How Much is 3db?

[jjptkd]

So in another thread there is debate over whether a 3db gain is even noticeable. If you look at the specs of two of my favorite Klipsch speakers, the Chorus II and the forte II it states that the max output of the Chorus II is 121db and the forte II is 119db. Anyone who has ever tested these side by side (I have many times) and really put the screws to them can tell you that the Chorus is quite noticeably louder with only a rated 2 db difference in max output.  Another case in point is the RSW-15 vs the RSW-12. Again, rated specs are 121db for the 15 and 119db for the 12 and the 15 is very noticeably louder than the 12. It seems in these two instances where there is only supposed to be a 2db difference in max output the differences are quite clear, am I missing something?

[jjptkd]

Ok so thinking about this a little more in my forte II / A-55g mid experiment when I measured the output of the stock driver compared to the new driver I found about a 3db difference. Audibly the mids of the lower output driver seemed softer and slightly recessed when compared to the stock driver, admittedly not by a lot but enough to notice, especially when side by side comparisons were done.

 

So what I've read on this is that 3db works pretty much on a sliding scale based on a percentage of total output. Going from 50db to 53db would be considerably more noticeable than going from 100db to 103db, which makes sense. It still does not explain the very noticeable differences between the speakers / subs listed in my OP unless I'm missing some other factor.

[derrickdj1]

Driver size, dispersion, polar response all come into play in what you are hearing.  A 3 db increase mean you double the sound energy physically.  Now what is the perception of the 3 db. increase.  This gets into how we hear certain sounds and frequencies.  For the majority of people, a 3 db increase is barely perceptible.  A 10X increase is perceived as twice as loud on average by the majority of people.

[Chris A]

It probably would be beneficial to read on the subject of "equal loudness contours" which is really what you're mixing with the idea of "dB". 

 

If you talk instead in terms of "phons" or the older "sones" scale instead of just straight SPL, then you'll find that these two measurement scales attempt to match perceived loudness to actual SPL, but it's a function of frequency and the multiplicity of tones (frequencies) present at the same time.

 

Chris

[Schu]

Im not understanding the conversation... if we had a discussion about individual drivers in relation to each other, I would understand.

[Chris A]

There is a difference between playing a single tone, band-passed pink noise, or wide-band pink or white noise in terms of perception and relative SPL (in dB).

[jjptkd]

I can understand why there would be a larger perceived difference between the Chorus and forte as I believe the mid frequencies, those most sensitive to our ears, are turned up a few notches on the Chorus so even if the differences are slight they are in the band that we would tend to notice more, at least according to the link which in fact matches what I have observed. That does not explain the perceived difference in the two subwoofers that basically play the same exact frequencies.

[derrickdj1]

Three db. is a lot in terms of power.  For example, to go from 3 to 6 db is double the power.  Chris mentioned the equal loudness curve which explain a lot.  It is much easier to hear frequencies in the 2 kHz area than in the 20 Hz area.  The 20 Hz sound/tone needs a large boost to be preceived as the same loudness of the 2 kHz tone.

[Chris A]

Look closely at the equal loudness contours...

 

 

Very small changes in SPL at very low frequencies sound much different in terms of the perceived loudness (...notice how the equal loudness curves are very close together at low frequencies).   Floyd Toole also pointed this out in a recent talk online.  It's also a factor in unmastering music.  You must be extremely careful when you get to frequencies below 100 Hz to get the SPL right or you'll hear either too much bass or too little.  But you may be dealing with only a dB or two in relative loudness (SPL that is...the type that's measured by a hand-held meter). 

[Chris A]

Small changes in dBSPL at these frequencies are perceived to be much bigger in terms of their phon-equivalent loudness.

[Chris A]

It's also why I'm a bit more vocal than perhaps others when I see large attenuation of bass during mastering of recordings.  Very small changes in bass SPL below 100 Hz leads to extreme changes in perceived loudness.

 

We have a generation or two (or three) of people who have only listened to bass-attenuated recordings--instead of the real thing on non-amplified instrumentation--that have become used to extremely weak bass performance of classic music compositions, I've found.

 

If you go listen to the real thing...in real life...it's a whole lot different than most recordings.  Restoring that bass is a big, big deal, and requires a bit of trial and error to get it right.  It would be nice if the mastering engineers would simply leave that area of the audible spectrum alone...

 

Chris

[derrickdj1]

Some good questions have been ask.  This is a lot more interesting than the usual what speaker should I get in a lot of threads.

[JohnA]

Generally, 3dB is just noticeable in a noisy environment and a pretty easy to notice change in a quiet environment, but not huge.   Getttta sound meter and play with it. You can download one for your phone.  I don't know if Rat Shack still sells one. 

[Ski Bum]

I think it makes a bit more sense to attribute the differences in output to the sensitivity of the speakers, rather than their max output ratings, which are just the point where some specific, measurable amount of thermal compression occurs.  Not many of us flog our speakers that hard, and if we do, it's time to step up to some Cinema speakers.  

 

So, the respective sensitivities of the fortes and chorus are ~95 db/w and ~98 db/w (based on the limiting factor of each, the sensitivity of the woofs).  With twice as much power into the fortes, making up that 3db difference, and they should play approximately the same spl level.  They'll dig slightly deeper, too.

 

This doesn't account for changes such as the modding the mids to lower their level, at frequencies where our ears are extremely sensitive (small changes are quite audible), or bass extension, where our ears are far less sensitive.  That F-M curve aspect is involved in all of this, but it seems to me that it's more relevant to proper equalization given the normal listening level (if you listen at quiet levels most of the time, you may want a bit of F-M type eq to preserve a natural sounding tonal balance; if you tend to blast it regularly, you won't need the additional bass boost for a natural tonal balance). 

[Chris A]

I've had more than one director tell the orchestra or wind symphony that the double bass, tuba and low brass sections really never get to play soft--instead it's various levels of "loud".  You can see why in the above curves.

 

Getting back to the subject at hand: when comparing subwoofers and their loudness levels, it's something that you should consider carefully.  Undersizing subwoofers for the rooms that they're in is a prime reason why you see so many instances of "my subwoofer doesn't work anymore" threads--because they basically play very loud all the time.  It's better to get 2 or 3 subwoofers that have significant headroom at the 100-120 dB level in-room.  Horn-loaded subwoofers do have a reason for being there--even in small rooms.

 

Paul Klipsch wasn't just joking around when he spent so much time on the Klipschorn bass bin design.  It was a big deal back in the 1940s to be able to reproduce--cleanly--bass below 100 Hz, and it still is...

 

Edit: I would recommend to Klipsch to improve their product line when it comes to smaller form factor horn-loaded subwoofers that can play below 20 Hz for HT LFE reproduction.  Direct radiating subwoofers (all of them) produce too much modulation, harmonic, compression, and phase/group delay distortion at those frequencies: horn loading is really required when you get below 30-40 Hz.  You really do need low moving mass and as high efficiency at those frequencies as you can get to be able to refer to extreme low frequency reproduction as "hi-fi".

 

Right now, Danley owns that marketplace (even though their prices are too high), along with for-profit enterprises that are selling plans for DIY. Bill Fitzmaurice comes to mind--but his designs are typically less wieldy than the tapped horn (TH) designs by Danley.

 

Chris

[chriswhotakesphotos]

This reminds me of how a little while ago (I think in Grizzog's H3/RF7-II/Chrous comparison thread) I'd remarked that I thought the Heresies (mine are KP-201s) were noticeably more sensitive than my RP-280Fs. They measure to be within 1 or 2 db of each other, but every time I've switched back and forth I've found the KPs to be louder. Maybe it's the difference in frequency response; they have a much stronger midrange to my ear, which is what the ear hears most of. Interesting!

[wvu80]

1 hour ago, Chris A said:

There is a difference between playing a single tone, band-passed pink noise, or wide-band pink or white noise in terms of perception and relative SPL (in dB).

 

I have a set of DIY speakers that Audyssey measures -3 db on one side of the L/R.  I can definitely hear a clear difference and I think 99% would also hear a difference.  My wife constitutes the other 1% of people who couldn't tell. 

[Peter P.]

12 hours ago, derrickdj1 said:

" Now what is the perception of the 3 db. increase.  This gets into how we hear certain sounds and frequencies.  For the majority of people, a 3 db increase is barely perceptible.  A 10X increase is perceived as twice as loud on average by the majority of people."

This is correct, and also true in my field, RF electronics.

[Quiet_Hollow]

On 11/29/2016 at 7:19 AM, jjptkd said:

So in another thread there is debate over whether a 3db gain is even noticeable.

No surprise. 3 dB of what?

 

The Decibel is a ratio, like a percent. To express anything in decibels, reference to the physical quantity must also be included. Such as:

dB SPL

dB W

dB m

dB V

 

Otherwise, it's simply a unit with no context.

 

If we assume SPL is the context, then yes 3 dB SPL is quite noticeable as a 0.5-1.0 dB change in SPL (at any audible frequency) is typically the threshold of perceptible change. You can experiment this with yourself using test tones and a typical stereo so long as the volume is calibrated in dB.

 

Things get muddy in conversation when there's a failure to respect the relationships ...ie. dB SPL to dB W.   Just because a dB symbol is hanging out there, doesn't mean the quantities are the same, or ever should be. Loudness, power, volts, intensity, are all very unique. They can be expressed as a relationship to one another using the decibel, but are not the same.

 

On 11/29/2016 at 7:19 AM, jjptkd said:

It seems in these two instances where there is only supposed to be a 2db difference in max output the differences are quite clear, am I missing something?

Yes. There's a lot more to hearing than just one number. Acoustics, as a topic of study, is comprehensive...too much to cover on any forum.

[mustang guy]

It never fails to amaze me just how complex the answers to simple questions about audio can be. I read through the responses, and I was prepared to stop reading after looking at the graphs Chris pointed to. Everything from the rule of thumb answers to the logarithmic scales is pertinent here. To have more precise answers, the question needs to be much more complex and specific. Perhaps simple questions are simply rhetoric questions. 

 

How do you ask a question that yields lower subjective response and more measurable and objective response? By asking very specific questions. How about this one taken right from the OP:

 

I have A/B compared the Chorus and Forte II. Why is the Chorus quite noticeably louder with only a rated 2 db difference in max output than the Forte II? 

 

The most likely and obvious answer to that specific question is that the Chorus is more efficient by 2 db.