel this type of setup/placement is rather diffuse in terms of soundstage and imaging.
If I use the speakers in near-field (~1m distance, I feel much more "immersed" into the music - almost like using headphones). I previously had the speakers in front of the TV furniture (1,5m apart) and my experience (sitting on the couch) was not immersive at all (but the sound was less diffuse). I prefer the way it is now (I now feel the existence of a "soundstage" but is less defined).

Different measurements were made (and I uploaded the .mdat file here for those who want to dig in ):

1. No filters (baseline)
2. With EQ/filters (with and without RePhase)
3. Time Align vs No Delay
4. Mic pointing front vs pointing up
5. PA5 vs Aiyima A07 vs Rotel RA-1062
6. No toe-in vs 45 degrees toe-in

So let's start!

1. Baseline (no-EQ) (green) vs EQ (Harmon 20hz-20khz) (orange)

1.1. Baseline (no-EQ) Left (green) vs Baseline (no-EQ) Right (blue)

2. Time Align/Delay Baseline (green) vs NO Time Align/Delay (purple)

3. EQ (Harmon 20hz-20khz) (orange) vs RePhase (red)

4. Mic front (dark purple) vs Mic Up (light purple)

5. Toe-in 35 degrees (red) vs NO Toe-in (orange)

6. Topping PA5 (purple) vs AIYIMA A07 (green)

6. Rotel RA-1062 (light green) vs AIYIMA A07 (dark green)

Conclusions:

• Introducion time align/delay improves the "experience" (more focused imaging) but changes the measurements. I cannot feel the voice in the center of me (even if I move within the sofa) without delay/time align. So it's a must for me.
• EQ turns the bass much more enjoyable and "tight/fast" and brings more focus to the singer.
• Using RePhase (I don't know if I am doing all the correct steps) "cleans" the sound and improves the resolution/clarity of the soundstage, but also changes the measurements
• The measurements with the mic at 0 or 90º (even with calibration files) present different measurements
• When toed-in, the "sweet spot" advances (from the sofa into the speakers wall), but starts to be very "narrow". The sound is also different (more "spacial" but not necessarily better).
• When toed-in, the dip in the cross-over area also deepens (as shown on the Stereophile measurements)
• PA5 and Aiyima A07 present a very similar response. Rotel has some differences at 100hz

Questions:
1. What can I do to improve this?
2. Should I point the mic to the ceiling or into the front of the listening position? (I get better sonic results if EQ based on the front measurements)
3. I am doing an individual measurement of each speaker, and applying a REW EQ filter for each of them (Harman target), with the same SPL target defined on REW (even though each speaker measures a different SPL at the listening position). Is this the right way to do it?
4. REW did not apply any filter in the 50hz, 120hz nor the 2500hz area. I understand that sharp dips should not be corrected, but should I force it/add it manually?
5. My previous experience between REW and Dirac Live (after EQ applied) was not much different. Do you think I should try it here again?
6. On RePhase, what I am doing is, for each channel, just 1) Go to "Tools" and import the XML generated EQ filters from REW 2) change to "linear-phase" in RePhase filters area 3) select "moderate" optimization and 48khz rate and 4) select generate with 32 IEEE mono .wav. I am missing anything?

Congratulations if you read everything until here!

Thank you for any tips

 

[tiago.trindade]

A few updates!

1. I am now able to have a "flat" curve (Harman) from 400khz-20khz.
2. The dip around the crossover (3khz) is now "corrected", but I have two huge dips at 200hz and 300hz
3. I was using rePhase in a wrong way. I should have selected "minimum phase" instead of "linear phase", and insert the cross-over on the "minimum phase filters" tab

1. old-EQ (orange) vs new EQ (red)

2. Right-EQ (brown) vs Left-EQ (blue)

3. Left (blue/green) and Right (brown/red) (EQ vs non-EQ)

The dips at 200hz and 300hz should be created by the room right? I see that the 200hz dips exists on both individual speakers, but the 300hz does not...

 

[Marc v E]

I must say that fq looks pretty good to me.

However, I think you can improve on what you have now. It looks like the speakers are each placed in the corner of 2 walls.

Ime you will get better imaging if you place the speakers on a speaker stand, somewhat away from the corners and not directly placed on a flat surface of tv furniture. That way you have less unwanted reflections.
Iirc B&W had some nice matching stands too.

Another point I would put my focus on is the listening position on the couch. I would try to minimize the reflections from the back wall.

 

[tiago.trindade]

I must say that fq looks pretty good to me.

However, I think you can improve on what you have now. It looks like the speakers are each placed in the corner of 2 walls.
View attachment 256078
Ime you will get better imaging if you place the speakers on a speaker stand, somewhat away from the corners and not directly placed on a flat surface of tv furniture. That way you have less unwanted reflections.
Iirc B&W had some nice matching stands too.

Another point I would put my focus on is the listening position on the couch. I would try to minimize the reflections from the back wall.

Great input regarding the listening position! I have never tought about reflection from the back wall.

Due to WAF (and 3 small kids running in the house), I can only move the speakers slightly along the TV wall.

What is very strange to me, is how different the sonic signature is from my two EQ/curves:

The last one (red) is much more focused (voice right on the center) and wide, but while less detail and more "air". The older one (orange) is more thin and with higher detail (but less body and "air")

If I use 24/96 tracks, I need to have a convultion filter (.wav) in 96khz rate? or I can keep the current 48khz I have now (processed by RePhase)?

 

[Marc v E]

Great input regarding the listening position! I have never tought about reflection from the back wall.

Due to WAF (and 3 small kids running in the house), I can only move the speakers slightly along the TV wall.

I understand. Much the same in my home.
I would then make sure the reflection survices are as small as possible and visually acceptable.

What is very strange to me, is how different the sonic signature is from my two EQ/curves:

The last one (red) is much more focused (voice right on the center) and wide, but while less detail and more "air". The older one (orange) is more thin and with higher detail (but less body and "air")

I'm can't directly correlate the frequency with your discription of air and detail. But I can say that a centered voice ime is a good start. The red curve from 700hz upwards to 20hz looks good to me. You have some dips lower than 700 hz that could be adressed with eq or subwoofers.

If I use 24/96 tracks, I need to have a convultion filter (.wav) in 96khz rate? or I can keep the current 48khz I have now (processed by RePhase)?

This is a good question for someone with more knowledge than I on this subject.

 

[tiago.trindade]

Thank you!

I'm can't directly correlate the frequency with your discription of air and detail. But I can say that a centered voice ime is a good start. The red curve from 700hz upwards to 20hz looks good to me. You have some dips lower than 700 hz that could be adressed with eq or subwoofers.

Don't subwoofers usually have a cut-out at around 100hz? I am "happy" with the response from 30h-120hz (the problem seems to be after that - 120hz until 400hz )

 

[Marc v E]

Correct. Can you correct the dips between 120 and 400 hz by eq?

 

[Dimifoot]

The 805N is not a far-field speaker. 4+ meters is too far away.

 

[tiago.trindade]

Correct. Can you correct the dips between 120 and 400 hz by eq?

Those measurements are already "eq" from 20hz-20khz

 

[tiago.trindade]

The 805N is not a far-field speaker. 4+ meters is too far away.

That can explain what I feel (like headphones when listening very close ~1meter; but with some lack of sharpeness/focus when listening away).

Why do you think that about this speaker? Are all bookshelfs not "far-field"?

I ask that because:
1) the measurements, with eq, "are not bad" (but I might be missing something)
2) and, with this eq , I feel the voice isolated in the "middle" of the room and there is some sense of soundstage (but feel that there is some "extra" space/air between the speakers)

 

[tiago.trindade]

A few updates!

I got a Dirac trial, and did a REW vs Dirac comparision

1. It's impressive how DIRAC was able to get the bass "stable" starting at 28hz!
2. The sound from DIRAC is more detailed (we can notice here that there is more "treble") and heterogeneous, but I get a better center imaging with REW.
3. The delay and gain detected from DIRAC are different from what I get with REW
4. I cannot understand why the time alignment / delay has such an impact in the frequency response (and listening experience) - we are talking about 0.2ms different with huge impact

1. DiRAC-EQ (red) vs REW-EQ (blue)

2. REW-EQ 0.5ms delay (green) vs REW-EQ 0.3ms delay (blue)

I am curious if will get different results using MMM method!

I am currently using this playlist to test the phantom center imaging/ soundstage and the differences between REW vs DIRAC are noticeable: https://open.spotify.com/playlist/2pcwQpUAGQcHlEBsWMQHRP?si=f2266fa5198c45c3

 

[Dimifoot]

Are all bookshelfs not "far-field"?

Exactly.
The 805N is OK up to 2m distance.
The size of the woofer is important.