Gearslutz

Gearslutz (https://www.gearslutz.com/board/)
-   Studio Building / Acoustics (https://www.gearslutz.com/board/studio-building-acoustics/)
-   -   Help needed, Problems at 70Hz (https://www.gearslutz.com/board/studio-building-acoustics/1321328-help-needed-problems-70hz.html)

schlorian 8th August 2020 06:45 PM

Help needed, Problems at 70Hz
 
3 Attachment(s)
Dear Gearslutz community,

I have been working on the acoustic optimisation of my room for some time now and in this way i have encountered a problem for which i currently know no solution. Maybe one of you knows what to do? Here are the facts:

The room
-Room dimensions LxWxH: 425x320x260cm
-the two long sides consist of approx. 20cm presumably brickwork
-the two short sides are exterior or load-bearing walls and consist of approx. 40cm of brickwork. The outside wall has additionally approx. 30cm heat insulation on the outside.
-the ceiling is plasterboard, probably with wooden beams and compacted gravel.
the floor consists of wooden planks, probably screed underneath and has a cellar (ground floor).

The target
-Mainly production of music, recording of various instruments via MIDI or DI, -reference listening of mixes/masters and recording of vocals
-Optimized frequency response, reverberation time from approx. 200ms to approx. 200Hz (roughly based on EBU Tech. 3276)
-Getting a grip on initial reflections in the listening position
-However, the room should not sound completely "dead" and be fun to make music in!

The idea
-sort of "Dead End" in the listening position
-sort of "Live End" at the other end of the room for vocal or acoustic guitar recordings

The implementation to date
-Measured the room modes diagonally across the room
-I tested the speakers at different places in the room and roughly determined the best position.
-then I tried to get the bass range under control with porous absorbers. To do this I stacked Sonorock packages in the corners and made measurements with REW in different setups. Except for the longitudinal mode at 41Hz I got most of it under control.
-There was a peak at about 130Hz. The Room Simulator of REW shows me that the octave of the floor-ceiling mode (0-0-2) is approximately there. At 65Hz (0-0-1) it shows more of a dip. So I put an absorber (16cm Sonorock) above the listening position.


The Problem
the ceiling absorber reduced the peak at 130Hz by about 5dB, but now I have a huge dip of 24dB compared to the measurement without ceiling absorber at about 70Hz.

My Guesses
-During my first measurements, from the floor (loudspeaker) to the diagonally opposite corner of the room (microphone), a large dip at 67Hz is already visible. This dip was well compensated by the superchunks in the corners, so it must have something to do with the ceiling absorber.
-In my opinion it can be ruled out that the valley at 70Hz is caused by SBIR, because the speakers would have to stand about 122cm away from the wall for a phase cancellation of 70Hz (Wavelength ~490cm).

-Since I can't hang anything from the ceiling (plasterboard), I put the ceiling absorber on battens at a height of about 2m. Is it possible that the height of the ceiling absorber plays an important role?
-If I play 70Hz as a sinus tone and roll on my office chair towards the back wall, the sound under the absorber is much quieter. The Spectrogram shows that the decay of 70Hz increases again in volume. Could it be that the difference in ceiling absorption between the front and back of the room leads to the valley at 70Hz? Should I therefore also equip the rear part of the ceiling with absorbers?

I have attached drawings, pictures and REW measurements.

I would be very happy if you could help me with my questions.
Thanks in advance and best regards,

schlorian

OpusOfTrolls 8th August 2020 10:08 PM

It could be second order reflections (involving 2 surfaces) caused by the wall/ceiling and wall/floor edges to the left and right of the speaker positions.

One idea is to make the speaker triangle wider. If this doesn't improve the null, then I would advise installing absorption in those room edges. They would need to be similar in size to the chunks you already have in the front wall/wall edges. You may be able to borrow the top portion of those chunks and set them on the floor where the floor and wall meet.

thethrillfactor 9th August 2020 12:18 AM

1 Attachment(s)
Can you measure left and right separately from 20hz and 20khz.

Also a null that is a deep notch is preferable to a wide null but not as deep. When its a notch that deep it wiil go away if you move the mic and measure at a different spot. This is why just going by a frequency response in one spot at one point in time is not reliable. You never keep your head in one location when mixing or listening, This is why its best if you are using frequency response as a tool, to do many measurements around the listening position and make an average to see what the situation really is like.

By the way, your minimum phase group delay analysis shows its most likely the result of some form of SBIR.

schlorian 10th August 2020 09:04 AM

1 Attachment(s)
Thank you guys for your advice! Greatly appreciated! I did some more measurements and attached them to this post.

Quote:

Originally Posted by OpusOfTrolls (Post 14904380)
It could be second order reflections (involving 2 surfaces) caused by the wall/ceiling and wall/floor edges to the left and right of the speaker positions.

One idea is to make the speaker triangle wider. If this doesn't improve the null, then I would advise installing absorption in those room edges. They would need to be similar in size to the chunks you already have in the front wall/wall edges. You may be able to borrow the top portion of those chunks and set them on the floor where the floor and wall meet.

I tried to make the speaker triangle wider (20cm) and made some more measurements. Absorption in those edges is a good Idea! I Will definitely consider this.




Quote:

Originally Posted by thethrillfactor
Can you measure left and right separately from 20hz and 20khz.

Also a null that is a deep notch is preferable to a wide null but not as deep. When its a notch that deep it wiil go away if you move the mic and measure at a different spot. This is why just going by a frequency response in one spot at one point in time is not reliable. You never keep your head in one location when mixing or listening, This is why its best if you are using frequency response as a tool, to do many measurements around the listening position and make an average to see what the situation really is like.

By the way, your minimum phase group delay analysis shows its most likely the result of some form of SBIR.

In the new measurements I measured from 20-20khz and also did multiple Mic Positions for one Speaker position. The Dip at 70Hz changed a bit in deepness and frequency. I also measured the Distance from the right speaker to the right wall, and it's 122cm, which could be the cause of the SBIR in my opinion.

What confuses me is that you can also clearly see a dip in this frequency region in the measurement which went diagonally through the Room. Maybe multiple Effects on top of eachother?


Thanks guys!
best regards,

schlorian

schlorian 10th August 2020 09:11 AM

1 Attachment(s)
Here Comes the 2nd Part of the 2nd Measurement.

akebrake 16th August 2020 04:11 PM

Hi Schlorian!

Quote:

-Measured the room modes diagonally across the room
Yes, that's fine! The first set of measurements contains 3 sweeps. One speaker on floor in the corner & Mic diagonally close to the ceiling?
Empty room? Corner traps? Cloud? Can you elaborate?

Quote:

-then I tried to get the bass range under control with porous absorbers. To do this I stacked Sonorock packages in the corners and made measurements with REW in different setups.
You need a lot more (large area & deep) in order to control frequencies under 100 Hz.

I guess Sonorock comes quite compressed in the delivery package.
How much area/weight is one package? GFR?

While convenient for testing (using delivered units) one can improve the bass absorption by open up the packages and re pack (less dense) in plastic bags.
That way y can get more traps :)

It's difficult to evaluate the result when one moves a couple of traps from one room position to another. Everything counts as sound is bouncing around at 343m/s...

IMHO It's better to measure, then add traps and measure again and watch what happens.
E.g you can test Opus idea of floor/wall and wall/ceiling trapping.

Best

thethrillfactor 17th August 2020 04:04 AM

Quote:

Originally Posted by schlorian (Post 14906372)





In the new measurements I measured from 20-20khz and also did multiple Mic Positions for one Speaker position. The Dip at 70Hz changed a bit in deepness and frequency. I also measured the Distance from the right speaker to the right wall, and it's 122cm, which could be the cause of the SBIR in my opinion.

What confuses me is that you can also clearly see a dip in this frequency region in the measurement which went diagonally through the Room. Maybe multiple Effects on top of eachother?


Thanks guys!
best regards,

schlorian

You can see in the new measurements that dip between 69hz-74hz is moving which is SBIR( also when you look at excess Group delay/minimum phase it shows the deep null in all of them). The only way to correct is to treat all the boundaries that are affecting it( sometimes this means you have to treat all of them except the ceiling).

schlorian 18th August 2020 10:23 PM

Quote:

Originally Posted by akebrake (Post 14918470)
Yes, that's fine! The first set of measurements contains 3 sweeps. One speaker on floor in the corner & Mic diagonally close to the ceiling?
Empty room? Corner traps? Cloud? Can you elaborate?

Hi akebrake, thank you for your answer!

Sorry, I just realized that I didn't translate the german text into english. The first set of measurements is (1) "ohne Deckenabsorber" = without ceiling absorber, (2) "mit Deckenabsorber" = with ceiling absorber, (3) "Diagonal durch den Raum" = diagonally across the room.

When I measured diagonally across the room wasn't fully empty (Two Shelfes and a couch on the long wall), but there was no absorption installed anywhere in the room. The diagonal Measurement was with the Speaker on the floor of the front-left Corner and the microphone in the back-right corner close to the ceiling.

Measurements (1) and (2) were with traps in all four corners, like on the image you can see in my first post.

Quote:

Originally Posted by akebrake (Post 14918470)
You need a lot more (large area & deep) in order to control frequencies under 100 Hz.

I guess Sonorock comes quite compressed in the delivery package.
How much area/weight is one package? GFR?

I guess you have a point there, porous absorbers aren't the best solution for low frequencies as far as I know. But surprisingly, in this room it did work quite well until down to around 70Hz in terms of reverberation time. I also hope that the performance will increase when I unpack the delivery packages.

One Package of Sonorock is round about 100x62,5x48cm, it contains 3,75 square meters of 80mm thick absorption. It has around 6 kPa.s/m2. Unfortunately I can't find the weight/m3.

Quote:

Originally Posted by akebrake (Post 14918470)
you can test Opus idea of floor/wall and wall/ceiling trapping.

Yeah that's a good idea! I will put trapping there and do some more measurements next.


Quote:

Originally Posted by thethrillfactor (Post 14919415)
You can see in the new measurements that dip between 69hz-74hz is moving which is SBIR( also when you look at excess Group delay/minimum phase it shows the deep null in all of them). The only way to correct is to treat all the boundaries that are affecting it( sometimes this means you have to treat all of them except the ceiling).

Hi thethrillfactor, thank you for your advice! I will treat all the boundaries this weekend and then provide more measurements.



Thanks guys!

Cheers,
Schlorian

thethrillfactor 18th August 2020 10:52 PM

Quote:

Originally Posted by schlorian (Post 14922446)

Hi thethrillfactor, thank you for your advice! I will treat all the boundaries this weekend and then provide more measurements.



Thanks guys!

Cheers,
Schlorian

Your best recourse is to treat down to the axial modes for the boundaries.

akebrake 21st August 2020 07:27 PM

2 Attachment(s)
Quote:

Originally Posted by schlorian (Post 14922446)
Hi akebrake, thank you for your answer!

You're welcome! :)

Quote:

But surprisingly, in this room it did work quite well until down to around 70Hz in terms of reverberation time. I also hope that the performance will increase when I unpack the delivery packages.
I guess: ”the performance will increase” means a flatter FR?
Just by adding more chunks in the corners? That’s wishful thinking unfortunately...gooof

By "repacking" one can get more bang for the bucks (re Topt at least).
The fluffy must be contained still. With plastic or other methods.

A lot more than corners are needed especially with your massive boundaries.
The Early reflections (SBIR) are strong and large untreated areas will interfere destructively with the direct sound.

Try improving the Spkr / Listener position.

Measuring dozens of sweps while moving mic and spkr a few cm (in between measurements) is interesting but can be quite tedious...

A much quicker method is using REW RTA together with REW Pink PN (Periodic Noice) No sweep, and watch the result in Real Time!
In the low end you’ll get very similar results. (Steady state measurement)
Example here from my TV room.
Pic 1 RTA window and REW Gen: Note RTA & Gen set/up parameters.
Pic 2 Pink PN & Sweep overlaid.

Best