• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Neumann KH310 in a small room (+subs, MA1, absorbers, AVAA)

oivavoi

Major Contributor
Forum Donor
Joined
Jan 12, 2017
Messages
1,651
Likes
1,648
Location
Oslo, Norway
You may be thinking of the depth required for resistive absorption to address bass frequencies. Toole discussing a RPG QRD: "At 7.9 in. (0.3 m) deep and exhibiting useful diffusion down to about 300 Hz, they appear to meet the requirements for wideband diffusers"

Geddes: "It appears then that the best rooms for serious listening would have a large amount of low frequency absorption accompanied by very low amounts of high frequency absorption... the directional source can be pointed in such a way that the first reflection actually arrives at the ear opposite the direct arrival. For the directional source, the secondary (after the first) reflection arrivals are virtually all lateral and behind the listener – a good thing." His solution is to have highly directional (90 degree coverage pattern) speakers toed-in severely to cross significantly in front of the listening position so that the first lateral reflection is actually contralateral and consequently later, not early.

Thanks a lot for clarifying!
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
1,712
Likes
1,203
Location
Canada
That's interesting, i could try this quite easily with my secondary KH120 and my RME UCX.
How exactly do you upmix? R to R surround and L to L surround, but less dB? How much less? Or a bit of L to R surround and vice versa? Do you add delay to simulate a reflection? Could it be beneficial not to place them in the back but on the sides?

Well, dunno if it's gonna be easy... depends on how one goes about it.

I myself am just using JRiver's native upmixer. I've read some people say it's similar to the old Dolby Pro Logic IIx algorithm, though I cannot be certain. I use it to upmix stereo LR content to 5.1 and 7.1 -- however, minus the center channel as I have no space for it.

My very vague memory of what is does (could be wrong in some of this!):

Copies LR signals in a lower level to the surrounds and inverts the IR of one of these for each surround. I believe the LR correlated signals will cancel each other out. Each surround channel produces a new decorrelated signal -- when both are summed acoustically (measured through a microphone), they (surrounds) also cancel each other out additionally lowering the level.

How it sounds for the side surrounds (20ms delayed) is a little bit different from the rears (30ms delayed). I'm not sure how it does this exactly, but it seems vocal ambience is a little clearer from the rears and instruments a bit more clear to the sides -- or that could purely be from the differences in frequency response and spectral decay characteristic (inevitably malformed by the room) when it finally arrives at my ears.



1635169433943.png


In my sofa setup (2.1/5.1ch), I have the surrounds maybe at a ~110-15 degrees angle from the front central axis.

For my desk setup (2.1/5.1/7.1ch), I have the side surrounds less than 5 degrees directly to the left and right from the desk chair MLP. The rears are maybe ~135 degrees located in the back corners of the room.

You will have to experiment to find out whatever positioning works best...

With regard to the sound levels, I have the surrounds EQ'd and toned down to be similar yet slightly lower than the front mains -- listening to pink noise helps in the matching process here -- but then, I also have additional saved presets in JRiver which allow me to change the focus of the sound:

1635169733973.png


Manual time and phase alignment is a challenge due to my mixing and matching of different speakers (ported and sealed) of different designs. Ultimately, I found it easier if I am able to mix both FIR and IIR EQs and xo with DSP -- since it's also desirable to have the phases from all loudspeakers sum coherently: wherein one is also able to vector average (time and magnitude sum) all channels, including the sub-LFE, with as little cancellation artifacts as possible.

1635201142362.png

edit: adjustments in phase to better match sub-LFE channel by changing the xo slope and freq.

This may be faster/easier to accomplish with the help of more advanced auto-calibration and full bass optimization features found in certain AVR processors.
 
Last edited:
OP
DjBonoBobo

DjBonoBobo

Senior Member
Joined
Jan 21, 2020
Messages
413
Likes
621
Location
any germ
Well, dunno if it's gonna be easy... depends on how one goes about it.

I myself am just using JRiver's native upmixer. I've read some people say it's similar to the old Dolby Pro Logic IIx algorithm, though I cannot be certain. I use it to upmix stereo LR content to 5.1 and 7.1 -- however, minus the center channel as I have no space for it.

My very vague memory of what is does (could be wrong in some of this!):

Copies LR signals in a lower level to the surrounds and inverts the IR of one of these for each surround. I believe the LR correlated signals will cancel each other out. Each surround channel produces a new decorrelated signal -- when both are summed acoustically (measured through a microphone), they (surrounds) also cancel each other out additionally lowering the level.

How it sounds for the side surrounds (20ms delayed) is a little bit different from the rears (30ms delayed). I'm not sure how it does this exactly, but it seems vocal ambience is a little clearer from the rears and instruments a bit more clear to the sides -- or that could purely be from the differences in frequency response and spectral decay characteristic (inevitably malformed by the room) when it finally arrives at my ears.



View attachment 161350

In my sofa setup (2.1/5.1ch), I have the surrounds maybe at a ~110-15 degrees angle from the front central axis.

For my desk setup (2.1/5.1/7.1ch), I have the side surrounds less than 5 degrees directly to the left and right from the desk chair MLP. The rears are maybe ~135 degrees located in the back corners of the room.

You will have to experiment to find out whatever positioning works best...

With regard to the sound levels, I have the surrounds EQ'd and toned down to be similar yet slightly lower than the front mains -- listening to pink noise helps in the matching process here -- but then, I also have additional saved presets in JRiver which allow me to change the focus of the sound:

View attachment 161353

Manual time and phase alignment is a challenge due to my mixing and matching of different speakers (ported and sealed) of different designs. Ultimately, I found it easier if I am able to mix both FIR and IIR EQs and xo with DSP -- since it's also desirable to have the phases from all loudspeakers sum coherently: wherein one is also able to vector average (time and magnitude sum) all channels, including the sub-LFE, with as little cancellation artifacts as possible.

View attachment 161359


This may be faster/easier to accomplish with the help of more advanced auto-calibration and full bass optimization features found in certain AVR processors.

Thank you very much for your detailed answer. I will think about it, even if it is unfortunately probably a bit more complicated than I first thought.
 

smcc

Member
Joined
Aug 18, 2020
Messages
16
Likes
14
...from what I've read, the reason people don't recommend diffusors for small rooms is that there isn't enough distance for them to work correctly...I do also remember that there are certain types that can…but not readily available, need to build…

…tbh your measurements look great, there’s many studios that wouldn’t look that good in data…hence why i was thinking maybe placement was a bit…tbh, u might just be hearing the limit of a three way…can’t beat coax for imaging…

…i like the 310s for mixing as FR is so linear and distortion is so low, and they have good headroom for tracking unprocessed instuments…but for just listening I’d go with a coax…or a set of sh50s even, not totally accurate, but pure joy…

I wouldn’t go crazy spending any more money on acoustics…think you’re at the point of very diminishing returns…maybe play with moving things around a bit…dabble with some eq…
 
Last edited:

youngho

Active Member
Joined
Apr 21, 2019
Messages
282
Likes
391
...from what I've read, the reason people don't recommend diffusors for small rooms is that there isn't enough distance for them to work correctly...I do also remember that there are certain types that can…but not readily available, need to build…
I have heard that, as well. https://www.audiosciencereview.com/...om-subs-ma1-absorbers-avaa.27288/#post-950938

Also:
 

sadburai

Member
Joined
Apr 17, 2021
Messages
10
Likes
3
very interesting! I think it would be even more interesting, if you showed not only the frequency response for every additional measure (i.e. your Steps 1-7), but also the time domain (e.g. Waterfall Plot or RT60 plots). I guess that the frequency response shows only half of the story to evaluate the effect of each treatment step.
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
1,712
Likes
1,203
Location
Canada
very interesting! I think it would be even more interesting, if you showed not only the frequency response for every additional measure (i.e. your Steps 1-7), but also the time domain (e.g. Waterfall Plot or RT60 plots). I guess that the frequency response shows only half of the story to evaluate the effect of each treatment step.

It might be informative to include spectral decay plots as well.

Related to what I wrote before i.e. bringing back some lost envelopment cues due to over-absorption by using upmixing DSP.


Here's some graphs showing the effect of upmixing a plain stereo L+R sine swept signal to surround channels:

1635710460013.png 1635710551750.png

The additional surround signal is added after 20ms delay time -- visible as a faint (but brighter) energy line in the horizontal plain in the second wavelet graph.


The decay graph (after upmixing) also slightly appears more smooth/even overall as the surround channels help fill-in many of the dips in the later spectra:

1635710774999.png 1635710780857.png

At the very least JRiver's upmixing process does not seem to be doing any obvious harm or alteration to the direct sound when set up correctly.
 

sadburai

Member
Joined
Apr 17, 2021
Messages
10
Likes
3
The angle itself is not important, but it has to be enough so the reflections are deflected from the listening position. [...] I use risers with a fixed height, so i can not really adjust the angle. You can try with a mirror on the desk and see how much you have to tilt it so you don´t see the speaker in it anymore.

Ergonomics is bad. I angle the desk only for listening to music. It is okay for surfing the internet while listening and i still can place a bottle or a cup on it, but i would not use it that way during work. I would not recommend a window behind the screen either. For working, i remove the risers

It is certainly possible to place absorbers on the desk instead of angling it. I don´think a mouse mat will be effective, though. I own a big mouse mat (90 x 40 cm), but it did not really help, i think, but I don´t have measurements of it.

I gave it a try to place absorbers on the table and I think it's much more convenient than to tilt the whole table. According to my measurements, it's sufficient to only treat areas on the desk according to the mirror method. At least it made a big difference, although I didn't try to tilt my desk for comparison.

I didn't have any absorbers at home, so instead I used:

1.) an IKEA TOFTBO floormat for the bath, folded to 5cm height and placed it on the table. It was slightly sloping downwards from the speakers to the edge of the table (mirror method).
2.) A small wooden board (approximately 30x40), angled such that I could not see the speaker from the listening/mic position in the mirror. The board was placed ca. 10cm away from the speaker, so that there is a gap. Otherwise, it's hard to not see the speaker in the mirror from the mic position.

You can still work at the desk relatively conveniently and the treatment can be added/removed within seconds.

Measurements:

Disclaimer: I used an uncalibrated Audyssey Mic, with the SPL dialed in by an Android App. The room is untreated.

The measurements show L+R channels simultaneously, but the effect is very similar for the individual speakers. There is not much difference between the IKEA mat and the angled wood in individual measurements. Both methods reduce the early reflection peaks in the impulse response and flatten the frequency response.

Impulse response:
- brown before treatment with strong reflections from the table after 1ms
- green after treatment with an IKEA mat (L) and an angled piece of wood (R)

Impulsantwort, L+R, Tischerstreflexion,gedämpft.jpg


Frequency response:
- much flatter after treatment (green). 1/48 smoothing
Frequenzgang, L+R, Tischerstreflexion,gedämpft.jpg
.
A mousepad on a flat table doesn't do much however.

Impulsantwort R, Mauspad.jpg
 
Last edited:
OP
DjBonoBobo

DjBonoBobo

Senior Member
Joined
Jan 21, 2020
Messages
413
Likes
621
Location
any germ
very interesting! I think it would be even more interesting, if you showed not only the frequency response for every additional measure (i.e. your Steps 1-7), but also the time domain (e.g. Waterfall Plot or RT60 plots). I guess that the frequency response shows only half of the story to evaluate the effect of each treatment step.

You can PM me if you want, and i send you my mdat-file. Too much work for me to post all the graphs ATM.
 

sadburai

Member
Joined
Apr 17, 2021
Messages
10
Likes
3
I can't edit my previous post anymore, but I still wanted to upload photos of my make-shift absorber & angled reflector to reduce the early reflections from my table. Super simple but effective. The gap between the angled board and the speaker makes it easier to place the board according to the mirror method such that no reflections arrive at the listening position.

table_absorber.jpg
 
Top Bottom