• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Why does DIRAC completely remove bass response (KEF R3 + sub + minidsp flex)

OP
G

Geertidow

Active Member
Joined
Nov 16, 2021
Messages
184
Likes
60
By the way slight repositioning helps that 80 hz dip when the xo is at 80 hz.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,446
Likes
7,955
Location
Brussels, Belgium
Looks like there is quite the messiness still with the crossover. But imma ignore that for a while.

Imma ask you to tune the gain of these filters using pink noise from the listening position, apply only negative gain, increase it till the offending frequencies disappear.

1671040451008.png


I noticed there are some very early reflections happening at these frequencies

1671040525893.png
 
OP
G

Geertidow

Active Member
Joined
Nov 16, 2021
Messages
184
Likes
60
Do you mean the 37 and the 115 hz? Thanks
Edit: I see now thee 646, 438 and 921 hz.
I will make time for this as soon as I can, but I think it'll be saturday!
@abdo123 these reflections are not really fixed at the LP?
 
Last edited:
OP
G

Geertidow

Active Member
Joined
Nov 16, 2021
Messages
184
Likes
60
Oke so in the new position I have run the same sweep at the preferred speaker positioning.
Can you recheck what RTA corrections I need to make? I assume the frequencies will be different
 

Attachments

  • 80hz_xo_samePEQ.zip
    3.5 MB · Views: 44

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,446
Likes
7,955
Location
Brussels, Belgium
This is what I do to check early reflections. I use frequency dependent windowing (in IR Windows), 10 cycles for 100Hz to 300Hz and 5 cycles for 300Hz and up.

Only use negative gain filters to treat areas strategically, always tune the gain by listening to pink noise. Under 100Hz what we measure is more or less what we hear, but above that that's not the case anymore.

You can also measure different listening positions and see which ones offer the least amount of peaks and dips. This makes a DRAMATIC difference in sound quality.

1671122098536.png


1671122054963.png


1671122065317.png
 

Attachments

  • 1671122077616.png
    1671122077616.png
    69.1 KB · Views: 30
OP
G

Geertidow

Active Member
Joined
Nov 16, 2021
Messages
184
Likes
60
That 100 hz dip is interesting since the speaker are symmetrical in the room. I will look for positioning with the best results and then upload a measurement.
 
OP
G

Geertidow

Active Member
Joined
Nov 16, 2021
Messages
184
Likes
60
Dips are not really a big deal but in my experience but peaks are literal distortion.
If I remove a cabinet that 100 hz peak is mostly removed but I still have to check if I can actually head it. I used the trimming of peaks up to about 900 hz.

However, does DIRAC do this for me or is it less optimal?
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,446
Likes
7,955
Location
Brussels, Belgium
If I remove a cabinet that 100 hz peak is mostly removed but I still have to check if I can actually head it. I used the trimming of peaks up to about 900 hz.

However, does DIRAC do this for me or is it less optimal?

Dirac only adjusts (besides the phase correction) the average amplitude response of all your measurements to the target curve you provide.

However, I don’t think these measurements represent how people actually hear above 100Hz, Dirac does miracles when it comes to the subbass but response wise above 100Hz it does more harm than good with good speakers.

You need to surgically treat issues caused by the location of the speaker in the room, and leave the issues caused by the location of the listener alone.
 
Last edited:
OP
G

Geertidow

Active Member
Joined
Nov 16, 2021
Messages
184
Likes
60
Dirac only adjusts (besides the phase correction) the average amplitude response of all your measurements to the target curve you provide.

However, I don’t think these measurements represent how people actually hear above 100Hz, Dirac does miracles when it comes to the subbass but response wise above 100Hz it does more harm than good with good speakers.

You need to surgically treat issues caused by the location of the speaker in the room, and leave the issues caused by the location of the listener alone.
And can you say anything about the 80 hz crossover?
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,446
Likes
7,955
Location
Brussels, Belgium
And can you say anything about the 80 hz crossover?

the 100Hz crossover seem to perform the best out of the measurements you send me, if you can somehow solve that 100Hz dip then that's even better.

plug the port of the speakers if you want to use crossover at 100Hz. 12db/oct BW high-pass and 24db/oct LR low-pass should yield a perfect LR 24db/oct crossover at 100Hz with the port plugged.
 
D

Deleted member 48726

Guest
Dirac only adjusts (besides the phase correction) the average amplitude response of all your measurements to the target curve you provide.

However, I don’t think these measurements represent how people actually hear above 100Hz, Dirac does miracles when it comes to the subbass but response wise above 100Hz it does more harm than good with good speakers.

You need to surgically treat issues caused by the location of the speaker in the room, and leave the issues caused by the location of the listener alone.

What do you base those claim of?
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,446
Likes
7,955
Location
Brussels, Belgium
What do you base those claim of?

I remember it boosting an SBIR null and making it sound pretty awful, I learned a lot during the last year and a half. Dirac support helped me by advising me to adjust the target curve not to boost that dip.

 
Last edited:
D

Deleted member 48726

Guest
I remember it boosting an SBIR null and making it sound pretty awful, I learned a lot during the last year and a half.

So it's your own experience with Dirac and not general.
 
D

Deleted member 48726

Guest
If you think so.

You're obviously not going to find an article from Dirac advising people not to use their products with good performing speakers.
No, not if I think so. It is, in fact, your own single experience you use as a generalization. -Isn't it?

Good performing speakers also benefit from EQ when placed in a room as the room color the sound. A speaker with a flat response (anechoic) will probably respond better to EQ than a speaker with wildly varying response. So if anything I would say that a good speaker is better to EQ than a "bad" one. Bad here meaning a speaker with a response that's all over the place.
 
D

Deleted member 48726

Guest
You can find a lot of sources on the internet from people with much more higher qualifications than I am saying you should not EQ a speaker with a flat on-axis response and smooth directivity above the transition frequency. Something that DIRAC does.

You don't have to listen to me per say.
Why would you not want to correct for room response and what are the negative effect you (and lots of sources on the internet) claim?
 

olegtern

Member
Joined
Sep 23, 2020
Messages
87
Likes
100
You can find a lot of sources on the internet from people with much more higher qualifications than I am saying you should not EQ a speaker with a flat on-axis response and smooth directivity above the transition frequency. Something that DIRAC does.
Dirac use a massive of 9 (as I remember) measurements, and if they are done correctly from many random spots in the listening area, a speaker with a flat on-axis response and smooth directivity would measure flat (with some HF decline, but Dirac also has decline in target curve).
 
OP
G

Geertidow

Active Member
Joined
Nov 16, 2021
Messages
184
Likes
60
O god, that DIRAC bug is back again.
I really feel like destroying this MiniDSP flex, I just want to measure in DIRAC and enjoy th emusic!
 

Attachments

  • dirac_error.jpg
    dirac_error.jpg
    125.4 KB · Views: 81
Top Bottom