Cookies are disabled. This site requires cookies turned on to run properly. Please enable cookies and reload this page.

Best noise reduction
for digital video

Treating Dust and Scratches in video clips with Neat Video

Neat Video includes a dedicated Dust and Scratches filter designed to reduce or fully remove certain defects of analog nature introduced into video data before or during its conversion to the digital representation.

See also:
Video tutorial on using Dust and Scratches filter in Neat Video 5

Dust and scratches

The Dust and Scratches filter reduces such defects as:

  • spots, dots, lines of relatively large size;
  • corrupted scan lines: a whole scan line of a video frame is incorrect;
  • random short lines affecting only a small part of a scan line.

Some of the areas containing such defects are marked by blue boxes in these sample frames:

Such defects are generally caused by:

  • physical dust and scratches present in original video film;
  • inaccurate recording/playing back of VHS materials;
  • damaged VHS film;
  • interference in analog TV transmission.

The common property of these defects is that each particular defect element affects a single frame in the clip rather than a series of frames. This property allows Neat Video to use information from adjacent frames to eliminate the defects present in the current frame.


For a live illustration of the Dust and Scratches filter's performance please play back this clip:


As you can see, the dust and scratches present in this clip are significantly reduced.

The Dust and Scratches filter has several controls and modes so you can adjust the amount and character of noise reduction to be applied.

Please see the User Guide for more information about the Dust and Scratches filter or check a video tutorial explaining how to use this feature.

Additional notes about dust and scratches

Note: For accurate reduction of such defects by the Dust and Scratches filter, make sure you do not include these defects into an analysis area during profiling (visually inspect the area selected by Auto Profile to ensure that). If you include these defects into an analysis area, then the regular noise reduction may become less accurate. The best profiling approach is to include only regular random noise in the sample area used by Auto Profile for analysis.