php - PHP/Imagick/PDFlib Flop Image Changes its Bit Depth

Question

I am Having PNG Image And Trying To Flop (Mirror) by imagick function of php It Gets Flop Exactly But The

Base Image is In Format 24 Bit RGB

and after Convertion It Gets To

8 Bit Pallated

. So the Main Problem is that when I use to place both images in my pdflib pages one of the image(converted) displays curly.... Original Image Output After Flop(Mirror) by Imagick and Rendered in PDFlib ->

My Code Is Simple ---->

$im = new Imagick($background_image);
$im->flopImage();
$im->writeimage($background_image."_flop.png");

---

Modified Date => 29 Oct 2013 Original Image -> Size 4.68 KB Bit Depth 32 Flopped Image -> Size 7.99 KB Bit Depth 64 Automatically Changes It's Properties ORIGINAL

Converted

Answer 1

Imagick is using the smallest format possible to save the image. Saving in these formats all produce the same image but have the sizes:

• Palette - 3.38kB
• RGBA 32bit - 6.14kB
• RGBA 64bit - 8.09kB

Saving to the smallest possible file is usually what people desire. However you can disable this in a couple of ways.

You can tell Imagick to use the same PNG format as the source image by setting the png:format option to png00. e.g.

$imagick = new Imagick(realpath("../images/FlopOriginal.png"));
$imagick->flopImage();
$imagick->setOption('png:format', 'png00');
$imagick->writeImage("../images/Flop.png");

The full options for png:format are png8, png24, png32, png48, png64, and png00.

Alternatively you can explicitly set the image format to use when saving the PNG file, through the png:bit-depth and png:color-type e.g.

$imagick = new Imagick(realpath("../images/FlopOriginal.png"));
$imagick->flopImage();
$imagick->setOption('png:bit-depth', '8');
$imagick->setOption('png:color-type', 6);
$imagick->writeImage("../images/Flop.png");

The color type values come from the libpng.h and are:

PNG_COLOR_TYPE_GRAY         0
PNG_COLOR_TYPE_RGB          2
PNG_COLOR_TYPE_PALETTE      3
PNG_COLOR_TYPE_GRAY_ALPHA   4
PNG_COLOR_TYPE_RGB_ALPHA    6

Both those methods produce a flopped image that is RGBA 32bit like the original image.