A way of gettin’ the bass right!

In every listening room I had until today standing waves were a problem, sometimes more sometimes less. The audible effect of standing waves is that there is too much or too little bass - or most of the time both at the same time!
Here I will show this effect using my current listening room. It's a square room in the basement with stone and concrete walls. and therefore bass problems are inevitable. Especially when I got my new speakers, which are quite big for the room and capable to produce "enough" bass with two 12“ drivers on each side and a - 6 dB point  at 33 Hz under free field conditions.
The first step with new speakers is always to find a position where they sound the best.
When I found that position I realized that the bass was not like what one would expect!
Bass was quite deep and around 35 Hz it was simply too much. But at the same time electric bass and drums lacked significantly body and impact!

This time I was not willing to accept this as given because I knew already what the speakers are capable of and wanted to analyze and address the problem. This requires measurements and so I got a free software (REW) and a calibrated USB microphone. REW is a very powerful program which allows all relevant measurements for speaker analysis and room acoustics. Until now I use perhaps 10 % of the possibilities!
https://www.roomeqwizard.com
REW offers also a room simulation, so why not start with that! To enter all relevant data took a few minutes and gave ma the following graph:


Everything above the blue line is too much bass and everything below is too little. The most prominent ones are +10 dB at ca. 36 Hz and the -8 dB at ca. 58 Hz plus many narrow deviations above 70 Hz.
This corresponds pretty well to what I hear but now it’s time for some real measurements with the microphone placed at the listening position. Here the results (imaginary zero line is now red):


Here I find remarkable:
- the bass extension in the room is pretty good down to 20 Hz!
- the modes at 35 Hz and 50-60 Hz are worse than in the simulation!
- above 65 Hz there is nothing to complain and the real response is much better than the simulation. This is due to the excellent directivity of my speakers down under 100 Hz, while the simulation assumes a omnidirectional radiation pattern.

So, what to do? Finding another position of the speakers and listening seat ist always a good idea, but the only place where the bass modes are minimal is with the listening seat in the middle of the room. I tried this, but it’s too close to the speakers and the sound isn’t really good. DSP room correction is omnipresent, but I want to preserve a fully analogue chain and the shown modes can’t be corrected with DSP.
But the problematic area here is only from 30 - 65 Hz! I decided to adress this with a pseudo-DBA.
According to Wikipedia a DBA is "a specific layout of subwoofers within a rectangular listening space. It removes unwanted room related resonances (modes) over a wide listening area.“
https://en.wikipedia.org/wiki/Double_bass_array
In short, the idea is the following: one array (grid) of subwoofers is placed at the front wall. This array produces a quite evenly (bass) sound wave which travels to the rear wall. At the rear wall is a 2nd identical array of subwoofers. This array gets a signal which is delayed by the time which the sound wave needs to travel from the front to the rear wall and inverted and therefore the sound wave is cancelled at the rear wall. The rear bass array does therefore not contribute to the heard sound but only to prevent the bass sound to be reflected and building standing waves. It might also be said, that the second bass array simulates an open rear wall for the bass.
In my case I wanted to use the main speakers are the first array, therefore I call it a pseudo-DBA. The second array are two passive subs which I placed a little elevated at the rear wall at 1/4 and 3/4 of the room width.
The signal path is as following: the power amp is connected as usual to the preamp. The signal goes parallel from the preamp into a miniDSP 2x4 HD where it is filtered to 80 Hz, delayed for about 9 ms (the distance between main speaker and subs/rear wall) and inverted. Then it goes into a small Class-D amp with volume knob to adjust the volume of the subs to the appropriate level which just is enough to cancel the bass of the main speakers.
It needed some tries to get the filtering, delay time and the volume right but then miraculously I measured the following:


The mode at 35 Hz is minimized and the „hole“ 50-60 Hz is completely gone!
A listening test confirmed the measurement. The bass goes very deep, is firm, tight and fast. Drums have attack and body.The resonance at 35 Hz is slightly audible, but only in a pleasant way!

mannitheear escreveu:

...


The mode at 35 Hz is minimized and the „hole“ 50-60 Hz is completely gone!
A listening test confirmed the measurement. The bass goes very deep, is firm, tight and fast. Drums have attack and body.The resonance at 35 Hz is slightly audible, but only in a pleasant way!

Its great that you've achieved such a good room correction ... but the sound isn't "dirty" on bass isolated lines as cellos or barroco violas ??? There's no "phamtom" sounds on electronica drum & bass lines ???...

Anyway, i'm impressed by the measurement results ... i've read something about the bass array but never heard it anywere ...

BRAVO

TD124 escreveu:

mannitheear escreveu:

...


The mode at 35 Hz is minimized and the „hole“ 50-60 Hz is completely gone!
A listening test confirmed the measurement. The bass goes very deep, is firm, tight and fast. Drums have attack and body.The resonance at 35 Hz is slightly audible, but only in a pleasant way!

Its great that you've achieved such a good room correction ... but the sound isn't "dirty" on bass isolated lines as cellos or barroco violas ??? There's no "phamtom" sounds on electronica drum & bass lines ???...

Anyway, i'm impressed by the measurement results ... i've read something about the bass array but never heard it anywere ...

BRAVO  

No, solo cello sounds very even and coherent to me! And on the deep bass lines of electronica or film scores I can't detect any strange or phantom sounds. All is quite focussed and clean!
Some people might even miss bass....until there is really some on the recording....

To be honest, it needed a lot of trying to get at this point The first attempts measured not bad but the bass seemed slow and blurred. At first I used 24 dB and 48 dB filters on the miniDSP but these filters have the same characteristics as if they were analogue and therefore the subs had a different phase response than the main speakers which results in a bad integration known from subwoofer - satellite systems.

Then I tried a very steep FIR Filter without any phase shift but the result was not much better!

Now it's a 6 dB filter to achieve the same phase behavior with the subs and the main speakers (which use 6 dB filters) but at first there was too much midrange energy on the subs! Therefore I filtered the midrange with two additional low pass high shelf filters and now it's quite listenable!

Hi Manfred,

I started using room EQ last year and will never go back.

I have a few questions and requests if you don't mind.

How did you perform the measurements, did you use the Moving Microphone Method (better for EQ'ing) or a sweep?

Did you measure one speaker at a time (there is no point in measuring both speakers at the same time because each will have to be EQ'ed separatly and when playing simultaneously they will produce cancelations/comb-filtering)?

Could you produce a plot from the "All SPL" tab of the room response from 20Hz to 20,000Hz with Variable Smoothing and reduce the vertical scale to 50Hz?

Have you tried creating the filters on REW and exporting them to you miniDSP?

Are you high-passing your mains?

I see no reason for your first-order + shelving filters, you've put the subs and mains reproducing the same frequencies over a wide-range.

Best wishes,
Ricardo

Hi Ricardo,

what I do is NOT room EQ!!!!

This has two reasons:
1. I want to preserve the analogue signal when I listen to vinyl
2. The main problem is a very severe suckout (cancellation due to room modes) between 45 and 65 Hz and it is not possible to fix that with EQ because it's a resonance problem, and regardless how much energy you pump into the speakers, the physical cancellation by room resonances remains! (I tried that many years ago with a digital EQ)
You can correct only a room-mode caused exaggeration, but not very good because if you reduce the level of the exaggerated frequency this will correct the steady state - but then there is too little energy in the first periods until the resonance has fully built up.

What I do is something very different: the main speaker signal remains untouched. This signal is routed to a DSP which applies:
1. a low pass (newest variant is a 12 dB Bessel filter at ca. 100 Hz) because I only need to fix the range 30...80 Hz.
2. a delay for the time which the sound needs to travel the distance between main speakers and the rear wall (10,5 ms)
3.a signal inversion
The processed signal feeds 2 subwoofers at the rear wall opposite the main speakers at ca. 1/4 and 3/4 of the room width.

The idea is, that the bass wave from the main speaker are cancelled at the rear wall by the inverted signal of the subwoofers which prevents the room mode to resonate at all.

It works good enough with 2 subwoofers. For a more DBA textbook operation you would need more subwoofers, for example 4 at the front wall and 4 on the back wall. But this has drawbacks because now the mains must be delayed with a DSP - which I do not want to use for the main signal.

Please ask if you have questions and it's still unclear!


Manfred

PS: in the mids and treble I have no need to correct anything, because my speakers are quite linear, have a very constant directivity from 100 Hz upwards and the crossing between the tweeter and the mid/bass is time aligned by the passive crossover. So, no need to correct anything - and this benign behaviour does not depend on the room.

mannitheear escreveu:

Hi Ricardo,

what I do is NOT room EQ!!!!

This has two reasons:
1. I want to preserve the analogue signal when I listen to vinyl
2. The main problem is a very severe suckout (cancellation due to room modes) between 45 and 65 Hz and it is not possible to fix that with EQ because it's a resonance problem, and regardless how much energy you pump into the speakers, the physical cancellation by room resonances remains! (I tried that many years ago with a digital EQ)
You can correct only a room-mode caused exaggeration, but not very good because if you reduce the level of the exaggerated frequency this will correct the steady state - but then there is too little energy in the first periods until the resonance has fully built up.

What I do is something very different: the main speaker signal remains untouched. This signal is routed to a DSP which applies:
1. a low pass (newest variant is a 12 dB Bessel filter at ca. 100 Hz) because I only need to fix the range 30...80 Hz.
2. a delay for the time which the sound needs to travel the distance between main speakers and the rear wall (10,5 ms)
3.a signal inversion
The processed signal feeds 2 subwoofers at the rear wall opposite the main speakers at ca. 1/4 and 3/4 of the room width.

The idea is, that the bass wave from the main speaker are cancelled at the rear wall by the inverted signal of the subwoofers which prevents the room mode to resonate at all.

It works good enough with 2 subwoofers. For a more DBA textbook operation you would need more subwoofers, for example 4 at the front wall and 4 on the back wall. But this has drawbacks because now the mains must be delayed with a DSP - which I do not want to use for the main signal.

Please ask if you have questions and it's still unclear!


Manfred

PS: in the mids and treble I have no need to correct anything, because my speakers are quite linear, have a very constant directivity from 100 Hz upwards and the crossing between the tweeter and the mid/bass is time aligned by the passive crossover. So, no need to correct anything - and this benign behaviour does not depend on the room.

Hi Manfred,

That's one interesting and original approach to room correction. I now understand Paulo's wariness.
You may be solving a problem in the frequency domain but creating another in the time domain, which is why I asked how you measured your speaker (are you still using ProAcs?).
If you've measured using a sweep, can you produce a waterfall?

Have you tried to digitise the signal then high-pass to the mains and low pass to the subs?
I think that the miniDSP can do that.

You are right that the nulls have to be treated acoustically with bass traps but if you EQ the room peaks you are actually preserving the analogue sound at the listening spot by making it closer to electrical signal.
And since you are already digitising the signal why not make the most of it?

I asked to see the response above 200Hz to get an idea of the balance at the listening spot.

Best wishes,
Ricardo

neo_2018 escreveu:

(...)

Hoje em dia, invisto apenas nas variáveis que produzem efectivamente alterações profundas no sistema: Colunas e DSP.

Recomendo vivamente que experimentem um sistema DSP dos que já referi, como Dirac Live, Room Perfect, etc.
Meia duzia de clicks, e o sistema corrige automaticamente a influência negativa do espaço, corrige os desequilíbrios temporais entre colunas, permitindo inclusive adaptar a resposta final ao nosso gosto.

P.ex. tenho duas calibrações diferentes para Cinema e Musica: para cinema, um reforço enorme da região 10-100Hz usando uma House Curve agressiva; para musica, um HPF aos 20Hz, e um House curve mais relaxada e natural.

Com o mote do neo_2018 e reabrindo este tópico do Mannitheear, até com a excelente experiência aqui relatada por ele, vamos lá fazer o favor de relatar outras experiências de sucesso na correção acústica.

Claro que este pedido é interesseiro, por 3 razões:

1. Sinto necessidade de fazer alguma correção acústica da minha sala principal e para o set principal;

2. Não estou disponível para fazer aquelas cenas de painéis acústicos, porque no dia em que fizer isso, tenho as malinhas à porta, pela Senhora Dona, desta casa.

3. A correção DSP é algo que não conheço e gostaria de conhecer, até porque o meu atual AV receiver tem essa possibilidade, não sei se é bom se é mau, mas apliquei e até me soa muitíssimo bem. Fiquei entusiasmado.

Goansipife escreveu:

até porque o meu atual AV receiver tem essa possibilidade, não sei se é bom se é mau, mas apliquei e até me soa muitíssimo bem.

Qual é o AV?

Note-se que a correcção efectuada pelo @mannitheear é algo diferente e um bastante mais complexo (tenho de ler com atenção antes de comentar )

Marantz SR5014

A correção das colunas foi feita por uma aplicação que se chama Audyssey

O Marantz SR5014 tem "apenas" a versão Audyssey MultEQ XT, que é inferior à versão Audyssey MultEQ XT32, nomeadamente em nº de filtros e capacidade de processamento:



No entanto, em cenários pouco complexos, mesmo a versão versão Audyssey MultEQ XT poderá ter excelentes resultados.

Ambas são inferiores a DSP's como Dirac Live, Anthem ARC, Room Perfect, etc, e superiores às soluções proprietárias da Onkyo, Pionner, Yamaha, etc.

Caro neo_2018,

Sim é verdade. Até aí eu já aprendi, mas 2 coisas:

1. Eu comecei por dizer que de DSP não percebo grande coisa, ou mesmo nada;
2. Este Marantz foi um excelente negócio de oportunidade e quando quis passar a ter Netflix.

Esta área do AV, que tentei entrar nos anos 90, nunca acabou por ser o meu maior must. Agora até aprecio mais do que na altura. Também é uma área que domino mal e já percebi que existem soluções bem melhores mas, calhando, ocupar-me-iam mais espaço, que também não tenho

Obrigado

Caro Goansipife, não conheço esse AV, mas os princípios são (com mais ou menos funcionalidades e precisão) aplicados a todos os sistemas de correção.

O que é que sente que quer corrigir?

Rstuyvesant escreveu:

Caro Goansipife, não conheço esse AV, mas os princípios são (com mais ou menos funcionalidades e precisão) aplicados a todos os sistemas de correção.

O que é que sente que quer corrigir?

O que sinto e quero corrigir não é tanto no set de AV, até porque com esse convivo bem e agora não é a minha maior prioridade.

O que quero corrigir é no set dedicado ao áudio. Tenho um ligeiro empolamento de graves, algum descontrolo, que gostaria de corrigir sem utilização de acessórios de tratamento de sala, que a cara metade não vai nisso

Há várias formas de corrigir um sistema stereo, algumas mais automáticas do que outras.

A forma mais simples e relativamente acessível é usar um DSPeaker Antimode 8033-S (aprox. 300€):
- http://www.dspeaker.com/en/products/anti-mode-8033.shtml

Basicamente é ligar e já está, ele corrige automaticamente entre os 16-250Hz.

Uma alternativa manual é o miniDSP 2x4HD (250€), que permite aplicar 10 bandas de PEQ, em 4 configurações alternativas.
O problema é que é necessário um microfone e usar o REW para calcular os filtros.

O miniDSP 2x4HD está também disponível em versão Dirac (450€), permitindo aplicar automaticamente o sistema Dirac, mas necessita de um microfone para efectuar as medições para aplicar a correcção.

Caro Goansipife, concordo com o caro Neo_2018 nas alternativas que apresenta. O anti-mode é o mais out of the box possível, sem complicações, e sendo boa ferramenta para qualquer modificação de futuro, o preço paga-se a si prórpio.
Mas... antes disso:
a) ouça um sweep de frequências (os cds de teste trazem isso como faixa), e identifique a(s) quais estão a contribuir para o empolamento (no limite pode ser só uma que esteja muito má mas porque se reflete na propagação dos seus harmónicos pode instaurar o caos);
b) altere o seu ponto de escuta pela sala para a frente e/ou para trás do ponto habitual;
c) descobrirá que haverá pontos de cancelamento e de incremento de frequências;
d) não há perfeição (mesmo em salas acusticamente tratadas procura-se um compromisso);
e) pode chegar à conclusão que uma posição de escuta diferente é um bom compromisso sem estar com gastos;
f) se a estética da sala for para manter e e) não poder ser possível.... avance para as alternativas do caro Neo.

Caros neo_2018 e Rstuyvesant,

Obrigado a ambos. Vou investigar.

Há algum tempo que ando de volta do miniDSP. Mas sim, vou dar mais umas "voltinhas na e com a sala".