End of Line Signals - m_axis_video_tlast, s_axis_video_tlast - 2.3 English

Video Processing Subsystem Product Guide (PG231)

Document ID
PG231
Release Date
2022-04-27
Version
2.3 English

The End-Of-Line signal, physically transmitted over the AXI4-Stream TLAST signal, marks the last pixel of a line. The EOL pulse is 1 valid transaction wide, and must coincide with the last pixel of a scan-line as shown in This Figure.

Figure 3-2:      Use of EOL

X-Ref Target - Figure 3-2

use_of_eol.JPG

The following deinterlacing algorithms are supported:

Bob (Line Doubling): The Bob algorithm takes the lines of each interlaced field (consisting of only even or odd lines of a progressive frame) and duplicates them, filling the entire frame. Line doubling prevents "combing" artifacts but causes a noticeable reduction in video quality. This is noticeable mostly on stationary or slowly moving objects since they seem to bob up and down. This algorithm does not require external frame buffers.

Weave: Weaving is done by assembling 2 consecutive fields together. This performs well when the image hasn't changed between fields, but any move will cause "combing" artifacts - when the pixels in one frame do not align with the pixels in the next frame, forming a jagged edge. Weave algorithm is a motion adaptive method and requires external frame buffers.

Vertical Temporal Linear Interpolation (VTLin): VTLin generates intermediate lines by linear interpolation of its 2 neighboring lines from the current field and 3 reference lines from the previous field. The generated lines are added to lines of the current field, forming the entire frame. VTLin performs better than Bob or Weaving for moving objects, but interpolation results in slight image blurring. VTLin is a motion adaptive algorithm and requires an external frame buffer.

Vertical Temporal Median (VTMed): VTMed makes the generated lines by taking the median of its 2 neighboring lines from the current field and the reference line from the previous field. The generated lines are added to lines of the current field, forming the entire frame. VTMed is a motion adaptive algorithm.

Median: The Median algorithm takes the median of the result from VTLin and VTMed in 2 consecutive fields and the reference line in the next field as the generated lines. The output frame is consisted of generated lines and reference lines of current field. Median is a motion adaptive method and requires an external frame buffer.

Bilinear Interpolation: This algorithm is done by performing bilinear interpolation on the current interlaced field. It fills generated lines by taking the average of 2 neighboring reference lines. The Bilinear Interpolation algorithm is not motion adaptive.