How to improve Clarity (C50) in Bass in (very) small Rooms

[Svarog98]

I've come across a very interesting thread which states that a C50 in Bass above 10db (even better 15db) is perceived as a tight and impactful bass
"A correlation exists and is related to better GD as well and time alignment around the sub-mains xo. Even if theoretically C50 is a metric supposedly more in line with speech. I already knew that, but my own listening tests agree with the improved measurements making the bass line “better”. It’s not just a difference in SPL or frequency amplitude by any means."

Measure your C50 in REW and tell us how your bass sounds!

You are all invited to participate in this experiment to test the idea that a high C50 value as measured by REW equates to good, tight sounding bass. The idea here is that some value of C50 may correlate to subjectively good sounding bass in listening rooms. We're specifically looking at the C50...

www.audiosciencereview.com

This is the latest measurement of my system in a light treated room:

How is clarity in bass measured and are absorbers the only way to increase clarity? What about double bass arrays and distributed bass arrays like "the swarm" (http://www.audiokinesis.com/the-swarm-subwoofer-system-1.html) Do they only flatten the bass response or does it do something for clarity too?
I'd be grateful if somebody could share a page or a paper that explains how clarity is measured.

 

[Curvature]

C50 is an simple ratio used in architectural acoustics for speech-dominated areas like classrooms and lecture halls. It compares the energy of the first 50ms and the energy of everything coming after 50ms to the end of the window. This is particularly important in large spaces, where direct sound can be very weak compared to reflections due to distance.

It's pointless to use it to judge your room and bass in particular.

 

[ernestcarl]

C50 is an simple ratio used in architectural acoustics for speech-dominated areas like classrooms and lecture halls. It compares the energy of the first 50ms and the energy of everything coming after 50ms to the end of the window. This is particularly important in large spaces, where direct sound can be very weak compared to reflections due to distance.

It's pointless to use it to judge your room and bass in particular.

Pointless would be to look at the C50 curve by itself to judge bass in small rooms -- same with RTx. Nevertheless, one should be able to see a good correlation between higher C50 values and better in-room transient response and reflection-decay behaviour. I'm sure you already know the magnitude response can be arbitrarily equalized to flat in-room yet this will not necessarily improve the objectively measured transient response. Yes, beyond a certain value, let's just say between 10-15 dB (as originally suggested by @Tim Link) it starts to quickly become immaterial. Personally, I prefer to look at the spectral decay and spectrograms.

 

[Philbo King]

I suspect a better test for 'bass tightness' would be a 10 cycle sine burst of bass at various freqs (perhaps bass nodal freqs) with a measurement mic to measure decay time.

 

[ernestcarl]

What about double bass arrays and distributed bass arrays like "the swarm" (http://www.audiokinesis.com/the-swarm-subwoofer-system-1.html) Do they only flatten the bass response or does it do something for clarity too?

Assuming resonances and nodes or nulls are successfully reduced with the help of multiple distributed subs and EQ, for sure, clarity levels will increase.

Excess phase correction in the bass (not always applicable) is probably the last step one should consider -- some would say it's the "icing on the cake":



There is an audible difference between the two, absolutely -- i.e. bass in the latter (FIR filtered) sounds louder/full and midrange is a little more clear/clean sounding.

 

[Curvature]

Pointless would be to look at the C50 curve by itself to judge bass in small rooms -- same with RTx. Nevertheless, one should be able to see a good correlation between higher C50 values and better in-room transient response and reflection-decay behaviour. I'm sure you already know the magnitude response can be arbitrarily equalized to flat in-room yet this will not necessarily improve the objectively measured transient response. Yes, beyond a certain value, let's just say between 10-15 dB (as originally suggested by @Tim Link) it starts to quickly become immaterial. Personally, I prefer to look at the spectral decay and spectrograms.

This doesn't reflect the way C50 is used or where it matters.

As room size goes up, reverb time goes up, as does the distance between the listener and each surface and the source, increasing as well the gap between direct sound, early and late reflections. In large "rooms" like concert halls these gaps are meaningful.

There's a related measure, C80, for music. Again the same calculation except using 80ms, and again meant for large rooms. C50 and C80 are highly correlated. In some seats C50/80 is negative. Increasing these to a few positive dB is a minimal design criteria. The current standard uses C50 for 500-2000Hz specifically. These are crude tools.

I don't consider either RT60 or C50/80 to have any real bearing on small rooms. The largest factor negatively impacting these ratios is increasing room size.

Make what you want of that. I've not found any interesting, meaningful research for small rooms that gives real guidance on reverberation while keeping all of the perceptual complexity in mind.

 

[ernestcarl]

I pretty much already agree that these are “crude tools”. They would not be the first things I would look at for reasons you have said. However, I’d at least want to know why clarity drops like a cliff in the low frequencies. And, despite these being “pointless” or not really created for small rooms — they still, yet again, correlate with other acoustically measurable anomalies or acoustic modifications (e.g. adding acoustic treatment).

 

[ozzy9832001]

I posted in the other thread when it was active.

My bass is super tight. Clarity increased significantly when I added the second sub. Is it super accurate because it's a small room? Probably not. However, there seems to be a correlation between higher numbers and tighter sound all around.

 

[Curvature]

And, despite these being “pointless” or not really created for small rooms — they still, yet again, correlate with other acoustically measurable anomalies or acoustic modifications (e.g. adding acoustic treatment).

There's no set way to evaluate those things. We are manipulating physical quantities and hearing changes without fully understanding the effects. Putting up treatment is easily audible but, "better?" I've not been able to make any sort of firm conclusion there beyond controlling SBIR, which is primarily an FR change.

However, I’d at least want to know why clarity drops like a cliff in the low frequencies.

50ms is very short for LF. 1 cycle at 20Hz. Decays are quite slow. The more similar the energy is on either side of the 50ms divide, the lower the ratio will be. It's not uncommon to see 1sec (20x longer) at LF, and then 400ms (8x longer) at MF and HF.

There's a paper by Fazenda showing lower LF decays are preferred, but it's the only paper I know of. I would not count your example of excess phase correction as proof given the difference in frequency response as well. Hard to disambiguate FR from decay, especially with high noise floors. Similarly, Fazenda's paper didn't have completely similar FR between different sub setups, including DBA.

One more factor relevant to reverb times is the Lombard effect, which everyone will have experienced in any loud environment. It's normally associated with speaking louder over ambient sound, as well as pitching up speech and few other things and, in turn, in social situations, everyone else at the bar or wherever will do the same, causing a feedback loop. In small rooms, with high reverb and high noise, the corollary is turning music up higher to hear the details, but that causes reverb to go up in level, initiating the loop. That and other things (unbalanced or inconsistent, or HF-dominated FR) are responsible for fatigue, for example.

Maybe if there is a place for treatment in small rooms it is to lower, not decay times, but the "reverb noise" floor you get when listening. In any case there is limited research here.

 

[Curvature]

I posted in the other thread when it was active.

View attachment 311386

My bass is super tight. Clarity increased significantly when I added the second sub. Is it super accurate because it's a small room? Probably not. However, there seems to be a correlation between higher numbers and tighter sound all around.

"Correlation"? A metaphorical correlation based on a memory and a sample of one?

What about the FR change when you added the second sub? It's well known that linear FR is perceptually tighter compared to resonant and highly varied FR. How well we can perceptually disambiguate linear FR that from lower reverb, and how certain we are that what we hear is better and not only different, is unknown.

Edit: Particularly when small rooms have no inherent clarity problems anyway.

 

[ernestcarl]

There's a paper by Fazenda showing lower LF decays are preferred, but it's the only paper I know of. I would not count your example of excess phase correction as proof given the difference in frequency response as well. Hard to disambiguate FR from decay, especially with high noise floors. Similarly, Fazenda's paper didn't have completely similar FR between different sub setups, including DBA.

I did not use my example as “proof” that an increase in C50 will cause a perceived improvement in transient response in all cases — other than in this one anecdotal case example. But boosting the bass also has a similar effect — minus the subtle effect of midrange clarity increase. Of course, the frequency response (magnitude) did not change. Excess phase “correction” means only the phase is manipulated using an inverse all pass filter. That’s usually what one means when we talk about frequency response… here, hardly anybody bothers with phase as that is also just another pointless metric according to relevant research, I’m sure.

There’s no significant “noise floor” included in my posted plots, but in my case, and and when I do such tests… I make sure that it is as low as possible.

I can create an ABX test if you want “proof” that I’m not imagining results other than pulling subjective descriptions out of my ass. But I’m also sure that’s not the kind of level proof you are seeking.

 

[ozzy9832001]

"Correlation"? A metaphorical correlation based on a memory and a sample of one?

What about the FR change when you added the second sub? It's well known that linear FR is perceptually tighter compared to resonant and highly varied FR. How well we can perceptually disambiguate linear FR that from lower reverb, and how certain we are that what we hear is better and not only different, is unknown.

Edit: Particularly when small rooms have no inherent clarity problems anyway.

Yes, the FR did get significantly better, as did the C50. In my case, there was a correlation or coincidence that one was linked to the other...whether directly or indirectly I don't know. In the original thread, the opposite appeared to be true as well, which could suggest to me that the room could be a determining factor. I don't claim that having a high C50 makes my bass super tight. What I'm saying is my bass is tight and I also appear to have a decent C50.

When examining my setup, I rarely, look at either the RT60 or Clarity.

 

[Curvature]

I did not use my example as “proof” that an increase in C50 will cause a perceived improvement in transient response in all cases — other than in this one anecdotal case example. But boosting the bass also has a similar effect — minus the subtle effect of midrange clarity increase. Of course, the frequency response (magnitude) did not change. Excess phase “correction” means only the phase is manipulated using an inverse all pass filter. That’s usually what one means when we talk about frequency response… here, hardly anybody bothers with phase as that is also just another pointless metric according to relevant research, I’m sure.

There’s no significant “noise floor” included in my posted plots, but in my case, and and when I do such tests… I make sure that it is as low as possible.

I can create an ABX test if you want “proof” that I’m not imagining results other than pulling subjective descriptions out of my ass. But I’m also sure that’s not the kind of level proof you are seeking.

No, I'm sure you heard what you heard. What I don't think is properly disambiguated is the reason why you heard it. In the wavelets you posted, the phase correction redistributes energy significantly below 150Hz between T0 and T20, including fixing a noticeable hole around 130Hz where you're cancelling a nonminimum phase issue, i.e., what is likely a destructive reflection.

I'd even say that phase correction is necessary where it fixes problems like that.

Any technique designed to reduce decay time will also cause a spectral change. It is far more likely the ear is responding to that and not to decay as such. Particularly in small rooms where direct sound is in your face and there are strong specular reflections.

My mentions of "noise floor" were referring to both:

• Normal room noise: HVAC, traffic, water in pipes. Whatever's happening in or on the other side of the wall.
• The pseudo-noise caused by music: the normal 400ms reverb will cause every musical phrase to persist for a short while at a low level, becoming a kind of noise floor. Lower reverb will mean a quieter room, overall.

 

[ernestcarl]

No, I'm sure you heard what you heard. What I don't think is properly disambiguated is the reason why you heard it. In the wavelets you posted, the phase correction redistributes energy significantly below 150Hz between T0 and T20, including fixing a noticeable hole around 130Hz where you're cancelling a nonminimum phase issue, i.e., what is likely a destructive reflection.

I'd even say that phase correction is necessary where it fixes problems like that.

Any technique designed to reduce decay time will also cause a spectral change. It is far more likely the ear is responding to that and not to decay as such. Particularly in small rooms where direct sound is in your face and there are strong specular reflections.

My mentions of "noise floor" were referring to both:

• Normal room noise: HVAC, traffic, water in pipes. Whatever's happening in or on the other side of the wall.
• The pseudo-noise caused by music: the normal 400ms reverb will cause every musical phrase to persist for a short while at a low level, becoming a kind of noise floor. Lower reverb will mean a quieter room, overall.

Yes, clarity values are merely indicative, which isn't really disputed.

 

[Flaesh]

What about double bass arrays .. Do they only flatten the bass response or does it do something for clarity too?

Front array in *BA (D or S) does not excite modes except longitudinal ones

absorbers the only way

An absorber is needed on the back wall; normal for SBA or active in DBA case. See DBA treads for details.

 

[Flaesh]

merely indicative

MFHF horns with direct radiating woofers, some thin absorbers\diffusers in room and painful for hearing low frequencies:

 

[ernestcarl]

MFHF horns with direct radiating woofers, some thin absorbers\diffusers in room and painful for hearing low frequencies:
View attachment 331930

Uhmmm... you said MFHF horns so does that mean there's not much low frequency energy/SPL to begin with?

Ah.... you did say direct radiating woofers... still, I'd rather see the full frequency response and/or IR information.

 

[Flaesh]

Uhmmm... you said MFHF horns so does that mean there's not much low frequency energy/SPL to begin with?

Ah.... you did say direct radiating woofers... still, I'd rather see the full frequency response and/or IR information.

Nothing good, peaks and deeps, almost no decay, PET jumps up and down. The owner has suspended work for the time being.
The amount of low-frequency energy is quite proportionate, but it sounds sad, and sometimes anecdotally contrasting.

 

[ernestcarl]

Nothing good, peaks and deeps, almost no decay, PET jumps up and down. The owner has suspended work for the time being.
The amount of low-frequency energy is quite proportionate, but it sounds sad, and sometimes anecdotally contrasting.

PET? I suppose you mean the peak energy time?

So I searched some much older files... but I could not find many D50 curves at 40% and below in a set of old IRs from gearspace/gearslutz I've previously saved.

Here are just some of the measurements with somewhat "okay" looking magnitude frequency response curves; however, they have rather obviously deficient clarity values in places:

Ex. A


Ex. B


Ex. C

 

[Flaesh]

peak energy time?

I thought this obvious in the context of the thread, but I deliberately wrote "jumping up and down"

about the shape of the graph.

PET also may be polyethylene terephthalate when discussing tweeters or room treatment. And then it can be confused with tweeter's or in room responce peak energy time

I could not find many D50 curves at 40% and below

Could there be some error in the measurements? For example, could too much noise be the cause? In any case, low frequencies sound very bad there.