What are the Key Frames?
In the field of video compression, calculating differences between frames usually gives better results. Video encoders can achieve better compression ratio by writing a full picture for the first frame, and then subsequent frames as a delta from the previous frame.
The frames with the full picture is called key frames, or I-frames.
The less key frames video encoders write into a video file, the better compression ratio it will be. However, too few key frames may cause problems when fast-forwarding or rewinding the video. Video creators choose the best balance how often they insert key frames in the video stream. For instance, a DVD video stream usually has two to three key frames in a second.
Delta frames come into an issue when you want to split a video file. Since delta frames cannot be played without previous frames, cutting at a delta frame breaks the video.
WMV Cut & Split first determines if the frame to cut is a key frame or not. If it is a key frame, WMV Cut & Split can cut the video file by constructing the appropriate WMV file structures. This process does not require reencoding and thus keeps the quality of the original video without any loss.
If it is not a key frame, WMV Cut & Split rewinds back to the last key frame before the cut point, decodes the following delta frames, and recompress them up to the next key frame. This process is called reencoding.
The reencoding process finishes when it reaches to the next key frame. WMV Cut & Split goes back to "no reencoding" mode for the following frames, so that the quality loss occurs only in a limited portion of the video file.
As noted above, reencoding is usually only for less than half a second, and all the rests are split without any quality loss.
Two-Pass Encoding for the Best Quality
When cutting at delta frames, the reencoding process is required until it hits the next key frame.
WMV Cut & Split uses the two-pass encoding technology when reencoding. It is true that the reencoding cannot be done without the loss in the quality, but the two-pass encoding can minimize the loss.
We know the quality of the video matters.