Which is the difference between field mode and frame mode in interlacing?

[mag93]

Which is the best one for moving objects?

[BruceAmmons]

An interlacing camera actually takes two pictures to create each picture you see. These pictures are taken a fraction of a second apart, which allows the object of interest to move.

First, the camera takes a picture using the odd lines of the sensor. This is the first field. About 1/60 of a second later, it takes a picture using the even lines of the camera. This is the second field. The two pictures are combined by the frame grabber by interlacing the lines, which creates the frame you normally see as the image.

In frame mode, the two fields are combined to make a single image. In field mode, you get the odd lines as an image, and the even lines as a separate image.

When capturing a moving object, it is best to use a progressive scan (non-interla
ced) camera, which takes a picture using all of the lines simultaneously.

If you only have an interlaced camera, your best option is probably field mode, because the object doesn't have a chance to move. There are a couple of drawbacks to field mode, though. The height of your image is only half what it should be, which distorts everything and makes it difficult for measurements and the like. Also, the odd and even images are shifted by one line, so the live image appears to vibrate slightly if there is a lot of detail. You can't compare two consective images, because they are odd and even fields.

I hope this helps.

Bruce

[mag93]

Hi Bruce, thank you very much for your answer but now I have another doubt about the difference between frame mode and field mode. I read on a tutorial by NI (Anatomy of a Camera)that "in frame mode only the even or odd field is repeatedly read from the image sensor while the other field is discarded", while in "field mode pairs of consecutive frame lines are read from the image sensor simultaneously using all frame lines" thus increasing the camera sensitivity by 2. I found this definition even on other documents but on other ones I read your definitions too...perhaps is there any confusion in the terminology to indicate different kinds of interlacing techniques? Is there anyone who can solve my doubts?

[BruceAmmons]

NI is probably talking about a camera in this case, where I am talking about the frame grabber.

Some analog cameras have the option to output frames (alternating odd and even fields) or fields (just the odd field, repeating it twice). This way, you won't accidentally get the even field when you want the odd field.

To the frame grabber it doesn't make a lot of difference. It all depends how you plan on using the images. For moving objects, fields will probably work best.

FYI, if you just want the odd or even fields, you can set up IMAQ to do a buffered acquisition with skip frame = 1, which skips every other frame (or field, actually). Another way to do it is to keep everything in frame mode, and deinterlace the resulting image. Thi
s way, you know which field is odd and which one is even.

Bruce

[mag93]

Then, in field mode, if I take an image and after 1/60 of a second I take another image I would get the odd field and the even field of the same frame, while if I take an image every 1/30 of a second I would get always the odd field or always the even field. Is that right? Why have I to skip one frame or field in a buffered acquisition in field mode?

Annalisa

[BruceAmmons]

The odd and even fields are offset by one line, so they don't match up perfectly. In an image with a lot of detail, it will look like it is shifting back and forth rapidly. If you aren't doing absolute position measurements, it won't hurt anything, but it is irritating to watch at lower playback speeds.

Bruce