When ImageCR needs to resize an image to a different resolution, it needs to interpolate the old image grid to a new one. This is resampling. ImageCR allows you to choose from one of fifteen pixel interpolation algorithms. The default filter, lanczos, produces resampled images of higher quality than most professional desktop applications.

The following code will create images using each resampling formula. We have ordered them here from slowest to fastest. This is generally from highest to lowest quality. One notable exception is lanczos, which we have always considered the best in visual quality, but which is not the absolute slowest.

<cfset algorithms="sinc,bessel,lanczos,mitchell,catrom">
<cfset algorithms="#algorithms#,cubic,quadratic,gaussian,blackman">
<cfset algorithms="#algorithms#,hamming,hanning,hermite,triangle,box,point">
<!--- trying to keep line length down so they fit well in documentation --->

<table>
  <tr>

    <cfoutput>
      <cfloop index="i" list="#algorithms#">
        <cfx_imagecr3 load="#expandpath('../examples/example.jpg')#" 
                      save="#expandpath('./resample-'&i&'.jpg')#" 
                      resample="#i#" 
                      resize="180x180" >
        <td>
          <img src="#imagecr.filename#"><br>
          #i#
        </td>               
      </cfloop>
    </cfoutput>

  </tr>
</table>
resample.cfm 

sinc	bessel	lanczos
mitchell	catrom	cubic
quadratic	gaussian	blackman
hamming	hanning	hermite
triangle	box	point