k-Wave

1. PiesangBrood2000
   

   Member
   Joined: Jun '14
   Posts: 6

   Hi,

   I'm struggling to get the right values for medium.alpha_coeff and medium.alpha_power. I'm working from the "Attenuation of sound in air" table from
   http://www.kayelaby.npl.co.uk/general_physics/2_4/2_4_1.html.
   From their data I got an approximate fitting formula of y=5.3273*x^1.4484 (for 50% humidity) where y is the attenuation coefficient in [dB/km] and x is the frequency in [kHz].
   I assume since medium.alpha_coeff is measured in [dB/(MHz^y cm]) I cant use 5.3273 as is. Would 0.053273 be the correct value for medium.alpha_coeff in [dB/(MHz^y cm)]?

   Kind regards,
   PiesangBrood2000

   Posted 9 years ago #

2. bencox
   

   Developer
   Joined: Mar '10
   Posts: 285

   Hi PiesangBrood2000,

   To convert from dB/km to dB/m requires multiplying the prefactor by 0.001. Then to convert from dB/m/kHz^y to dB/m/MHz^y requires multiplying the prefactor by a further 0.001^y. So to convert your 5.3273 dB/km/kHz^y to dB/m/MHz^y as required by k-Wave, you need to multiply by 0.001^(y+1).

   So 5.3273 dB/km/kHz^y = 5.3273*0.001^(y+1) dB/m/MHz^y = 2.4x10^-7 dB/m/MHz^y.

   Hope that helps,
   Ben

   Posted 9 years ago #

3. PiesangBrood2000
   

   Member
   Joined: Jun '14
   Posts: 6

   Hi Ben,

   Thanks, that does help.
   In the k-Wave manual it gives the units for alpha_coeff as [dB/(MHz^y cm)] not [dB/(MHz^y m)], is this correct? If so then it would only be necessary to multiply the 2.4X10^-7 by a factor of 0.01 right?

   Kind regards,
   PiesangBrood2000

   Posted 9 years ago #

4. PiesangBrood2000
   

   Member
   Joined: Jun '14
   Posts: 6

   Also this value of 2.4X10^-7 for alpha_coeff seems very small, compared to the 0.75 used in the example problems. I assume the examples aren't for waves in air (sound speed of 1500 m/s) as well as being in the MHz range.

   Posted 9 years ago #

5. Bradley Treeby
   

   Developer
   Joined: Mar '10
   Posts: 1,640

   Hi PiesangBrood2000,

   Yes you're correct - the units in k-Wave should be dB/cm/MHz^y. To convert from dB/km to dB/cm you need to multiply by 1e-4. The absorption in air in the audible frequency range is very low compared the absorption in tissue in the MHz range. Most of the examples in the toolbox use parameters relevant to ultrasound in tissue.

   Hope that helps,

   Brad.

   Posted 9 years ago #

6. PiesangBrood2000
   

   Member
   Joined: Jun '14
   Posts: 6

   Hi Bradley,

   That clears things up, thanks a lot!

   Kind regards,
   PiesangBrood2000

   Posted 9 years ago #

7. NikiKeyz
   

   Member
   Joined: Feb '17
   Posts: 7

   Hi!
   It seems to be a mistake here.
   dB/km/kHz^y = 1e-5 dB/cm/kHz^y = 1e-5 dB/cm/(1e-3 MHz)^y
   y = 1.4484
   (1e-3)^1.4484 = 4.5165e-5
   So 5.3273*x^1.4484 dB/km/kHz^y = 1.1795*x^1.4484 dB/cm/MHz^y

   Posted 7 years ago #

8. v3p
   

   Member
   Joined: Jun '17
   Posts: 3

   Hi,

   How can I get the right values for medium.alpha_coeff and medium.alpha_power for:

   1) Glycerol, it article absorption equal to ~ 3000 10−15 neper m−1 Hz−2?
   2) Any kind of foam or porous material!?

   Thank you

   Posted 6 years ago #

9. v3p
   

   Member
   Joined: Jun '17
   Posts: 3

   In addition how can I change layer alpha coeff. if it should be a scalar? :/
   I am trying to simulate layered medium with different alpha coedd.

   Posted 6 years ago #

10. Bradley Treeby
    

    Developer
    Joined: Mar '10
    Posts: 1,640

    Hi v3p,

    medium.alpha_coeff can be defined as a matrix (see the heterogeneous medium examples included in the MATLAB help).

    To make your conversion:
    - multiply by 8.686 to convert from Nepers to dB
    - divide by 100 to convert from per m to per cm
    - multiply by 1e12 to convert from per Hz^2 to per MHz^2

    This gives:

    medium.alpha_coeff = 0.26;

    Remember to set medium.alpha_power to 2.

    Brad

    Posted 6 years ago #

11. v3p
    

    Member
    Joined: Jun '17
    Posts: 3

    Thank you Bradley!
    It was my mistake, as I understand now only alpha power should be a scalar.
    Might I have one more assist from you. In new simulation I tried to make heterogeneous medium where one layer is skull bone and second brain. i took data from this site http://www.kayelaby.npl.co.uk/general_physics/2_4/2_4_6.html

    When I convert units I received a huge number for alpha coefficient, for example:
    Brain: 0.068*10^2*8.686(db)*10^(6*1.3)(Hz)/100(cm)=~37*10^6...

    And second is possible to calculate alpha coefficient for skull when I know only alpha at certain frequency 9@3Mhz

    Posted 6 years ago #

12. Bradley Treeby
    

    Developer
    Joined: Mar '10
    Posts: 1,640

    The conversion should be:

    (0.068e2) Np / m / MHz^-b = (0.068e2 * 8.686 / 100) dB / cm / MHz^-b = (0.59) dB / cm / MHz^-b

    You will need to assume a value of alpha_power if you want to calculate alpha_coeff from the absorption at a single frequency.

    Posted 6 years ago #

13. alisiddiqi87
    

    Member
    Joined: Jul '20
    Posts: 1

    Hi Bradley,

    I think NikiKeyz is right.

    >>To convert from dB/km to dB/cm you need to multiply by 1e-4.
    It should be 1e-5.

    >>Then to convert from dB/m/kHz^y to dB/m/MHz^y requires multiplying the prefactor by a further 0.001^y
    I think it should be 'dividing' the prefactor by a further 0.001^y

    Thanks and regards,
    Ali

    Posted 3 years ago #

RSS feed for this topic

Reply

You must log in to post.