Digital Video (DV)
Digital Video (DV) is a digital video format created by Sony, JVC, Panasonic and other video camera producers, and launched in 1995. In its smaller tape form factor MiniDV, has since become a standard for home and semi-professional video production; it is sometimes used for professional purposes as well, such as filmmaking and electronic news gathering (ENG). The DV specification (originally known as the Blue Book, current official name IEC 61834) defines both the codec and the tape format. Features include intraframe compression for uncomplicated editing and good video quality, especially compared to earlier consumer analog formats such as Video8, Hi8 and VHS-C. DV now enables filmmakers to produce movies inexpensively, and is strongly associated with independent film and citizen journalism.
The high quality of DV images, especially when compared to Video8 and Hi8 which were vulnerable to an unacceptable amount of video dropouts and “hits", prompted the acceptance by mainstream broadcasters of material shot on DV. The low costs of DV equipment and their ease of use put such cameras in the hands of a new breed of videojournalists. Programs such as TLC’s Trauma: Life in the E.R. and ABC News’ Hopkins: 24/7 were shot on DV. CNN’s Anderson Cooper is perhaps the best known of the generation of reporter/videographers who began their professional careers shooting their own stories.
There have been some variants on the DV standard, most notably Sony's DVCAM and Panasonic's DVCPRO formats targeted at professional use. Sony's consumer Digital8 format is another variant, which is similar to DV but recorded on Hi8 tape. Other formats such as DVCPRO50 utilize DV25 encoders running in parallel.
MiniDV tapes can also be used to record a high-definition format called HDV in cameras designed for this codec, which differs significantly from DV on a technical level as it uses MPEG-2 compression. MPEG-2 is more efficient than the compression used in DV, in large part due to inter-frame/temporal compression. This allows for higher resolution at bitrates similar to DV. On the other hand, the use of inter-frame compression can cause motion artifacts and complications in editing. Nonetheless, HDV is being widely adopted for both consumer and professional purposes and is supported by many editing applications using either the native HDV format or intermediary editing codecs.
Technical standards
Before compression
To avoid aliasing, optical low pass filtering is necessary (although not necessarily implemented in all camera designs). Essentially, blurry glass is used to add a small blur to the image. This prevents high-frequency information from getting through and causing aliasing. Low-quality lenses can also be considered a form of optical low pass filtering.
Before arriving at the codec compression stage, light energy hitting the sensor is transduced into analog electrical signals. These signals are then converted into digital signal by an analog to digital converter (ADC or A/D). This signal is then processed by a digital signal processor (DSP) or custom ASIC and undergoes the following processes:
Processing of raw input into (linear) RGB signals: For Bayer pattern-based sensors (i.e. sensors utilizing a single CCD or CMOS and color filters), the raw input has to be demosaiced. For Foveon-based designs, the signal has to be processed to remove cross-talk between the color layers. In pixel-shifted 3CCD designs, a process somewhat similar to de-mosaicing is applied to extract full resolution from the sensor.
Matrix (for colorimetry purposes): The digital values are matrixed to tweak the camera's color response to improve color accuracy and to make the values appropriate for the target colorimetry (SMPTE C, Rec. 709, or EBU phosphor chromaticities). For performance reasons, this matrix may be applied after gamma correction and combined with the matrix that converts R'G'B' to Y'CbCr.
Electronic white balance may be applied in this matrix, or in the matrix operation applied after gamma correction.
Gamma correction: Gamma correction is applied to the linear digital signal, following the Rec. 601 transfer function (a power function of 1/0.45). Note that this increases the quantization error from before.
Matrix (R'G'B' to Y'CbCr conversion): This matrix converts the gamma-corrected R'G'B' values to Y'CbCr color space. Y'CbCr color space facilitates chroma subsampling. This operation introduces yet more quantization error, in large part due to a difference in the scale factors between the Y' and Cb and Cr components.
Chroma Subsampling: Since human vision has greater acuity for luminance than color, performance can be optimized by devoting greater bandwidth to luminance than color. Chroma subsampling approximates this by converting R'G'B' values into Y'CbCr color space. The Cb and Cr color difference components are stored at a lower resolution than the Y' (luma) component.
Sharpening is often used to counteract the effect of optical low pass filtering. Sharpening can be implemented via finite impulse response filters.
The data is now compressed using one of several algorithms including discrete cosine transform (DCT), adaptive quantization (AQ), and variable-length coding (VLC).
Video compression
DV uses DCT intraframe compression at a fixed bitrate of 25 megabits per second (25.146 Mbit/s), which, when added to the sound data (1.536 Mbit/s), the subcode data, error detection, and error correction (approx 8.7 Mbit/s) amounts in all to roughly 36 megabits per second (approx 35.382 Mbit/s). At equal bitrates, DV performs somewhat better than the older MJPEG codec, and is comparable to intraframe MPEG-2. (Note that many MPEG-2 encoders for real-time acquisition applications only use intraframe compression [I-frames only], but not interframe compression [P and B frames].) DCT compression is lossy, and sometimes suffers from artifacting around small or complex objects such as text. The DCT compression has been specially adapted for storage onto tape. The image is divided into macroblocks, each consisting of 4 luminance DCT blocks and 1 chrominance DCT block. Furthermore 6 macroblocks, selected at positions far away from each other in the image, are coded into a fixed amount of bits. Finally, the information of each compressed macroblock is stored as much as possible into one sync-block on tape. All this makes it possible to search video on tape at high speeds, both forward and reverse, as well as to correct faulty sync blocks very well.
Audio
DV allows either 2 digital audio channels (usually stereo) at 16-bit resolution and 48 kHz sampling rate, or 4 digital audio channels at 12-bit resolution and 32 kHz sampling rate. For professional or broadcast applications, 48 kHz is used almost exclusively. In addition, the DV spec includes the ability to record audio at 44.1 kHz (the same sampling rate used for CD audio), although in practice this option is rarely used. DVCAM and DVCPRO both use locked audio while standard DV does not. This means that at any one point on a DV tape the audio may be +/- ⅓ frame out of sync with the video. However, this is the maximum drift of the audio/video sync; it is not compounded throughout the recording. In DVCAM and DVCPRO recordings the audio sync is permanently linked to the video sync.
Application software support
Most DV players, editors and encoders only support the basic DV format, but not its professional versions. DV Audio/Video data can be stored as raw DV data stream file (data is written to a file as the data is received over FireWire, file extensions are.dv and.dif) or the DV data can be packed into AVI container files. The DV meta-information is preserved in both file types.
Most Windows video software only supports DV packed into AVI containers, as they use Microsoft's avifile.dll, which only supports reading avi files. A few notable exceptions exist:
* Apple Inc.'s QuickTime Player: QuickTime by default only decodes DV to half of the resolution to preserve processing power for editing capabilities. However, in the "Pro" version the setting "High Quality" under "Show Movie Properties" enables full resolution playback.
* DVMP Basic & DVMP Pro: full decoding quality. Plays AVI (inc DVCPRO25 and DVCAM) and.dv files. Also displays the DV meta-information (e.g. timecode, date/time, f-stop, shutter speed, gain, white balance etc)
* The VLC media player (Free software): full decoding quality
* MPlayer (also with GUI under Windows and Mac OS X): full decoding quality
* muvee Technologies autoProducer 4.0: Allows editing using FireWire IEEE 1394
* Quasar DV codec (libdv) - open source DV codec for Linux
Type 1 and Type 2 DV AVI files
There are two types of DV-AVI files:
* Type 1: The multiplexed Audio-Video is kept in its original multiplexing and saved together into the Video section of the AVI file
o Does not waste much space (audio is saved uncompressed, but even uncompressed audio is tiny compared to the video part of DV), but Windows applications based on the VfW API do not support it.
* Type 2: Like type 1, but audio is also saved as an additional audio stream into the file.
o Supported by VfW applications, at the price of little increased file size.
Type 1 is actually the newer of the two types. Microsoft made the "type" designations, and decided to name their older VfW-compatible version "Type 2", which only furthered confusion about the two types. In the late 1990s through early 2000s, most professional-level DV software, including non-linear editing programs, only supported Type 1. One notable exception was Adobe Premiere, which only supported Type 2. High-end FireWire controllers usually captured to Type 1 only, while "consumer" level controllers usually captured to Type 2 only. Software is and was available for converting Type 1 AVIs to Type 2, and vice-versa, but this is a time-consuming process.
Many current FireWire controllers still only capture to one or the other type. However, almost all current DV software supports both Type 1 and Type 2 editing and rendering, including Adobe Premiere. Thus, many of today's users are unaware of the fact that there are two types of DV AVI files. In any event, the debate continues as to which – Type 1 or Type 2 – if either, is better.
Source: Wikipedia
Digital Video (DV) is a digital video format created by Sony, JVC, Panasonic and other video camera producers, and launched in 1995. In its smaller tape form factor MiniDV, has since become a standard for home and semi-professional video production; it is sometimes used for professional purposes as well, such as filmmaking and electronic news gathering (ENG). The DV specification (originally known as the Blue Book, current official name IEC 61834) defines both the codec and the tape format. Features include intraframe compression for uncomplicated editing and good video quality, especially compared to earlier consumer analog formats such as Video8, Hi8 and VHS-C. DV now enables filmmakers to produce movies inexpensively, and is strongly associated with independent film and citizen journalism.
The high quality of DV images, especially when compared to Video8 and Hi8 which were vulnerable to an unacceptable amount of video dropouts and “hits", prompted the acceptance by mainstream broadcasters of material shot on DV. The low costs of DV equipment and their ease of use put such cameras in the hands of a new breed of videojournalists. Programs such as TLC’s Trauma: Life in the E.R. and ABC News’ Hopkins: 24/7 were shot on DV. CNN’s Anderson Cooper is perhaps the best known of the generation of reporter/videographers who began their professional careers shooting their own stories.
There have been some variants on the DV standard, most notably Sony's DVCAM and Panasonic's DVCPRO formats targeted at professional use. Sony's consumer Digital8 format is another variant, which is similar to DV but recorded on Hi8 tape. Other formats such as DVCPRO50 utilize DV25 encoders running in parallel.
MiniDV tapes can also be used to record a high-definition format called HDV in cameras designed for this codec, which differs significantly from DV on a technical level as it uses MPEG-2 compression. MPEG-2 is more efficient than the compression used in DV, in large part due to inter-frame/temporal compression. This allows for higher resolution at bitrates similar to DV. On the other hand, the use of inter-frame compression can cause motion artifacts and complications in editing. Nonetheless, HDV is being widely adopted for both consumer and professional purposes and is supported by many editing applications using either the native HDV format or intermediary editing codecs.
Technical standards
Before compression
To avoid aliasing, optical low pass filtering is necessary (although not necessarily implemented in all camera designs). Essentially, blurry glass is used to add a small blur to the image. This prevents high-frequency information from getting through and causing aliasing. Low-quality lenses can also be considered a form of optical low pass filtering.
Before arriving at the codec compression stage, light energy hitting the sensor is transduced into analog electrical signals. These signals are then converted into digital signal by an analog to digital converter (ADC or A/D). This signal is then processed by a digital signal processor (DSP) or custom ASIC and undergoes the following processes:
Processing of raw input into (linear) RGB signals: For Bayer pattern-based sensors (i.e. sensors utilizing a single CCD or CMOS and color filters), the raw input has to be demosaiced. For Foveon-based designs, the signal has to be processed to remove cross-talk between the color layers. In pixel-shifted 3CCD designs, a process somewhat similar to de-mosaicing is applied to extract full resolution from the sensor.
Matrix (for colorimetry purposes): The digital values are matrixed to tweak the camera's color response to improve color accuracy and to make the values appropriate for the target colorimetry (SMPTE C, Rec. 709, or EBU phosphor chromaticities). For performance reasons, this matrix may be applied after gamma correction and combined with the matrix that converts R'G'B' to Y'CbCr.
Electronic white balance may be applied in this matrix, or in the matrix operation applied after gamma correction.
Gamma correction: Gamma correction is applied to the linear digital signal, following the Rec. 601 transfer function (a power function of 1/0.45). Note that this increases the quantization error from before.
Matrix (R'G'B' to Y'CbCr conversion): This matrix converts the gamma-corrected R'G'B' values to Y'CbCr color space. Y'CbCr color space facilitates chroma subsampling. This operation introduces yet more quantization error, in large part due to a difference in the scale factors between the Y' and Cb and Cr components.
Chroma Subsampling: Since human vision has greater acuity for luminance than color, performance can be optimized by devoting greater bandwidth to luminance than color. Chroma subsampling approximates this by converting R'G'B' values into Y'CbCr color space. The Cb and Cr color difference components are stored at a lower resolution than the Y' (luma) component.
Sharpening is often used to counteract the effect of optical low pass filtering. Sharpening can be implemented via finite impulse response filters.
The data is now compressed using one of several algorithms including discrete cosine transform (DCT), adaptive quantization (AQ), and variable-length coding (VLC).
Video compression
DV uses DCT intraframe compression at a fixed bitrate of 25 megabits per second (25.146 Mbit/s), which, when added to the sound data (1.536 Mbit/s), the subcode data, error detection, and error correction (approx 8.7 Mbit/s) amounts in all to roughly 36 megabits per second (approx 35.382 Mbit/s). At equal bitrates, DV performs somewhat better than the older MJPEG codec, and is comparable to intraframe MPEG-2. (Note that many MPEG-2 encoders for real-time acquisition applications only use intraframe compression [I-frames only], but not interframe compression [P and B frames].) DCT compression is lossy, and sometimes suffers from artifacting around small or complex objects such as text. The DCT compression has been specially adapted for storage onto tape. The image is divided into macroblocks, each consisting of 4 luminance DCT blocks and 1 chrominance DCT block. Furthermore 6 macroblocks, selected at positions far away from each other in the image, are coded into a fixed amount of bits. Finally, the information of each compressed macroblock is stored as much as possible into one sync-block on tape. All this makes it possible to search video on tape at high speeds, both forward and reverse, as well as to correct faulty sync blocks very well.
Audio
DV allows either 2 digital audio channels (usually stereo) at 16-bit resolution and 48 kHz sampling rate, or 4 digital audio channels at 12-bit resolution and 32 kHz sampling rate. For professional or broadcast applications, 48 kHz is used almost exclusively. In addition, the DV spec includes the ability to record audio at 44.1 kHz (the same sampling rate used for CD audio), although in practice this option is rarely used. DVCAM and DVCPRO both use locked audio while standard DV does not. This means that at any one point on a DV tape the audio may be +/- ⅓ frame out of sync with the video. However, this is the maximum drift of the audio/video sync; it is not compounded throughout the recording. In DVCAM and DVCPRO recordings the audio sync is permanently linked to the video sync.
Application software support
Most DV players, editors and encoders only support the basic DV format, but not its professional versions. DV Audio/Video data can be stored as raw DV data stream file (data is written to a file as the data is received over FireWire, file extensions are.dv and.dif) or the DV data can be packed into AVI container files. The DV meta-information is preserved in both file types.
Most Windows video software only supports DV packed into AVI containers, as they use Microsoft's avifile.dll, which only supports reading avi files. A few notable exceptions exist:
* Apple Inc.'s QuickTime Player: QuickTime by default only decodes DV to half of the resolution to preserve processing power for editing capabilities. However, in the "Pro" version the setting "High Quality" under "Show Movie Properties" enables full resolution playback.
* DVMP Basic & DVMP Pro: full decoding quality. Plays AVI (inc DVCPRO25 and DVCAM) and.dv files. Also displays the DV meta-information (e.g. timecode, date/time, f-stop, shutter speed, gain, white balance etc)
* The VLC media player (Free software): full decoding quality
* MPlayer (also with GUI under Windows and Mac OS X): full decoding quality
* muvee Technologies autoProducer 4.0: Allows editing using FireWire IEEE 1394
* Quasar DV codec (libdv) - open source DV codec for Linux
Type 1 and Type 2 DV AVI files
There are two types of DV-AVI files:
* Type 1: The multiplexed Audio-Video is kept in its original multiplexing and saved together into the Video section of the AVI file
o Does not waste much space (audio is saved uncompressed, but even uncompressed audio is tiny compared to the video part of DV), but Windows applications based on the VfW API do not support it.
* Type 2: Like type 1, but audio is also saved as an additional audio stream into the file.
o Supported by VfW applications, at the price of little increased file size.
Type 1 is actually the newer of the two types. Microsoft made the "type" designations, and decided to name their older VfW-compatible version "Type 2", which only furthered confusion about the two types. In the late 1990s through early 2000s, most professional-level DV software, including non-linear editing programs, only supported Type 1. One notable exception was Adobe Premiere, which only supported Type 2. High-end FireWire controllers usually captured to Type 1 only, while "consumer" level controllers usually captured to Type 2 only. Software is and was available for converting Type 1 AVIs to Type 2, and vice-versa, but this is a time-consuming process.
Many current FireWire controllers still only capture to one or the other type. However, almost all current DV software supports both Type 1 and Type 2 editing and rendering, including Adobe Premiere. Thus, many of today's users are unaware of the fact that there are two types of DV AVI files. In any event, the debate continues as to which – Type 1 or Type 2 – if either, is better.
Source: Wikipedia
