How to resize a scanned image for video output

(without stretching it or losing quality)

 

1.            Start with an image scanned at medium to high resolution, in the 1000 to 4000 pixel range.

2.            Open the image in Photoshop. Make any changes you wish, such as dust and scratch removal or color correction. Save As… to a new filename, such as scan_repaired.psd.

3.            Before proceeding, save another copy of the file, so you don’t accidentally overwrite your high resolution repaired scan. Save As… to a new filename, such as scan_video.psd.

4.            Figure out what digital video format you are exporting to. The two primary NTSC formats are:

CCIR-601:      720 x 486 x 0.9      (for Zap-It, or Video Lab high-end systems)
DV:                  720 x 480 x 0.9      (for Speed Razor or FireWire)

5.            Resize the image, remembering that 10% of the image will be lost on all four sides when the image is viewed on video. To account for this, you must make the image smaller than the full frame of video.

Select Image > Image Size from the main menu. Look at the Pixel Dimensions. If the image is taller than it is wide, it is a portrait aspect. If it is wider than it is tall, it is a landscape aspect.

For portrait images, resample the image to be 480 pixels tall. Make sure Resample is on and Constrain Proportions is on. Let Photoshop figure out how many pixels wide the image should be.

For landscape images, resample the image to be 640 pixels wide. Make sure Resample is on and Constrain Proportions is on. Let Photoshop figure out how many pixels tall the image should be.

6.            You need to control the background color for the image. To do this effectively, you must have two layers, one for the image and one for the background.

Unlock the current image layer. Double-click the layer to get the New Layer dialog, and click OK. The layer is now unlocked.

Create a new empty layer, and drag it to the bottom of the layer stack.

7.            Select Image > Canvas Size from the main menu. Enter 720 pixels for the new canvas width, and 540 pixels for the new canvas height.

8.            The image you scanned is now a floating layer above a transparent background layer. Select the background layer and use the Paint Bucket or Gradient tools to create a background. Black is a good choice for video.

9.            Save the file, overwriting scan_video.psd.

10.       Now you must resample the image again in order to compensate for the non-square pixels of digital video. Select Image > Image Size. Turn OFF the Constrain Proportions option. This lets you resample in only one dimension. Under Pixel Dimensions > Height, enter 480 or 486 pixels tall, depending on whether you’re outputting to DV or CCIR-601. The Width must remain at 720 pixels.

You now have an image which is either 720 x 480, or 720 x 486, and looks slightly stretched from side to side. This is normal. When you output it to video, it will look OK.

By setting up your canvas to 720 x 540, when you resample it to 720 x 486 at this later stage, you are downsampling instead of upsampling. This gives you slightly better quality than if you set up your canvas to 648 x 486, and then upsampled to 720 x 486.

11.       Save As… to a new filename, such as scan_dv.psd or scan_601.psd.

12.       Save As… a copy of the file to the TGA (Targa) format, so that the video hardware can accept it. Make sure to choose 24-bit in the Targa file format options. Don’t use compression. (Targa’s compression is lossless, but some video hardware may not accept that format.)

13.       You now have five files:

a.     original scan at high resolution

b.     Photoshop file, retouched scan at high resolution

c.      Photoshop file, with layers, resized to 720 x 540

d.     Photoshop file, with layers, resized to 720 x 486 (or 720 x 480)

e.     Targa file, no layers, no alpha channel, 720 x 486 (or 720 x 480)

Create a CD-R of all five files to turn in for the assignment. Also, save all five files to your home directory.

NOTE: You can now bring your Targa file into Premiere or your editing program of choice. Make sure your project is set up for the correct format, either CCIR-601 or DV. The Targa file should drop right into Premiere with no distortion and no resizing within the editing application.


Return to home