Deinterlacing is the process of converting interlaced video, such as common analog television signals or 1080i format HDTV signals, into a non-interlaced form.
Interlaced video frame consists of two sub-fields taken in sequence, each sequentially scanned at odd and even lines of the image sensor; analog television employed this technique because it allowed for less transmission bandwidth and further eliminated the perceived flicker that a similar frame rate would give using progressive scan. CRT based displays were able to display interlaced video correctly due to its complete analogue nature. All of the newer displays are inherently digital in that the display comprises discrete pixels. Consequently the two fields need to be combined into a single frame, which leads to various visual defects which the deinterlacing process should try to minimise.
Deinterlacing has been researched for decades and employs complex processing algorithms; however, consistent results have been very hard to achieve.[1]
Contents |
Both video and photographic film capture a series of frames (still images) in rapid succession; however, television systems read the captured image by serially scanning the image sensor by lines (rows). In analog television, each frame is divided into two consecutive fields, one containing all even lines, another with the odd lines. The fields are captured in succession at a rate twice that of the nominal frame rate. For instance, PAL and SECAM systems have a rate of 25 frames/s or 50 fields/s, while the NTSC system delivers 29.97 frames/s or 59.94 fields/s. This process of dividing frames into half-resolution fields at double the frame rate is known as interlacing.
Since the interlaced signal contains the two fields of a video frame shot at two different times, it enhances motion perception to the viewer and reduces flicker by taking advantage of the persistence of vision effect. This results in an effective doubling of time resolution as compared with non-interlaced footage (for frame rates equal to field rates). However, interlaced signal requires a display that is natively capable to show the individual fields in a sequential order, and only traditional CRT-based TV sets are capable of displaying interlaced signal, due to the electronic scanning and lack of apparent fixed resolution.
Most modern displays, such as LCD, DLP and plasma displays, are not able to work in interlaced mode, because they are fixed-resolution displays and only support progressive scanning. In order to display interlaced signal on such displays, the two interlaced fields must be converted to one progressive frame with a process known as de-interlacing. However, when the two fields taken at different points in time are re-combined to a full frame displayed at once, visual defects called interlace artifacts or combing occur with moving objects in the image. A good deinterlacing algorithm should try to avoid interlacing artifacts as much as possible and not sacrifice image quality in the process.
There are several techniques available that extrapolate the missing picture information from the available picture information. These fall into the category of intelligent frame creation rather than true deinterlacing. The degree to which they succeed is a function of the processing power that is applied to the task. Further, the visible quality of the final video information is dependent on the quality of the input video (though poor input video is not in itself an inherrent defect of the technique). Currently, the biggest obstacle to an acceptable final picture is compression artifacts in the original video feed. Some of these systems are now good enough that even high bitrate (10 Mb/s or greater) 1080/25p video can be converted into 1080/50p video for display such that it is difficult to distinguish from original 1080/50p video.
Deinterlacing techniques require complex processing and thus can introduce a delay into the video feed. While not generally noticeable, this can result in the display of older video games lagging behind controller input. Many TVs thus have a "game mode" in which minimal processing is done in order to maximize speed at the expense of image quality. Deinterlacing is only partly responsible for such lag; scaling also involves complex algorithms that take precious milliseconds to run.
Interlaced video can carry progressive scan signal, and deinterlacing process should consider this as well.
Typical movie material is shot on 24 frames/s film; when converting film to interlaced video using telecine, each film frame can be presented by two progressive segmented frames (PsF). This format does not require complex deinterlacing algorithm because each field contains a part of the very same progressive frame. However to match 50 field interlaced PAL/SECAM or 59.94/60 field interlaced NTSC signal, frame rate conversion should be performed using various "pulldown" techniques; most advanced TV sets can restore the original 24 frame/s signal using an inverse telecine process. Another option is to speed up 24-frame film by 4% (to 25 frames/s) for PAL/SECAM conversion; this method is still vastly used for DVDs, as well as television broadcasts (SD & HD) in the PAL markets.
DVDs can either encode movies using one of these methods, or store original 24 frame/s progressive video and use MPEG-2 decoder tags to instruct the video player on how to convert them to the interlaced format. Most movies on Blu-ray discs have preserved the original non interlaced 24 frame/s motion film rate and allow output in the progressive 1080p24 format directly to display devices, with no conversion necessary.
Some 1080i HDV camcorders also offer PsF mode with cinema-like frame rates of 24 or 25 frame/s. The TV production can also use special film cameras which operate at 25 or 30 frame/s; such material does not need framerate conversion for broadcasting in the intended video system format.
Deinterlacing requires the display to buffer one or more fields and recombine them into full frames. In theory this would be as simple as capturing one field and combining it with the next field to be received, producing a single frame. However, the originally recorded signal was produced as a series of fields, and any motion of the subjects during the short period between the fields is encoded into the display. When combined into a single frame, the slight differences between the two fields due to this motion results in a "combing" effect where alternate lines are slightly displaced from each other.
There are various methods to deinterlace video, each producing different problems or artifacts of its own. Some methods are much cleaner in artifacts than other methods.
Most deinterlacing techniques can be broken up into three different groups all using their own exact techniques. The first group are called field combination deinterlacers, because they take the even and odd fields and combine them into one frame which is then displayed. The second group are called field extension deinterlacers, because each field (with only half the lines) is extended to the entire screen to make a frame. The third type uses a combination of both and falls under the banner of motion compensation and a number of other names.
Modern deinterlacing systems therefore buffer several fields and use techniques like edge detection in an attempt to find the motion between the fields. This is then used to interpolate the missing lines from the original field, reducing the combing effect.[2]
Line doubling is sometimes confused with deinterlacing in general, or with interpolation (image scaling) which uses spatial filtering to generate extra lines and hence reduce the visibility of pixelation on any type of display.[3] The terminology 'line doubler' is used more frequently in high end consumer electronics, while 'deinterlacing' is used more frequently in the computer and digital video arena.
Best picture quality can be ensured by combining traditional field combination methods (weaving and blending) and frame extension methods (bob or line doubling) to create a high quality progressive video sequence; the best algorithms would also try to predict the direction and the amount of image motion between subsequent sub-fields in order to better blend the two subfields together.
One of the basic hints to the direction and amount of motion would be the direction and length of combing artifacts in the interlaced signal. More advanced implementations would employ algorithms similar to block motion compensation used in video compression; deinterlacers that use this technique are often superior because they can use information from many fields, as opposed to just one or two. This requires powerful hardware to achieve realtime operation.
For example, if two fields had a person's face moving to the left, weaving would create combing, and blending would create ghosting. Advanced motion compensation (ideally) would see that the face in several fields is the same image, just moved to a different position, and would try to detect direction and amount of such motion. The algorithm would then try to reconstruct the full detail of the face in both output frames by combining the images together, moving parts of each subfield along the detected direction by the detected amount of movement.
Motion compensation needs to be combined with scene change detection, otherwise it will attempt to find motion between two completely different scenes. A poorly implemented motion compensation algorithm would interfere with natural motion and could lead to visual artifacts which manifest as "jumping" parts in what should be a stationary or a smoothly moving image.
Deinterlacing of an interlaced video signal can be done at various points in the TV production chain.
Deinterlacing is required for interlaced archive programs when the broadcast format or media format is progressive, as in EDTV 576p or HDTV 720p50 broadcasting, or mobile DVB-H broadcasting; there are two ways to achieve this.
When the broadcast format or media format is interlaced, real-time deinterlacing should be performed by embedded circuitry in a set-top box, television, external video processor, DVD or DVR player, or TV tuner card. Since consumer electronics equipment is typically far cheaper, has considerably less processing power and uses simpler algorithms compared to professional deinterlacing equipment, the quality of deinterlacing may vary broadly and typical results are often poor even on high-end equipment.
Using a computer for playback and/or processing potentially allows a broader choice of video players and/or editing software not limited to the quality offered by the embedded consumer electronics device, so at least theoretically higher dinterlacing quality is possible – especially if the user can pre-convert interlaced video to progressive scan before playback and advanced and time-consuming deinterlacing algorithms (i.e. employing the "production" method).
However, the quality of both free and commercial consumer-grade software may not be up to the level of professional software and equipment. Also, most users are not trained in video production; this often causes poor results as many people do not know much about deinterlacing and are unaware that the frame rate is half the field rate. Many codecs/players do not even deinterlace by themselves and rely on the graphics card and video acceleration API to do proper deinterlacing.
The European Broadcasting Union has argued against the use of interlaced video in production and broadcasting, recommending 720p 50 fps (frames per second) as current production format and working with the industry to introduce 1080p50 as a future-proof production standard which offers higher vertical resolution, better quality at lower bitrates, and easier conversion to other formats such as 720p50 and 1080i50.[4][5] The main argument is that no matter how complex the deinterlacing algorithm may be, the artifacts in the interlaced signal cannot be completely eliminated because some information is lost between frames.
"Deinterlacing modes:" Bob, Linear, Yadif (2x) (v1.1.0+), Phosphor(v1.2.0+), best of them and the slowest are QTGMC (AviSynth).
These algorithms display the video at the original half-picture rate, which is typically 50 (PAL) or 60 (NTSC) half-pictures per second. This is double the full picture rate, hence the name. This approach to deinterlacing is also known as field rendering.This group takes into account that the half-pictures of a true interlaced video were intended to be displayed at different times. This can make the motion look very smooth.Simple doublers (Bob and Linear) display only one half-picture at a time. Nevertheless, the quick alternating display produces a convincing illusion of full vertical resolution while playback is running.Some more advanced doublers (such as Yadif (2x)) are based on interpolators (see below), and attempt to generate full pictures to display. When interpolators are used in this way, which field is kept, alternates just like in the simple doublers.The last doubler (Phosphor) does not fit into either of these categories, but attempts to simulate a traditional CRT TV.All doublers can be used with both true interlaced and telecined video.
Here is the following methods from VLC:
| Algo | 4:2:0 | 4:2:2 | Algo type | Interpolation (if applic.) | Notes |
|---|---|---|---|---|---|
| C, H, FR | C, H, FR | ||||
| Discard | 0, h, 1x | 0, f, 1x | interpolator | none | keeps only top field; each line is repeated |
| Mean | 0, h, 1x | 2, h, 1x | blender | half-resolution blender; 1) | |
| Blend | 0, f, 1x | 0, f, 1x | blender | full-resolution blender; 2) | |
| Bob | 0, f, 2x | 0, f, 2x | doubler | none | each line is repeated; 3) |
| Linear | 0, f, 2x | 2, f, 2x | doubler | simple linear | first/last line copied; others interpolated; 3) |
| X | 0, f, 1x | 2, f, 1x | interpolator | MC + edge-oriented | keeps only top field in interlaced parts; 4) |
| Yadif | 0, f, 1x | 2, f, 1x | interpolator | spatial/temporal | keeps only top field in interlaced parts |
| Yadif (2x) | 0, f, 2x | 2, f, 2x | doubler | spatial/temporal | Yadif and Yadif (2x) come from MPlayer |
| Phosphor | 5), f, 2x | 2, f, 2x | doubler | CRT TV simulator; 6) | |
| IVTC | 0, f, 7) | 2, f, 7) | inverse telecine |
|
||||||||||||||