- Joined
- Jan 15, 2020
- Messages
- 6,873
- Likes
- 16,838
It isn't though above 200 Hz see post #685.
-
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!
Share your in-room measurements?
- Thread starter thefsb
- Start date
I believe you, it's just uncanny. I've seen you measurements in other threads and always assumed they were corrected to 1khz.
It's more of a testament to how great the LS50Ms are!
My LS 50 Metas + single SB 2000 crossed over at 100 Hz with PEQ and Dirac Live. Magenta is PIR with full Dirac Correction. Green is uncorrected measurement.
Last edited:
Hi, i am not familiar with dirac live, and i see it does two things that are usually not recommended, at least to newbie humans:
- fill dips (that one at 100 hz)
- eq the full band
Is this ok or because dirac is so smart or should we prevent it to do it?
Thanks for your insight!
Now that’s how a constant directivity speaker ‘should’ measure imo @thewas.
There is barely any correction happening above 200Hz anyway so it shouldn’t be a big deal.
The boost at 100Hz is a little on the risky side, but since it’s handled 50\50 by the sub and the speaker it might sound okay. Usually Dirac eliminates all microphone position related dips (when it averages out all the impulse responses) so this is likely to be SBIR.
Edit: wait a minute that’s a crossover dip!
However @phoenixdogfan you should probably limit the correction to below 10Khz because I wouldn’t trust the directivity of the mic at that point, but it probably won’t make a big difference anyway. Are the speakers roughly 3.5-4 meters away from the wall behind them?
No about 30". The correction is so small above 300hz, I just let it ride. Suppose I could have put a curtain there, but I didn't see the need. Also, it's crossed over at 100 hz with a Linkwitz Reilly, so given I'm working with a 600 wpc amp in the SB2000, and a 400 wpc amp (Purifi Eval1) in the Metas, I think it can handle the dip, and there's only really one big one anyway. It does sound really good, btw.
- Joined
- Jun 5, 2020
- Messages
- 4,800
- Likes
- 3,744
That's an almost 10 dB dip though which will kill headroom. Fine if you don't crank things too loud, of course.
It really looks like a crossover dip though, it’s better to get it sorted out.
The metas have very little headroom at 100Hz too.
Did you check the acoustic alignment of the 100 Hz x-over? If the filter is a Linkwitz-Reilly the acoustic result could be completely different.
ernestcarl
Major Contributor
I would appreciate it you post a link of the before and after correction mdat file of your measurements (same EQ from latest report) -- if you have REW... It would be interesting to compare the latest report to a third party software like REW. I am aware that Genelec uses 1/12 octave smoothing below 250 Hz and 1/6 above. However, I am curious how similar their early and late frequency response filtering compares with REW's windowing and/or smoothing.
- Joined
- Jan 15, 2020
- Messages
- 6,873
- Likes
- 16,838
As said the LS50 is constant directivity only above 3 kHz. Also the original LS50 has not as high neutrality on the LW and on the directivity so that example doesn't tell much.
Also as I have repeatedly written that LP response depends on the room reverberation function and the listening distance.
For example the room of John Atkinson loudspeakers like the LS50 give also a flat response till some frequency, just in his room the the kink is in a higher frequency, approximately flat till 5 kHz:
Fig.7 KEF LS50 Meta, spatially averaged, 1/6-octave response in JA's listening room (red), and of original LS50 (blue).
Source: https://www.audiosciencereview.com/forum/index.php?threads/kef-ls50-bookshelf-speaker-review.11144/
That's also the reason why the correction above room transition frequency to some predefined targets usually doesn't work well, its better to EQ in that region based on anechoical measurements.
As said the LS50 is constant directivity only above 3 kHz. Also the original LS50 has not as high neutrality on the LW and on the directivity so that example doesn't tell much.
Also as I have repeatedly written that LP response depends on the room reverberation function and the listening distance.
For example the room of John Atkinson loudspeakers like the LS50 give also a flat response till some frequency, just in his room the the kink is in a higher frequency, approximately flat till 5 kHz:
Fig.7 KEF LS50 Meta, spatially averaged, 1/6-octave response in JA's listening room (red), and of original LS50 (blue).
Source: https://www.audiosciencereview.com/forum/index.php?threads/kef-ls50-bookshelf-speaker-review.11144/
That's also the reason why the correction above room transition frequency to some predefined targets usually doesn't work well, its better to EQ in that region based on anechoical measurements.
Thank you for the response. But please don’t take this personally in anyway (feeling some bad juju from the tone of the post), it just (understandably) stood out to me.
Even in JK’s measurements there is no 6dB nose dive at 1KHz but it starts rather at 6KHz which can be more reasonably justified.
- Joined
- Jan 15, 2020
- Messages
- 6,873
- Likes
- 16,838
Don't worry, nothing personal, I just explain different room responses which I have seen in the now more 10 years I experiment on DRC.
Don't really see why that should be more justified and the example shows also how strongly it depends on the room acoustics, Anyway, we all agree that above transition frequency equalisation should be done based on anechoic measurements so it is not really an issue. Also below unfortunately the target not only depends on the room but also on taste/recordings, due to the audio's circle of confusion unfortunately, like Toole says with a good relatively neutral basis someone can adjust bass level according to the recording.
ernestcarl
Major Contributor
Well, the stuff above the transition or HF is largely irrelevant if you're not equalizing it -- or when there's some gross unexpected non-linearity that needs to be figured out. These are all in-room measurements -- and not even quasi-anechoic axial responses -- where there's always going to be any number of unknown variables... so some leeway in judgment of what is truly correct or optimal should be applied (esp. when speaking of conformity to any so-called "target").
So, if I wanted to address that, probably go with something like a 12db per octave high pass for the Metas?
After adding 2 huge towers of absorbers (2+2 pieces of 64 x 64 x 124 cm each, 32 kg/m3 density rock wool) on the corners, they really did wonders on the bass decay. Now the room is sounding very nice!
Before, with 14 panels of rock wool absorbing >150 Hz, but huge bass decay untouched.
After moving the speakers a little to tame a peak, before EQ (so ignore the frequency response itself), just look how the decay on the bass is a lot lower.
Last edited:
- Joined
- Jun 5, 2020
- Messages
- 4,800
- Likes
- 3,744
Indeed that is some massive resonance in the bass. It's normal to have some in the really low frequencies, but not so broad:
This room is small and has no absorbers other than a few recliners and the carbon lifeform that sometimes inhabits it.
This room is small and has no absorbers other than a few recliners and the carbon lifeform that sometimes inhabits it.
Isn't there a 5-6 dB acoustic slope between the first 100-50 Hz and then around 20 dB for the octave between 50-25 dB?
An x-over 24 dB/octave for the low-pass (bass) should be 18 dB/octave for high-pass to yield 24 dB acoustic for both, at least during one octave around the x-over point.
Similar threads
