Google Dfine 2 basics

Cure your noise with Google Dfine 2

It’s one of the less glamorous plug-ins in the Google Nik Collection, but Google Dfine 2 still fills a valuable function. It’s a noise-reduction plug-in that goes a few steps further than the noise reduction options built into regular applications, so here’s a quick guide to how it works and what it can do.

Google Dfine 2 basics

I’m trying it out on this picture (above) I took hand-held on Westminster Bridge at ISO 3200 with a Nikon D4. The D4 has exceptional high-ISO noise control, but there’s still room for improvement. You won’t be able to see it at this magnification, but if you click on the screenshots for the walkthrough below to see the full-size versions, you should be able to see the noise in the original and the difference that the noise reduction tools make.

01 Automatic Profile

Google Dfine 2 basics

Dfine works in two steps. In Measure mode it scans the image to find representative areas to sample for noise, then generates an automatic noise ‘profile’ to correct it. You can see a number of squares on this picture to show the areas that the software has chosen for sampling.

02 Loupe panel

Google Dfine 2 basics

To see the effect this is having up close, use the Loupe panel at the bottom of the right sidebar. You can click the pin (circled) then click on an area in the image (also circled) to examine a specific region.

03 Manual profile

Google Dfine 2 basics

The automatic noise profile may not always work. Occasionally, Dfine may miss out one of the noisiest areas in the picture, or it may smooth over areas you’re not too worried about, so the solution here is to switch to manual using the pop-up ‘Method’ menu in the tools panel.

04 Reset the profile

Google Dfine 2 basics

You can add new areas to the automatic profile, or you can click the Reset button to start again and only sample those areas you choose.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.