Top: result after subtracting the mean background and using the PynPoint PCA algorithm to subtract the speckle noise with 17 principal components. Bottom: result after using the PCA algorithm to subtract the background with 40 principal components and PynPoint to subtract the speckle noise with 15 principal components. The inner 0.3′′ were masked for the speckle subtraction. We convolved both images with a 0.5λ/D width Gaussian to improve the visibility of the planet signals for this figure. The images shown here are typical results for both background subtraction schemes, where all planets can be identified, but they do not necessarily exhibit the highest achievable S/N for all planets. The scale is in arbitrary linear units.