Permalink
Fetching contributors…
Cannot retrieve contributors at this time
665 lines (411 sloc) 21.7 KB

Transforming images on the fly

What you as an end-user of an Imbo installation will be doing most of the time, is working with images. This is what Imbo was originally made for, and this chapter includes details about all the different image transformations Imbo supports.

All image transformations can be triggered by specifying the t query parameter. This parameter must be used as an array so that you can provide several image transformations. The transformations will be applied to the image in the same order as they appear in the URL. Each element in this array represents a single transformation with optional parameters, specified as a string. If the t query parameter is not an array or if any of its elements are not strings, Imbo will respond with HTTP 400.

Below you will find all image transformations supported "out of the box", along with their parameters. Some transformations are rarely used with HTTP GET, but are instead used by event listeners that transform images when they are added to Imbo (HTTP POST). If this is the case it will be mentioned in the description of the transformation.

Auto rotate image based on EXIF data - t[]=autoRotate

This transformation will auto rotate the image based on EXIF data stored in the image. This transformation is rarely used per request, but is typically used by the :ref:`auto-rotate-image-event-listener` event listener when adding images to Imbo.

Examples:

  • t[]=autoRotate

Blur the image - t[]=blur

This transformation can be used to blur the image.

Parameters:

mode

The blur type (optional). Defaults to gaussian. Possible values are:

gaussian
When adding gaussian blur, the radius and sigma parameters are required.
adaptive
When adding adaptive blur, the radius and sigma parameters are required. Adaptive blur decrease the blur in the part of the picture near to the edge of the image canvas.
motion
When adding motion blur, the radius, sigma and angle parameters are required.
radial
When adding radial blur, the angle parameter is required.
radius
The radius of the Gaussian, in pixels, not counting the center pixel.
sigma
The standard deviation of the Gaussian, in pixels.
angle
The number of degrees to rotate the image.

Examples:

  • t[]=blur:radius=1,sigma=2
  • t[]=blur:type=adaptive,radius=2,sigma=4

Add an image border - t[]=border

This transformation will apply a border around the image.

Parameters:

color
Color of the border in hexadecimal. Defaults to 000000 (You can also specify short values like f00 (ff0000)).
width
Width of the border in pixels on the left and right sides of the image. Defaults to 1.
height
Height of the border in pixels on the top and bottom sides of the image. Defaults to 1.
mode
Mode of the border. Can be inline or outbound. Defaults to outbound. Outbound places the border outside of the image, increasing the dimensions of the image. inline paints the border inside of the image, retaining the original width and height of the image.

Examples:

  • t[]=border
  • t[]=border:mode=inline
  • t[]=border:color=000
  • t[]=border:color=f00,width=2,height=2

Expand the image canvas - t[]=canvas

This transformation can be used to change the canvas of the original image.

Parameters:

width
Width of the surrounding canvas in pixels. If omitted the width of <image> will be used.
height
Height of the surrounding canvas in pixels. If omitted the height of <image> will be used.
mode
The placement mode of the original image. free, center, center-x and center-y are available values. Defaults to free.
x
X coordinate of the placement of the upper left corner of the existing image. Only used for modes: free and center-y.
y
Y coordinate of the placement of the upper left corner of the existing image. Only used for modes: free and center-x.
bg
Background color of the canvas. Defaults to ffffff (also supports short values like f00 (ff0000)).

Examples:

  • t[]=canvas:width=200,mode=center
  • t[]=canvas:width=200,height=200,x=10,y=10,bg=000
  • t[]=canvas:width=200,height=200,x=10,mode=center-y
  • t[]=canvas:width=200,height=200,y=10,mode=center-x

Compress the image - t[]=compress

This transformation compresses images on the fly resulting in a smaller payload. It is advisable to only use this transformation in combination with an image type in the URL (for instance .jpg or .png). This transformation is not applied to images of type image/gif.

Parameters:

level
The level of the compression applied to the image. The effect this parameter has on the image depends on the type of the image. If the image in the response is an image/jpeg a high level means high quality, usually resulting in larger files. If the image in the response is an image/png a high level means high compression, usually resulting in smaller files. If you do not specify an image type in the URL the result of this transformation is not deterministic as clients have different preferences with regards to the type of images they want to receive (via the Accept request header).

Examples:

  • t[]=compress:level=40

Change image contrast - t[]=contrast

This transformation can be used to change the contrast of the colors in the image.

Parameters:

alpha
Used to adjust the intensity differences between the lighter and darker elements of the image. Can also be negative. Note: this parameter was named sharpen in Imbo 1.x.
beta
Where the midpoint of the gradient will be. This value should be in the range 0 to 1. Default: 0.5.

Examples:

  • t[]=contrast:alpha=3

Note

If you are getting different results than expected when using negative alpha values, your imagick extension is probably built against an old version of ImageMagick.

Convert the image type - .jpg/.gif/.png

This transformation can be used to change the image type. It is not applied like the other transformations, but is triggered when specifying a custom extension to the <image>. Currently Imbo can convert to:

  • image/jpeg
  • image/png
  • image/gif

Examples:

  • curl http://imbo/users/<user>/images/<image>.gif
  • curl http://imbo/users/<user>/images/<image>.jpg
  • curl http://imbo/users/<user>/images/<image>.png

Crop the image - t[]=crop

This transformation is used to crop the image.

Parameters:

x
The X coordinate of the cropped region's top left corner.
y
The Y coordinate of the cropped region's top left corner.
width
The width of the crop in pixels.
height
The height of the crop in pixels.
mode

The crop mode (optional). Possible values are:

center
When using the center mode the x and y parameters are ignored, and the center of the cropped area is placed in the center of the original image.
center-x
Center the crop on the x-axis. Use the y parameter to control the upper edge of the crop.
center-y
Center the crop on the y-axis. Use the x parameter to control the left edge of the crop.

Examples:

  • t[]=crop:x=10,y=25,width=250,height=150
  • t[]=crop:width=100,height=100,mode=center
  • t[]=crop:width=50,height=50,mode=center-x,y=15
  • t[]=crop:width=50,height=50,mode=center-y,x=15

Make a gray scaled image - t[]=desaturate

This transformation desaturates the image (in practice, gray scales it).

Examples:

  • t[]=desaturate

Draw points of interest - t[]=drawPois

This transformation will draw an outline around all the POIs (points of interest) stored in the metadata for the image. The format of the metadata is documented under the :ref:`smartSize <smartsize-transformation>` transformation.

Parameters:

color
Color of the border in hexadecimal format. Defaults to ff0000 (You can also specify short values like f0f (ff00ff)).
borderSize
Width of the border in pixels. Defaults to 2.
pointSize
The diameter (in pixels) of the circle drawn around points of interest that do not have a height and width specified. Defaults to 30.

Examples:

  • t[]=drawPois
  • t[]=drawPois:borderSize=10
  • t[]=drawPois:color=0f0
  • t[]=drawPois:color=00f,borderSize=10,pointSize=100

Note

This transformation has a bug/limitation: all coordinates are based on the original image. In other words, applying this at the end of a transformation chain which resizes/crops/rotates the image can lead to unexpected results. This will hopefully change in the future.

Make a mirror image - t[]=flipHorizontally

This transformation flips the image horizontally.

Examples:

  • t[]=flipHorizontally

Flip the image upside down - t[]=flipVertically

This transformation flips the image vertically.

Examples:

  • t[]=flipVertically

Generate a histogram of the image - t[]=histogram

This transformation will convert the image into a histogram of the image itself, with a size of 256x158 pixels. The size of the generated image can be overridden by using one or more of the supported parameters.

Parameters:

scale
The amount to scale the histogram. Defaults to 1.
ratio
The ratio to use when calculating the height of the image. Defaults to 1.618.
red
The color to use when drawing the graph for the red channel. Defaults to #D93333.
green
The color to use when drawing the graph for the green channel. Defaults to #58C458.
blue
The color to use when drawing the graph for the blue channel. Defaults to #3767BF.

Examples:

  • t[]=histogram
  • t[]=histogram:scale=2
  • t[]=histogram:red=f00,green=0f0,blue=00f

Adjust levels of the image - t[]=level

This transformation will adjust the levels of an image. You are able to specify individual channels to adjust - by default it will apply to all channels.

Parameters:

channel
The channel to adjust. r (red), g (green), b (blue), c (cyan), m (magenta), y (yellow), k (black) and all (all channels) are available values. These channels can also be combined, if multiple channels should be adjusted. Defaults to all.
amount
The amount to adjust by. Range is from -100 to 100. Defaults to 1.

Examples:

  • t[]=level
  • t[]=level:channel=r,amount=30
  • t[]=level:channel=rg,amount=-45

Enforce a max size of an image - t[]=maxSize

This transformation will resize the image using the original aspect ratio. Two parameters are supported and at least one of them must be supplied to apply the transformation.

Note the difference from the :ref:`resize <resize-transformation>` transformation: given both width and height, the resulting image will not be the same width and height as specified unless the aspect ratio is the same.

Parameters:

width
The max width of the resulting image in pixels. If not specified the width will be calculated using the same aspect ratio as the original image.
height
The max height of the resulting image in pixels. If not specified the height will be calculated using the same aspect ratio as the original image.

Examples:

  • t[]=maxSize:width=100
  • t[]=maxSize:height=100
  • t[]=maxSize:width=100,height=50

Modulate the image - t[]=modulate

This transformation can be used to control the brightness, saturation and hue of the image.

Parameters:

b
Brightness of the image in percent. Defaults to 100.
s
Saturation of the image in percent. Defaults to 100.
h
Hue percentage. Defaults to 100.

Examples:

  • t[]=modulate:b=150
  • t[]=modulate:b=120,s=130,h=90

Make a progressive image - t[]=progressive

This transformation makes the image progressive.

Examples:

  • t[]=progressive

Resize the image - t[]=resize

This transformation will resize the image. Two parameters are supported and at least one of them must be supplied to apply the transformation.

Parameters:

width
The width of the resulting image in pixels. If not specified the width will be calculated using the same aspect ratio as the original image.
height
The height of the resulting image in pixels. If not specified the height will be calculated using the same aspect ratio as the original image.

Examples:

  • t[]=resize:width=100
  • t[]=resize:height=100
  • t[]=resize:width=100,height=50

Rotate the image - t[]=rotate

This transformation will rotate the image clock-wise.

Parameters:

angle
The number of degrees to rotate the image (clock-wise).
bg
Background color in hexadecimal. Defaults to 000000 (also supports short values like f00 (ff0000)).

Examples:

  • t[]=rotate:angle=90
  • t[]=rotate:angle=45,bg=fff

Apply a sepia color tone - t[]=sepia

This transformation will apply a sepia color tone transformation to the image.

Parameters:

threshold
Threshold ranges from 0 to QuantumRange and is a measure of the extent of the sepia toning. Defaults to 80

Examples:

  • t[]=sepia
  • t[]=sepia:threshold=70

Sharpen the image - t[]=sharpen

This transformation can be used to change the sharpness in the image.

Parameters:

radius
The radius of the Gaussian operator in pixels. Defaults to 2.
sigma
The standard deviation of the Gaussian, in pixels. Defaults to 1.
gain
The percentage of the difference between the original and the blur image that is added back into the original. Defaults to 1.
threshold
The threshold in pixels needed to apply the difference gain. Defaults to 0.05.
preset

Different presets that can be used. The presets are:

  • light (radius = 2, sigma = 1, gain = 1, threshold = 0.05)
  • moderate (radius = 2, sigma = 1, gain = 2, threshold = 0.05)
  • strong (radius = 2, sigma = 1, gain = 3, threshold = 0.025)
  • extreme (radius = 2, sigma = 1, gain = 4, threshold = 0)

When using any of the presets the different parameters can be overridden by specifying radius, sigma, gain and/or threshold. Not specifying any parameters at all is the same as using the light preset.

Examples:

  • t[]=sharpen
  • t[]=sharpen:preset=light (same as above)
  • t[]=sharpen:preset=extreme,gain=10 (use the extreme preset, but use a gain value of 10 instead of 4)
  • t[]=sharpen:radius=2,sigma=1,gain=1,threshold= 0.05 (same as using t[]=sharpen:preset=light, or simply t[]=sharpen)

Smart size the image - t[]=smartSize

This transformation is used to crop the image based on a point of interest (POI) provided either as a transformation parameter or from the image metadata.

Metadata format

The smart size transformation supports reading the POI from the metadata of the image. The POI information is expected to be stored on the poi property in metadata. Below is an example of a valid metadata object containing a 600,240 POI:

{
  "poi": [
    {
        x: 600,
        y: 240
    }
  ]
}

Note

The smart size transformation currently takes only the first object into account when cropping the image, but the POIs is stored as an array of objects in order to be easy to expand with more information for a more sophisticated smart size algorithm in the future.

Parameters:

width
The width of the crop in pixels.
height
The height of the crop in pixels.
poi
The POI coordinate x,y to crop around. The parameter is optional if the POI exists in metadata.
crop

The closeness of the crop (optional). Possible values are:

close medium wide

Examples:

  • t[]=smartSize:width=250,height=250,poi=300,200
  • t[]=smartSize:width=250,height=250,poi=300,200,crop=close

Strip image properties and comments - t[]=strip

This transformation removes all properties and comments from the image. If you want to strip EXIF tags from the image for instance, this transformation will do that for you.

Examples:

  • t[]=strip

Create a thumbnail of the image - t[]=thumbnail

This transformation creates a thumbnail of <image>.

Parameters:

width
Width of the thumbnail in pixels. Defaults to 50.
height
Height of the thumbnail in pixels. Defaults to 50.
fit
Fit style. Possible values are: inset or outbound. Default to outbound.

Examples:

  • t[]=thumbnail
  • t[]=thumbnail:width=20,height=20,fit=inset

Create a vertical mirror image - t[]=transpose

This transformation transposes the image.

Examples:

  • t[]=transpose

Create a horizontal mirror image - t[]=transverse

This transformation transverses the image.

Examples:

  • t[]=transverse

Add a vignette to the image - t[]=vignette

This transformation can be used to add a vignette to the image.

Parameters:

inner
Color at the center of the image, in hexadecimal. Defaults to none, which means transparent. (You can also specify short values like f00 (ff0000)).
outer
Color at the edge of the image, in hexadecimal. Defaults to 000.
scale
Scale factor of the vignette. 2 will create a vignette twice the size of the original image. Defaults to 1.5.

Examples:

  • t[]=vignette
  • t[]=vignette:outer=ccc
  • t[]=vignette:scale=1,outer=333

Add a watermark to the image - t[]=watermark

This transformation can be used to apply a watermark on top of the original image.

Parameters:

img
Image identifier of the image to apply as watermark. Can be set to a default value in configuration by using <setDefaultImage>.
width
Width of the watermark image in pixels. If omitted the width of <img> will be used.
height
Height of the watermark image in pixels. If omitted the height of <img> will be used.
position
The placement of the watermark image. top-left, top-right, bottom-left, bottom-right and center are available values. Defaults to top-left.
x
Number of pixels in the X-axis the watermark image should be offset from the original position (defined by the position parameter). Supports negative numbers. Defaults to 0
y
Number of pixels in the Y-axis the watermark image should be offset from the original position (defined by the position parameter). Supports negative numbers. Defaults to 0
opacity
Can be an integer between 0 and 100 where 0 is fully transparent, and 100 is fully opaque. Defaults to 100

Examples:

  • t[]=watermark:img=f5f7851c40e2b76a01af9482f67bbf3f
  • t[]=watermark:img=f5f7851c40e2b76a01af9482f67bbf3f,width=200,x=5
  • t[]=watermark:img=f5f7851c40e2b76a01af9482f67bbf3f,height=50,x=-5,y=-5,position=bottom-right,opacity=50

If you want to set the default watermark image you will have to do so in the configuration:

<?php
return [
    // ...

    'eventListeners' => [
        'watermark' => function() {
            $transformation = new Imbo\Image\Transformation\Watermark();
            $transformation->setDefaultImage('some image identifier');

            return $transformation;
        },
    ],

    // ...
];

When you have specified a default watermark image you are not required to use the img option for the transformation, but if you do so it will override the default one.