25% (~900x600) 33% (~1200x800) 50% (~1800x1200) .

Data - Information - Digital Proc. - Other pics

Scanned with a Polaroid SS35 at 2700dpi and processed with Pleiades PixInsight v1.0.0.50b

[mouseover]

Comparason with the final result

Aligned and cropped the image, I averaged them. The result is shown above.

[mouseover]

Here is a comparison between the result of substracting a AV mask generated throught DBE and the "raw" image. Move the mouse over image to see the original image.

After adjusting the dinamical range's end points, a first curves adjustement corrected the overall color balance.

A PIP transform increased the contrast in the shadows and middtones.

Sin reducción v/s Wavelets -- Wavelets v/s Deconvolución --
Deconvolución v/s SGBNR_L -- SGBNR_L v/s SGBNR_ab
SGBNR_ab v/s SCNR -- Sin reducción v/s SCNR
[mouseover]

The different steps attacks diferent scales of the noise in the image. A Trous Wavelets fights the hot pixels and the high frecuency noise, in all the RGB channels. The deconvolution regenerates the shape of the faint stars. SGBNR removes first the uneven illumination, acting over the L channel. Then, applied to the a and b channels, it removes the chromatic differences. Finally, a mild SCNR to the green channel decreases the green cast and some halos in the stars.

You can read the Processing Report of the noise reduction here: noise.rtf

Also here is another sample of NR, a PSD image: pleiades.rar
(it was another, but succesfull procedure)

Here is the result of the previous SCNR application.

As usual, a histograms and a curves adjustements are needed to recover the color balance and use all of the dinamical range. Also a mild SCNR was applied to the blue channel.

A last PIP transform revealed more hidden nebulosities.

Finally, a curves adjustement to the image generated the final appearance of it (including a saturance incrementation).

The last step was a USM to recover definition in the edges. It was a very soft one.

.