Created by sebastien.popoff on 28/04/2013


From the bimodal distribution to the quarter circle law

[A. Goetschy and A. D. Stone, Phys. Rev. Lett., 1304.5562, (2013)]

Almost thirty years ago, theoreticians predicted that the distribution of the transmission values of a multiple scattering sample should follow a 'bimodal distribution'.  Physically, that means that, in the diffusive regime, there is a large number of strongly reflected channels - the closed channels - and a small number of channels of transmission close to one - the open channels. The existence of these open channels regardless of the thickness of the medium is of big interest for researchers, especially for imaging or communication applications. Nevertheless, such channels have not yet been directly observed. A investigation on those channels requires a measurement of the entire transmission matrix of a lossless scattering medium. For practical reasons (open geometry, limited numerical aperture, noise...) one usually has access to a subpart of the total transmission matrix. In recent experimental measures of the transmission matrix in optics [S.M. Popoff et al., Phys. Rev. Lett., 104, 100601, (2010)] the distribution of the transmission values follows a 'quarter circle law', characteristic of totally uncorrelated systems. This means that the fraction of the transmission matrix measured shows no effect of the correlations at the origin of the bimodal distribution due to the loss of information. In this paper, A. Goetschy and D. Stone theoretically study the effect of the loss of information or the imperfect control on the statistics of the transmission matrix of a scattering system.

See full post
Created by sebastien.popoff on 26/04/2013


Retrieving an optical scale resolution with light focusing guided by ultrasound

[B. Judkewitz et al., Nat. Photon., 7, 300, (2013)] 

To focus light in or through a scattering medium using wavefront shaping techniques, one needs a way to probe the intensity or the field at the target position. To avoid having to insert a probe in the medium, Xu et al. proposed in 2011 the use of an ultrasonic focused beam to select a target area by photo-acoustic effect [X. Xu, H. Liu and L.V. Wang, Nat. Photon., 5, 154, (2011)]. This technique allows focusing light on a spot of the size of the ultrasound focused beam, which is typically at least one order of magnitude larger than the optical wavelength. In this new study, B. Judkewitz and co-authors used an innovative method to be able to focus light on a much smaller scale.

See full post
Created by sebastien.popoff on 24/04/2013


A pioneering experiment: Focusing through scattering media using wavefront shaping

[I.M Vellekoop and A.P. Mosk, Opt. Lett., 15, 2309, 2007]

In 2007 I.M. Vellekoop and A.P. Mosk published their work on the first demonstration of focusing light through a highly scattering medium. Most techniques to image or focus through scattering media relied on selecting only the part of the light that has not been scattered - the ballistic light. The ballistic signals decay exponentially with the thickness of the medium, limiting drastically the depth at which light can be focused. The idea developed by the authors is to use the scattered waves, that are randomly mixed, to focus light through the medium. A scattering sample illuminated by a coherent wave gives rise to a so-called speckle pattern, that results from the interference of the scattered waves. Using a spatial light modulator (SLM), the authors are able to control independently the phase of the different parts of the incident beam. Each segment gives an output seemingly random complex field. By testing different values of the phase for each segment, they are able to put in phase all the contributions, giving rise to a very bright focus spot.

See full post
Created by sebastien.popoff on 18/04/2013

Tutorials Spatial Light Modulators

How to control an SLM with Matlab/Octave using Psychtoolbox

Most spatial light modulators (SLMs) available are controllable like a normal computer monitor and are plugged on a computer with a DVI cable.  Some SLMs are now sold with a dedicated card or can be controlled via USB. If you possess such a device, this tutorial is not for you. The first requirement to control the SLM with a DVI/HDMI cable is to have a graphic card with two monitor outputs, one for your screen, one for your SLM. Once plugged to the computer, the SLM is then handled by the operating system as a secondary monitor. No software is required to display an image on the SLM. For that reason, the constructor does not provide any code to use the SLM with Matlab/Octave or other software. One solution to send images with Matlab is to display an array in a figure that fits the size of the secondary monitor. Nevertheless, this technique presents some drawbacks due to the fact that you do not control directly the pixels of the SLM. For instance, the border of the figure, which may be different depending on the operating system, has to be taken into account. More importantly, the scaling of the figure does not guarantee that one pixel of the image displayed corresponds to one pixel of the SLM. For application where a very good resolution is needed, a blurred image on the SLM can be detrimental.

I present how to control directly the pixels of the SLM using Psychtoolbox, a free toolbox for Matlab and Octave that uses GPU acceleration. I show here a tutorial for Matlab, but the toolbox also exists for Octave and seems to work a similar way.

See full post