|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
In Reply to: Sensitivity Hypothesis! Correct me if I'm wrong............ posted by Eric U. on May 08, 1999 at 18:12:49:
It is relativly simple of course.Say you have a typical speaker with a mid-band impedance of 5 ohms(a very common value).And you have a 1.5mH inductor in series with it(another common value).Let us say the manufacturer used the 'standard line' from Madisound(about what most OEM's do)the DCR would be .75 ohms,cost=$3.50.And you buy the 'Sledgehammer' inductor from Madisound the DCR would be .114 ohms,cost=$12.00.This forms a voltage divider.With the first example you have lost 13.1% of your voltage across the DCR.With the second you have lost 2.2% of your voltage.A 10.9% reduction in voltage is exactly 1dB.Since mtry and the ABX crowd think a .1dB match is needed so as not to influence the subject by the volume difference I think we can safely assume that 1dB would be audible.Changing capacitor types can have an even larger effect on mid and high frequency balance.A friend of mine replaced all the capacitors in his Thiel networks and the speakers sounded completely different and required re-voicing,not only pad changes but capacitor value changes too(mainly decreasing values).
However, the audibility of 1db difference would also depend at what frequency it takes place and what amount of the bandwidth is affected as your threashold to differentiate volume differences is not linear across the frequency bandwidth nor the amount of the bandwidth affected. But you knew this, so even this may not be audible at the right frequencies.
I am tempted to give you partial credit on your point.If you look at equal loudness contours you will see that at low frequencies a 1dB change is more perceptable than it is at higher frequencies(the curves are closer together at low freq and further apart at high freq).I once made a trade-off in an OEM design of a woofer that increased the network cost by $2 but saved over $20 off the cost of the magnetic circuit on the woofer.The difference at retail is usually 6x parts cost so it allowed the retail price to be $200 less on a pair($900 vs $1100).
This goes beyond the equal loudness curve as that compares the loudness level perception to a standard frequency of 1khz. While the perceived threashold for a difference may be less at the lower frequencies compared to the higher frequencies, the band width that this difference occurrs is as important in tha a broadband difference is more noticable than a narrow band difference where a larger difference is needed.
Perhaps the David Clark paper, 'High Resolution Subjective Testing Using a Double-Blind Comparator,' JAES May 82, pg 330-338, fig 2, will shed some light on this.
I just was letting him be also aware that the Speakers "Qes" is proportional to "Re/BL^2"but in reality the "Re" includes the "Rg" (resistance of everything between the amp and the woofer). Thereby:
Qes = c1*(Re + Rg)/BL^2
Qts = Qms /( 1 + Qms*BL^2/c1*(Re + Rg) )Given a generic woofer:
BL=10
Qes=1/2
Re=5
Qms=10Rg = .114 translates into a 2% increase in Qts
Rg = .75 translates into a 14% increase in QtsThis may be enough to significantly change the required box size & other sonic characteristics.
It is more ominous than you think.In a vented box the size goes with the square of Qts so if the Qts is .314 for .114DCR it will be .35 for .75DCR in this example squaring the Qts gives us an increase of 25% in box size to keep the same bass response.Also consider what happens when you go from an Rg=.05 ohms(solid state)to Rg=.6 ohms(pentode)or Rg=1.2 ohms(triode).You may have to double the box size(and people wonder why tubes sound different).
This post is made possible by the generous support of people like you and our sponsors: