User Manual 4.0.cwk

Transcript

Multi-Standard and CBS Color Converter User and Technical Manual Copyright 2004-2007 DAH Revision 4.0 7 June, 2007 Introduction Introduction This manual covers the operation and technical aspects of the Multi-Standard Converter and the CBS Color Converter. The Multi-Standard converter can be controlled to operate in an NTSC or PAL/SECAM input mode by an internal, user selectable jumper. The CBS Color converter only comes in an NTSC input version. Other differences between the different models will be called out when appropriate. Features • Compact, low power, surface mount design • Fully automatic operation • Front panel multi-function user push button and Status LED • Rear panel user switch selectable output standard control • Internal user switch selectable options control • Zoom feature to output a 4:3 image from a 16:9 source (electronic standards only) • Extremely stable output: +/- 3% levels, +/- 50ppm timing • Output clock line locked to input clock for perfect conversions • 10 bit video D/A for greater than 54db dynamic range • 32Mb FLASH Image Memory for storing custom images (Multi-Standard Only) - up to 2 custom images plus one fixed test pattern for each standard • 300K gate equivalent FieldProgrammableGateArray (100K CBS Color) • FLASH memory for FPGA firmware • Extremely accurate algorithms used for conversions: - Full FIR filtering for image scaling, 3 to 81 tap depending on standard - Proprietary reverse 3:2 pull-down algorithm for perfect 24fps standards - All internal calculations done to a minimum 9 bit precision • Three full frames of static RAM for temporal scaling • Versatile I/O: - Composite Video Input (NTSC/PAL, 1Vpp, 75 ohm) - S-Video Input (NTSC/PAL, Y- 1 Vpp, 75 ohm C- 600mVpp, 75 ohm) - Composite Video Output (selectable standard, 1Vpp, 75 ohm or 2Vpp 20K depending on standard) - Stereo Audio I/O (16bit, 35kHz, 3Vpp, 20k input, 200 ohm output) - DC power (6Vdc maximum, 500ma) 3 Introduction Front Panel The front panel controls are shown below: Power Switch Status LED User Pushbutton Status LED: The status LED conveys the current operating state of the converter. A solid light indicates the converter is locked while a flashing light indicates the converter is unlocked as follows: Flashing Red: No video input signal detected No external mechanical sync detected (mechanical standards only) Solid Red: Converter locked to video input No external mechanical sync detected (mechanical standards only) Flashing Yellow: No video input signal detected External mechanical sync detected (mechanical standards only) Solid Yellow: Converter locked to video input External mechanical sync detected (mechanical standards only) Flashing Green: Converter locked to video input Loss of cadence lock (24fps based standards only) Solid Green: Converter locked to video input Converter locked to 3:2 material (24fps based standards only) Toggling Red: and Green Data being stored into Image Memory 4 Introduction User Pushbutton (Multi-Standard Only): The User Pushbutton controls several functions of the converter, and operates in one of three modes as follows: The first mode is obtained when valid video is being supplied to the converter (solid Status LED). This is the Normal mode. In this mode, when an electronic output standard is selected and the button is push momentarily, the Zoom function will be toggled. Most electronic video output standards have an image aspect ratio of 4 wide to 3 high, or referred to as a 4:3 aspect ratio. (some very old formats are slightly different at 6:5) Current analog television standards, NTSC and PAL/SECAM, also have an image aspect ratio of 4:3. When the converter is supplied with an input video that has a 4:3 aspect ratio, it simply converts this input video to the selected output standard. With the advent of HD television, many DVD’s, tapes, and off-air shows are using the new wide screen 16:9 aspect ratio in what is called a “letterbox” mode. This is where the entire 16:9 image is displayed on a 4:3 television with black bars at the top and bottom of the screen. This is done to show the entire image as filmed, however on older 4:3 televisions, most with small CRT’s, a very small usable image results. By using the Zoom feature, the converter will take the center 4:3 section of the 16:9 image and expand it to fill the screen. This can actually result in a larger, sharper image since the input video can have much more resolution than the selected output standard. For instance, a PAL image has 720X576 pixels. A letterboxed 16:9 image on a PAL input represents 720 X 432 pixels. Cutting a 4:3 section out of this yields 540X432 pixels which is still much higher than the final resolution of a 405/25i standard. This means that even though the converter only used a portion of the input video, it still needs to scale it down to the 405/25i standard. While this is not true for every input versus output standard, it is a very useful feature when viewing 16:9 material on an older television. The Zoom function can be toggled at any time desired, however it will always resort to Off when the unit is first powered up. Also in the Normal mode when a mechanical output standard is selected, the framing of the image can be accomplished electronically without having to use the method provided by the mechanical scanning disk, usually a knob that rotates the motor or lamp assembly. (of course you can still use the mechanical framing method) In this mode, when the button is push momentarily, the image will move by one line down or to the left, depending on the scanning direction of the selected standard. If the button is held down, the image will repeatedly move at a 1Hz rate by one line up or to the right, depending on the scanning direction of selected standard. By pushing or holding this button down, a mechanical scanning disk can be easily framed. The push button has no effect on electronic standards in this mode. The second mode is when valid video is not being supplied to the converter (flashing Status LED). This is the Default mode. In this mode, the converter will freerun at the selected output standard, and one of several images stored in the internal FLASH memory, or a test pattern can be displayed as a default image. This mode is useful for showing off a television when no source video is available, or for providing an image source during repair and testing. In this mode, when the push button is held down, the output will rotate through the stored images and test pattern at a 1Hz rate. This mode is available for all standards, electronic and mechanical. Additionally, as in 5 Introduction the Normal mode, if the push button is pressed momentarily and a mechanical standard is selected, the image will repeatedly move by one line down or to the left depending on the scanning direction of the selected standard. Pressing the button momentarily in this mode will have no effect if an electronic standard is selected. The third mode is used for clearing the internal Image Memory and for storing custom default images into this FLASH memory. This is the Image Storing mode. This mode is entered by holding the push button down while turning the unit on, and then releasing the push button. The Status LED will now additionally blink at a slow rate to show the Image Storing mode is active. Care must be taken in this mode as previously stored images can be overwritten. Once in this mode, with no video input attached, the push button can be held down until the desired position in the Image Memory is reached as previously described in the Default mode. (Note: the test patterns are fixed and can not be changed) No actual storing or erasing of the Image Memory is done at this time. Once the desired location is reached, a valid video input is connected to the converter. The converter is now ready to store the image into it’s internal memory. Momentarily pressing the push button will erase the existing image and store the new image into the Image Memory. This may take less than a second for standards with small image size, and up to several seconds for standards with large image sizes. During the storing process, the Status LED will toggle between Red and Green to show the image is being stored. The output from the converter will be blanked during the storing process. (Note: if a mechanical standard was selected, it may need to be re-framed when the storing process is finished, and the output returns to normal) When the storing process is finished, the output and Status LED will return to normal. Additional images can be stored for the selected standard by once again removing the video input, holding the push button down until the desired location is reached, and reapplying the video input. (Note: remember there are up to two (2) programmable locations for each of the electronic standards, and only one (1) programmable location for each of the mechanical standards. The test patterns are fixed and can not be changed.) When all desired images are stored, the unit should be switched off to exit the Image Storing mode. It can be immediately turned back on and used as normal. Finally, the entire Image Memory can be erased all at once. Since this mode is catastrophic to all stored images for all standards, it was purposely made very difficult to initiate. To start a Full Image Memory Erase, hold the push button down while turning the unit on, then without releasing the push button, change the Standards Selector switch to any other standard from where it currently is at. (in other words rotate the switch in any direction) The converter will sense this and enter the Full Image Memory Erase mode. The Status LED will toggle between Red and Green while the Image Memory is being erased, and the output will be blanked. The Full Image Memory Erase procedure can take up to 1 minute to complete. When finished, the converter will automatically enter into the Image Storing mode and operate as previously described. 6 Introduction Rear Panel The rear panel connections are shown below: Composite Composite Audio Input Output Output S-Video Audio Standard Power Input Input Selector 2 0 E 4 6 8 C A Power: The converter requires a power source of between 5.5 and 6.0 volts DC at 500 mA. Use of an unregulated power supply that provides a voltage over 6.0V can cause damage to the converter and is not recommended. Composite / S-Video Input: A video source conforming to the NTSC or PAL video standard can be supplied to either the Composite (RCA or BNC) or S-Video (MiniDIN) input connectors. The converter will switch back and forth between the two inputs until a valid signal is found. Once a valid signal is found, the other input is ignored. If the signal is lost, the converter will once again start searching between the two inputs until a valid signal is found. It is always recommended to use the higher quality S-Video input when possible. Composite Output: This RCA or BNC connector provides the video output for all standards. For all electronic standards, this output should terminate into a 75 ohm load. For all mechanical standards, this output should terminate into a >20K load. For complete information about the characteristics of this output, please refer to the Specifications section found later in this manual. 7 Introduction Audio Input: This is a stereo capable, 1/8” audio input connector. It is used to compensate for the delay in the video caused by the internal processing of the converter by delaying the audio a similar amount. Further, when a mechanical standard is selected, the left audio channel becomes the optional mechanical synchronization input. In this mode an external sine wave can be used to synchronize the converter to the same power source as a mechanical television. Please refer to the Theory of Operation and Specifications sections found later in this manual. Audio Output: This is a stereo capable, 1/8” audio output connector. It is used to compensate for the delay in the video caused by the internal processing of the converter by delaying the audio a similar amount. Further, when a mechanical standard is selected, the left audio channel becomes the optional mechanical synchronization output. In this mode the converter can drive an external amplifier with a sine wave to synchronize a mechanical television. Please refer to the Theory of Operation and Specifications sections found later in this manual. Standard Selector (Multi-Standard Only): This switch is used to select the desired output standard. A list of these standards and their characteristics can be found in the Output Standards section found later in this manual. It is recommended that this setting be made before making connection to the converter, or powering it up. Internal Options Switch (Multi-Standard Only) The internal Options switch is shown below: U1 C26 C27 ON 1 2 3 4 U9 SW1 C5 C56 C66 8 Front of unit (Power Switch / LED) Back of unit (I/O connectors) Introduction The internal Option Switch (SW1) has four controls allowing the user to set the operating mode of the converter. In order to change the switch settings, the cover must be removed from the unit. To do this, first remove all cables from the unit, including the power cable. Place the unit on it’s top, and remove the two, small phillips screws from the bottom. Flip the unit back over, and remove the top being careful not to remove the front or back panels. Caution! In the subsequent steps, make sure you are touching one of the outer metal rings on the phono or audio jacks to discharge any static electricity that may be present before proceeding. There are static sensitive devices inside the converter that can be damage if subjected to a static discharge. Using a small tool such as a paper clip, carefully slide the desired switch to it’s new position being careful to not put an undue amount of force on the switch that might damage it or the circuit board. Once the desired switch settings have been achieved, replace the cover on the unit, and reinstall the two phillips screws. Switch 1 - NTSC - PAL/SECAM: This switch is used to selected the input mode and supported standards of the converter. When the switch is in the ON position the converter is in the NTSC mode. In this mode the converter will accept NTSC video and output any of the 16 standards listed for NTSC mode. Conversely when the switch is in the OFF position, the converter is in the PAL/SECAM mode, and output any any of the 16 standards listed for PAL/SECAM mode. Only the output standards listed for a given mode are available in that mode. If an output standard is only listed for one mode, than it is not possible to output in the other mode. Switch 2 - Mechanical Output Gamma Correction: This switch is used to enable/disable Gamma Correction when a mechanical output standard is selected. Standard NTSC or PAL/SECAM video has a gamma correction of approximately 0.45 applied to it at the source. This is done to compensate for the inherent gamma of a CRT of approximately 2.2. Neon and LED lamps used for mechanical televisions have a gamma very close to 1.0, so without a correction applied to the video output of the converter, the resulting image would have an incorrect gamma. When this switch is OFF, no gamma correction is applied to mechanical standards, and the resulting video output from the convert will have a 0.45 gamma. When this switch is ON, a gamma correction is applied to the video output resulting in a nearly 1.0 gamma. If the lamp driver used for the mechanical television already has gamma correction built in, this switch should be OFF. If the driver is linear and does not have gamma correction, this switch should be ON. 9 Introduction Switch 3 - 819/25i, 441/25i, 405/25i, 343/30i Equalization Pulses: This switch is used to enable/disable equalization pulses in the vertical (frame) sync for the 819/25i, 441/25i, 405/25i and 343/30i formats. When these formats were first created, they did not include equalization pulses in the vertical (frame) sync. Because of this, poor interlacing of the image can result due to difficulty of the vertical (frame) oscillator in the television properly locking to the signal. Equalization pulses are added before and after the serration (broad) pulses to improve the vertical (frame) oscillator's ability to lock to the sync signal. All modern analog video formats utilize equalization pulses in the sync. When this switch is ON, equalization pulses will be added to the listed formats for better synchronization of the television. While this does create a better interlaced image, it is not historically accurate, and does not represent the formats as they would have originally appeared. This should be the default position of this switch in most cases. When this switch is OFF, the equalization pulses will not be added, resulting in the formats as they would have been. This may result in a poorly interlaced image on the television, exactly as it would have been when originally broadcast. This mode can be used to better convey what these formats would have looked like originally. Switch 4 - Not Currently Assigned: This switch is not currently assigned, and is for future expansion. Typical Connections In normal usage, the supplied power adapter is connected to the converter and to the AC power source. The desired output standard is then selected on the Standard Switch. A valid NTSC or PAL video source should then be connected to one of the video inputs, preferably the S-Video input as it provides higher quality video than Composite does. The video source can be anything from a VCR to a DVD to an offair broadcast. For the best quality, a DVD is recommended. Since the converter will act as a TimeBaseCorrector in all modes, even poor signals like those from a consumer VCR can be used which normally can not be used with early television equipment due to their instability. This means that VHS tapes that would normally tear or be unstable on an early television can be played through the converter. This is especially true of the 525/30i (NTSC) or 625/25i (PAL) output standard which was included for this express purpose. Additionally, since early monochrome televisions do not contain chroma traps, the 3.58MHz (NTSC) or 4.43MHz (PAL) chroma subcarrier can appear as noise in the image. The converter will also strip this chroma signal out so that it can not interfere with the output video signal. The video output from the converter conforms to all applicable standards, and can 10 Introduction normally be connected to an RF modulator for the electronic standards, or an amplifier to drive the neon or LED lamp for the mechanical standards. If a direct connection is made to the video amplifier in most tube televisions, care should be taken so that no high voltages will be induced on the video signal or ground connection. This can be dangerous and cause damage to the convertor. The audio from the source device (VCR, DVD, etc.) should be routed through the converter and then on to the modulator. As with the video, if a direct connection to the audio amplifier is to be made on a tube television, care should be taken so that no high voltages will be induced on the video signal or ground connection. This can be dangerous and cause damage to the convertor. If a mechanical standard is selected, the left channel is no longer used for audio purposes, but for the synchronization signal for the mechanical television. In this case, a splitter cable should be used to connect mono audio to the right channel input, while the right channel output is connected to the audio load. While it is not totally necessary to use the “delayed audio” ability of this converter with electronic standards as the delay is not that great, it is highly recommended to use this feature in conjunction with mechanical standards as the delay can be very noticeable. The use of high quality video cables is recommended for best results. Cables conforming to 75 ohm impedance should be used on the video inputs and outputs. Cables of lesser quality can be used for the audio channels. There should now be a solid status light on the front panel indicating a locked video signal as described previously, and a stable image on the television. To help aid in setup, when no video input is presented to the converter, it will output a test pattern in the selected format. This can be useful in making final adjustments to the television. Theory of Operation In order to convert between different video standards, two main goals need to be realized, spatial correction and temporal correction. The first, spatial correction involves changing the resolution, size and aspect ratio of the incoming video to the output video format. This can be easily achieved through standard digital methods utilizing scalers and FIR filters. The second, temporal correction is much more complex, and involves the use of PLL’s and digital techniques to change the rate at which the video is output. Both corrections will be discussed in detail. It was decided that no off the shelf components existed that would provide the desired functionality, so a FieldProgrammableGateArray, or FPGA, was chosen to provide all the digital functionality. By adding input/output circuitry, memory, and ancillary circuitry such as PLL’s to the FPGA, the entire system could be realized. The basic building blocks to the design are; FPGA, video decoder (ADC), video DAC, video PLL, frame memory, image FLASH memory, audio codec and amplifiers, and multiple power supplies. A brief description of each part follows: 11 Introduction FPGA: Multi-Standard Converter Only: Xilinx XC2S300e-7pq208 300K gate equivalent 64Kb block RAM 1.8V core / 3.3V I/O CBS Color Only Converter: Xilinx XC2S100e-6pq208 100K gate equivalent 40Kb block RAM 1.8V core / 3.3V I/O Video Decoder: Philips SAA7113H 9bit ADC’s Line Locked Clock Composite and S-Video inputs Multiplexed 8bit YCrCb output bus Video DAC: TI THS5651A 10bit DAC 125 MSPS maximum conversion rate 79 db SNR Differential current output Frame Memories: Cypress CY7C1049CV33-15VC 512M x 8 Static RAM Image Memory: Atmel AT49BV322AT-70CI 4M X 8 FLASH ROM Video PLL: MicroClock MK1575-01 1.5 - 80MHz range Internal or External Feedback Counters Audio CODEC: TI PCM3008T 16 bit ADC/DAC 64X over sampled delta-sigma modulation 88 db SNR 8 - 48 KHz range 12 Introduction Topology A block diagram of the circuitry is shown below: Video PLL Standard Selector/ PushButton Audio Codec and Amplifier Composite Input Video ADC and Decoder Stereo Input Stereo Output Composite Output Video DAC, Filter and Driver FPGA S-Video Input Frame Memory A Frame Memory B Frame Memory C Image Memory (FLASH) The incoming video is digitized and processed by the SAA7113 using a 24.576MHz reference crystal to the ITU-601 (formerly known as CCIR601) specification. All internal timing is generated using this crystal. The video is quantized, processed for brightness, contrast, chroma gain and hue, among others, and output at the ITU rate of 27MHz on an 8 bit, time multiplexed bus, with alternating luma and chroma samples. No other signals are required from this circuit as the ITU specification describes a method for encrypting the horizontal and vertical timing information directly into the digital data using timing reference makers, or TRS’s. A brief description of the ITU-601 specification is as follows: Fundamental quantization frequency: 13.5MHz Pixel Resolution: 720 H x 486 V NTSC / 720 H x 576 PAL Image Aspect Ratio: 4:3 Pixel Aspect Ratio: 1.1 NTSC / 0.9 PAL Horizontal Frequency: 15,734 Hz NTSC / 15,625 Hz PAL Vertical Frequency: 29.97 Hz NTSC / 25 Hz PAL Clocks per Line: 1716 NTSC / 1728 PAL (27MHz clock) Clocks per Frame: 900900 NTSC / 1080000 PAL (27 MHz clock) Notice that the vertical frequency is 29.97Hz for NTSC, not 30Hz as expected. This is due to the NTSC color system that was first described in 1953. All monochrome television transmissions prior to this standard used 13 Introduction exactly 30Hz, or 30 frames per second, so as to be in sync with the AC line frequency. This was done to reduce distortions in the image due to induced AC fields or “hum” from the power supplies of these early sets. In order to devise a “compatible” color system that would show a monochrome signal on existing sets, RCA proposed a method of modulating the color components of the video signal onto a subcarrier in the video. For reasons beyond the scope of this manual, a frequency needed to be chosen so that no standing patterns in the color signal would result. This required lowering the vertical frequency from 30Hz to 29.97Hz. While this change caused no adverse side effects on televisions, it has created a legacy of problems for modern video equipment. Instead of being able to use integer numbers like 24, 25 and 30, we now have to include 29.97 which makes many calculations and conversion extremely difficult. For digital processing, the ratio 1000/1001 has been established as the conversion between 30 and 29.97 video. This will become important when discussing the temporal correction. The digital video data is then routed to the FPGA where it is further processed. It is first passed through a multiple tap FIR filter before horizontal downscaling. The FIR filter is required to remove the high frequency components from the video before downscaling, otherwise alias artifacts would be visible in the downscaled image. The amount of filtering and downscaling is controlled by which output standard is selected. The lowest amount being no filtering or downscaling (525/30i NTSC / 625/25i PAL standard), to a maximum of an 81 tap filter and 16:1 downscaling (24/15p standard). No vertical downscaling is done at this point. The filtered and scaled data is then sent to the frame memories in round robin fashion. The three frame memories are each large enough to hold a complete frame of video for any standard, so there are always three frames of video stored at any given time. All processing is done to the luma component only, except in the case of the 405/72i (CBS Color) standard which involves the additional step of color space converting the incoming YCrCb data to RGB data before processing and storing into the frame memories. All previous processing was done synchronous to the ITU clock, so no additional clocks were necessary. Before any further processing can be done, a new clock, synchronous to the output standard but integrally related to the ITU clock must be generated. The MK1575 is used for this purpose. By running it in it’s external feedback loop mode, the exact frequency and phase of it’s output can be determined by programmable counters in the FPGA. Depending on the selected output standard, the appropriate integer numerator and denominator for the PLL are selected. This will be discussed in more detail later. With this new output synchronous clock, a video timing generator, or flywheel is created in the FPGA to generate all timing signals for the selected output standard. All horizontal, vertical, pixel count and line count generation is done in this process. Additional for electronic standards, frame timing signals in the form of equalization and serration pulses are generated. Finally, for the 405/72i CBS Color standard, additional color sequence and “red field” equalization signal are generated. This is the 14 Introduction main “heartbeat” process for the entire design. Using these timing signals, the video data that was stored in the frame memories synchronous to the incoming 27MHz video clock can now be clocked out synchronous to the new output clock. Since the video has already been downscaled horizontally, it only needs to be further downscaled vertically before being output to the video DAC. The data is again routed though a FIR filter and downscaler that is controlled by the selected standard. This downscaler includes ratios from 1:1 (525-30i standard) to 20:1 (24/15p) including fractional ratios such as 3:4 for the 405/72i (CBS Color) standard. For mechanical standards, one additional step is taken before outputting the video. Since CRT displays have a characteristic gamma of 2.2, it was decided in the very early days of television to do a 0.45 “gamma correction” at the camera to compensate the image. The result is the brightness on the CRT closely matches the brightness of the original scene. Mechanical sets however use neon, or newer LED replacement lamps for illumination that have a gamma close to 1.0. Because of this, displaying raw video on one of these sets would result in an overall gamma of 0.45 causing the gray levels to be severely distorted. To eliminate this problem, all video goes through a 2.2 “gamma corrector” in the FPGA before being output. This gives the desired 1.0 gamma at the surface of the neon or LED lamp. With all the above timing now generated, the output video can be generated. The signals from the flywheel are routed to the video output DAC at the appropriate times in the signal, while the processed video from the frame memories are routed to the video DAC during the active portions of the video signal. Simultaneously audio from the audio input connectors is quantized at approximately 35KHz, delayed by the appropriate amount of time to offset the delay of the video due to internal processing, and sent out the audio output connectors. If a mechanical standard is selected, the left channel of the audio is used for timing purposes. A sine wave of the correct frequency and phase is output on the left audio channel to drive a phonic coil or AC inverter to keep the mechanical set in synchronism with the video. Additionally the left audio input channel can be used as a reference timing signal input to force the video output into synchronism with this external source. Detailed Analysis Generating an output clock: As shown in the previous section, a clock, synchronous to the output standard, but related to the incoming video clock must be generated. Using standard PLL techniques, it is a simple matter to generate this clock. Below is the block diagram of the PLL circuitry: 15 Introduction FPGA 27MHz Reference Counter Phase Comparator Feedback Counter Loop Filter VCO By carefully selecting the Reference and Feedback counter values, the desired synchronous frequency for the selected output standard can be achieved. Using the 405/72i CBS Color standard as an example, the counter values can be calculated as follows: First, it is desirable to keep this clock near the 27MHz incoming video clock for the best utilization of the available bandwidth. Using this fact, and a target horizontal blanking area of 18%, the best fit for the pixels per line is around 360 for: 72fps * 405 lines * 360 pixels/line * 1.18 HBL = 12.4MHz Using a 2X clock would yield a 24.8MHz rate, or 360 * 2 = 720 clocks for the active horizontal area and 720 * 18% = 130 clocks for the horizontal blanking area, for 850 clocks total per line. Knowing there are 900,900 clocks per frame in the NTSC signal, and the frame lasts 1/29.97 of a second, multiplying by 29.97/71.928 (output standards conform to the 1000/1001 ratio) yields exactly 375,375 NTSC video clocks per 405/72i frame. Finding the factors of 375,375 yields for the reference counter: 375,375 = 3 * 5 * 5 * 5 * 7 * 11 * 13 Calculating the number of clocks per frame in the 405/72i output standard yields the feedback counter of: 850 clocks/line * 405 lines = 344,250 clocks/frame Finding the factors of 344,250 yields: 344,250 = 2 * 3 * 3 * 3 * 3 * 5 * 5 * 5 * 17 16 Introduction Eliminating common terms yields: 2 * 3 * 3 * 3 * 3 * 5 * 5 * 5 * 17 3 * 5 * 5 * 5 * 7 * 11 * 13 Resulting in a final ratio of the reference counter to the feedback counter: 918 1001 Using these counter values and the 27MHz NTSC video clock, we can generate a 24.8MHz output clock that has exactly 344,250 clocks per frame, and is in exact synchronism with the NTSC rate. Spatial Correction: In order to convert between different video standards, the video image must first be spatially converted between the two standards. Many aspects need to be taken into account such as image aspect ratio, and the number of active lines. The distinction of image aspect ratio is made here to differentiate it from pixel aspect ratio which only has to do with how the analog data is quantized. On first inspection, it would appear that spatial correction can easily be achieved by simple adding or dropping pixels to get the desired result. For example, if you have 720 pixels in the input, and need 360 pixels in the output, dropping every other pixel would appear to be adequate. Unfortunately, half the original information is lost, not being included in the output in any way, and since the input pixels where sampled at a much higher frequency than the pixels are being output, aliasing of the image will occur. To avoid this situation, the pixel data must first be run through a low pass filter to remove any high frequency components above what the output frequency will be, then this filtered data can be resampled for the desired number of output pixels. In the analog world, this is done with a typical low pass filter. In the digital world, this is done with an FIR, or Finite Impulse Response filter. Both achieve the same end result, reduction of the high frequency components. Now that the data has been filtered, it can be scaled. For most of the electronic output standards this is a relatively easy task since they share a common 4:3 image aspect ratio. Using the 405/72i CBS Color format as an example, we know from the previous discussion that there are 360 active pixels per line, and 405 total lines. Using a vertical blanking area of 10% yields a desired active vertical area of 364 lines. This means that the active area for this standard is 360 x 364 pixels. Since ITU-601 NTSC has 720 x 486 pixels, we need to scale the horizontal by 2:1 and the vertical by 3:4. A three tap FIR filter is used to eliminate the high frequency components of the original quantized video data, and then sent to a 2:1 scaler. For the unique case of the CBS Color standard, the only color format supported, the data is also put through the following matrices to convert between the YCrCb and RGB color spaces: 17 Introduction R = Y + 1.371(Cr - 128) G = Y - 0.698(Cr -128) - 0.336(Cb - 128) B = Y + 1.732(Cb - 128) The data is then sent to the frame memories. During the active video areas in the output video, the data is read out of the frame memories, additionally scaled by 3:4 vertically to arrive at the final output resolution of 360 x 364. When working on mechanical standards, or any standard that varies from the incoming 4:3 aspect ratio, this also needs to be considered in the scaling of the video. For example, the 240/24p electronic standard has a 6:5 aspect ratio. By targeting 360 X 240 pixels for the output, we come close to the desired standard, with the exception that if this 4:3 image were to be displayed in it’s entirety on the 6:5 aspect ratio display, it would appear squeezed horizontally. Because of this, we discard 36 pixels on each line and 4 lines to yield a final resolution of 324 x 236 which has the correct 6:5 aspect ratio. Note: the original pixel aspect ratio of 1.1 must be a taken into account in this calculation to yield the correct result. Temporal Correction: As previously noted, temporal correction can be much more involved than spatial correction. Making the situation even more difficult is that fact that most high definition standards are interlaced. This means that a frame of video can not be handled as a single image since the two fields are actually separated in time. Since each field in the original video is in temporal order, they must stay in the correct temporal order during playback, or severe stuttering or jumping of the images will occur. There is no perfect solution for rate conversion, as all forms are a trade off between complexity/cost and the quality of the conversion. Many different approaches have been taken to solve this problem, and the best one depends on many different factors. Several different methods are employed by this converter depending on the source of the input video, and the selected standard for the output video. First, for any standard where the output frame rate matches the input frame rate, i.e. 525/30i NTSC to 343/30i or 625/25i PAL to 405/25i, no correction is required. The video is simply scaled spatially, and formatted to the output standard. Additionally, if the output is an even multiple of the input, i.e. 15 fps output from a 30fps input, simple correction is all that is required. In this case, frames are simply dropped to achieve the desired output rate. This process is known as decimation. While information in the original video is discarded, (that being the frames that are decimated) it is not missed since the resulting effect is no different then running a camera at a higher shutter speed than the frame rate. As long as each frame being output represent what the image should be at that point in time, the full temporal information can be conveyed. The second type to consider is when the selected output standard is of the progressive type. In this mode the input video is first decimated on a field basis to achieve the desired output frame rate. While this process can create frames that 18 Introduction consist of fields from two different original frames, and would be temporally out of order if played back to an interlaced standard, the process of de-interlacing, or converting the image to a progressive one, removes this temporal distortion. This is similar to the above method, and while image data from the original video is discarded, no visual loss occurs as long as the outputted frames are temporally correct. In the most complex scheme, an interlaced electronic output is selected that is not an even multiple of the video input frame rate i.e. 525/30i input to 405/25i output. In this situation, there are several methods that can be employed with varying degrees of success. The first method is to simply have the output video follow the input video at whatever the rate difference is. In this scheme, the output video will switch between input video fields at any arbitrary time in the frame. For instance, if the output is halfway through a field, and the input changes from one field to the next, the output would immediately change. While this is the simplest method, and used by most consumer converters and multi-standard VCR’s, due to the way interlaced video works, portions of the output will be temporally out of order, resulting in what looks like stuttering or tearing across the image. Due to of all these factors, this is not a good choice for this converter. Skipping to the most advanced method which is known as motion estimation, it offers the highest degree of image quality most of the time, but at an enormous cost in complexity and price. In this method, two frames are analyzed by a high speed computer algorithm which attempts to dissect the image into moving and stationary objects. The motion of these moving objects is calculated and their position at an arbitrary point in time calculated. The new frames for the output rate are constructed from this data. As can be seen, this is an extremely complicated process, and can be prone to errors. Many factors can complicate this method such as; what does the object look like; is the background moving, etc. In it’s best case, this method can provide excellent results obtaining a smooth, natural looking output. In it’s worst case, objects can temporarily disappear, or be placed in the wrong position. Because of all these factors, this is not a good choice for this converter. This is the method used by some extremely high end video processors costing upwards of $100,000 USD (2004). The method chosen for this converter is one that provides excellent looking output images, with minimal (usually imperceptible) motion artifacts and image degradation during fast motion. In this preferred method, interpolated fields in the output are made up of combining several fields from the input. This allows for extremely smooth and natural looking video, with minimal motion artifacts, and softening of the image. For static or slowly moving images, this method does not degrade the image at all, and for fast moving objects, it has the tendency to decrease the image sharpness very slightly. This is the method used by most professional frame rate converters, and with a proprietary algorithm unique to this converter, provides the best all around image quality and cost of implementation. There is one more method that can be used when in the NTSC mode of the converter, and exploits a phenomenon unique to the standard. As it turns out, discarding one field every five to yield a desired 24fps or 72fps output rate from the 30fps input will result in no loss of data in most cases. This is due to a fact that was 19 Introduction discovered very early in television development, and is described below. As can be seen by the supported standards of this device, early American electronic television used a 24fps frame rate to match what was typically used by film. This was done to simplify the broadcasting of films which was the major source of material at the time. Two problems arose using this frame rate. The first is that the frame rate is so low, the perceived flicker in the image is objectionable to a human viewer. It was found early in the development of film projectors that any rate less than about 48fps caused objectionable flicker to a viewer. In film, this problem was solved by placing a “cutter” wheel in front of the lens that spun at twice the rate the film was running at, fooling the viewer into perceiving 48fps. The second problem with using 24fps for electronic televisions is that most AC power grids run at a rate of 60Hz. This means that any stray AC fields, or unfiltered hum that gets through the televisions power supplies will result in rhythmic distortion of the image. Because of these two issues, it was quickly realized that a higher frame rate would be required, but higher frame rates require equally higher bandwidths. To overcome this new problem, interlaced scanning was devised that draws the image on the screen in two successive passes, typically 60 fields/sec or 30 full frames/sec, fooling the viewer into seeing a higher field rate, but keeping the required bandwidth down. This solved both the flicker and AC interference problems, but now films could not be shown without some kind of conversion. At first it would seem obvious to just duplicate every fourth film frame so that four film frames create five video frames, yielding the desired 24fps to 30fps conversion. (shown below) 4 Film Frames 24 fps A B C D A 1 A 2 B1 B2 C1 C2 D1 D2 D1 D2 5 Video Frames 30 fps Unfortunately the human eye is very sensitive to this duplication of images, and this was found to be unacceptable. However, since there are actually 60 fields/sec in interlaced video, a single field every four can be duplicated with much less visibility to the human eye. Since this yields a pattern of 2 fields followed by 3 repetitively, it is called “3:2 pulldown”, and is still how film is shown on NTSC to this day. (shown below) 20 Introduction 4 Film Frames 24 fps A B C D A 1 A 2 B1 B2 B1 C2 C1 D2 D1 D2 5 Video Frames 30 fps Because 2 fields are identical every 5 fields, (i.e. B1 and D2) this process can be detected and removed from the video. By comparing successive fields as they are quantized, the original 24fps film can be completely reconstructed and used for the 24fps based output standards. Because of this process, even though one field is discarded every five fields, no data is lost from the original material. It should be noted that while this method is very effective at recovering the original 24fps material, it only works if the “cadence” or ordering in which the fields were created is in the proper order, and does not break sequence. This is true of any commercial DVD, or uncut movie on broadcast television. Unfortunately, many network television shows are edited after this 3:2 pulldown process is performed, meaning if the editor is not extremely careful, the cadence will be broken at each edit point. Additionally, to cram more commercials into network television, a method called “time compression” is performed on the original video to speed it up slightly so more commercials can be placed in the broadcast. This has the effect of destroying the cadence of the original material, rendering this method totally ineffective. When this happens, the output image will appear to stutter, and then smooth out repeatedly as it obtains and then looses cadence lock. Additional Mechanical Processing: Because video signals for mechanical television do not contain synchronization pulses as electronic television does, mechanical televisions had two methods for synchronization to the transmitter. The first method was to use a fully synchronous AC motor for both the transmitter and receiver. If both devices were operated on the exact same power grid, they would be inherently in synchronization with each other. This was the simplest method of synchronization used for mechanical television, but has the obvious severe limitation of the receiver needing to be on the exact same power grid as the 21 Introduction transmitter. The second method was to recover a portion of the received video signal, filter out it’s strong characteristic horizontal frequency, amplify and apply this signal to a small synchronous motor mounted to a larger motor. This smaller synchronous motor was known as a “phonic” coil. This allowed a large, powerful, non-synchronous motor to spin the disk up to near-synchronous speed, and then have the smaller, phonic coil pull the disk the remaining amount, and into lock. This method worked very well, and did not require the transmitter and receiver to be on the same power grid, allowing reception over a greater distance. To accommodate both of these methods, this converter supplies two different mechanisms. First, the left audio channel is not used for regular audio when a mechanical standard is selected. Instead, an appropriate synchronous sine wave of about 3Vp-p is generated, which can then be amplified and sent to a phonic coil, or sent to an AC inverter to run a synchronous motor. While running a phonic coil in this manner is a simple matter, obtaining an AC inverter with the required power and frequency input is not a simple matter. Because of this, an additional method is employed. This involves supplying a reference sine wave to the left audio input channel to cause the video output to be retimed. This method allows an AC synchronous motor type mechanical television to be run directly off of any AC main circuit with a reference of this AC sent to the left audio input of the converter. This reference should be about 6-9 VAC, and connected through a series limiting resistor of approximately 4.7K. Additionally a small capacitor can be added to reduce false triggering on spurious signals. Audio can still be routed to the right channel input of the converter for use with the internal audio delay compensation feature of the converter. A typical circuit is as follows: 4.7K AC Line (Mains) 6 to 9 VAC Transformer Optional 0.1uF 1/8” Jack from Audio 22 1/8” Plug to Converter Specifications Specifications Video Inputs: Supported Standards: NTSC M, 29.97fps / PAL B,G,H,I, 25fps / SECAM 25fps Video Quantization: 9bit A/D, 8 bit data Video Input: Composite - 1Vpp, 75 ohm impedance S-Video Y - 1Vpp, 75 ohm impedance S-Video C - 600mVpp, 75 ohm impedance Video Output: Video Output: Composite - 1Vpp into 75 ohm for electronic standards 2Vpp into 20k for mechanical standards Video Quantization: 10 bit D/A Video Levels: +/- 3% of selected output standard Video Timing: +/- 50 ppm, crystal/PLL locked, < 2ns jitter typical Video SNR: > 54db typical Audio I/O: Audio Sample Rate: 35.1kHz Audio Quantization: 16 bit Audio Input: Unbalanced, > 20K impedance 3Vpp maximum Audio Output: Unbalanced, < 200 ohm impedance 3Vpp maximum Audio Response: 20Hz to 16kHz, +/- 2db Audio SNR: > 85db typical 23 Specifications General: Dimensions: 4.50” X 3.00” X 1.25” (177mm X 76mm X 32mm) Power Requirements: 6Vdc maximum input, 4 watts maximum Humidity: 20% - 80% non-condensing Temperature: 10C - 45C ambient (50F - 110F) 24 Supported Output Standards Supported Output Standards - NTSC Mode Multi-Standard Converter Selector Switch Setting, NTSC Input Mode: 0 - 525/30i interlaced electronic (TimeBaseCorrector, NTSC 1941-Present) 1 - 441/30i interlaced electronic (pre-war U.S. standard, 1937-1941) 2 - 343/30i interlaced electronic (RCA experimental, 1934-1936) 3 - 819/25i interlaced electronic (France, 1949-1983) * 4 - 625/25i interlaced electronic (PAL, 1950-Present) * 5 - 405/25i interlaced electronic (EMI, U.K., 1936-1984) 6 - 240/24p progressive electronic (RCA experimental, 1933-1934) 7 - 120/24p progressive electronic hybrid (RCA experimental, 1931-1932) 8 - 60/20p progressive mechanical (RCA/Jenkins, 1930-1934) 9 - 48/20p progressive mechanical (GE experimental, 1931) 10 - 48/15p progressive mechanical (Jenkins, 1928-1931) 11 - 45/15i triple interlaced mechanical (Western Television/Sanabria, 1929-1932) 12 - 24/15p progressive mechanical (GE experimental, 1928) 13 - 32/12.5p progressive mechanical hybrid (NBTV, 1995-Present) 14 - 30/12.5p progressive mechanical (Baird, U.K, 1928-1932) 15 - 30/12.5p progressive mechanical (TeKaDe, Germany, 1930) Note: Electronic standards for the NTSC mode have two (2) custom image locations except those denoted with an * which only have one (1) location. Mechanical standards have one (1) custom image location. 25 Supported Output Standards Supported Output Standards - PAL/SECAM Mode Multi-Standard Converter Selector Switch Setting, PAL/SECAM Input Mode: 0 - 525/30i interlaced electronic (NTSC, U.S. 1941-Present) * 1 - 441/30i interlaced electronic (pre-war U.S. standard, 1937-1941) * 2 - 819/25i interlaced electronic (France, 1949-1983) 3 - 625/25i interlaced electronic (TimeBaseCorrector, PAL 1950-Present) 4 - 441/25i interlaced electronic (Germany/France, 1935-1956) * 5 - 405/25i interlaced electronic (EMI, U.K., 1936-1984) 6 - 240/25p progressive electronic hybrid (Baird, U.K., 1936) * 7 - 180/25p progressive electronic (TeKaDe, 1934-1939) * 8 - 120/25p progressive mechanical (TeKaDe/Fernseh-A.G., 1932) 9 - 96/25p progressive mechanical (Telefunken, 1932) 10 - 90/25p progressive mechanical (TeKaDe, 1932) 11 - 60/25p progressive mechanical (France, 1935) 12 - 50/25p progressive mechanical (Marconi experimental, 1932) 13 - 32/12.5p progressive mechanical hybrid (NBTV, 1995-Present) 14 - 30/12.5p progressive mechanical (Baird, U.K, 1928-1932) 15 - 30/12.5p progressive mechanical (TeKaDe, Germany, 1930) Note: Electronic standards for the PAL/SECAM mode have two (2) custom image locations except those denoted with an * which only have one (1) location. Mechanical standards have one (1) custom image location. 26 Supported Output Standards Supported Output Standards - CBS Color Model CBS Color Converter: 405/72i interlaced electronic/mechanical color hybrid (CBS 1950-1953) 27 NTSC Mode Multi-Standard Output Specifications NTSC Mode Multi-Standard Output Specifications Mode 0 - 525/30i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 1.13 [0.88] 27.0 MHz 720 / 480 15,734 Hz / 29.97 Hz Electronic Interlaced 6.5 MHz [5.25 MHz], 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio Mode 1 - 441/30i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 1.35 [1.04] 22.7 MHz 720 / 400 13,217 Hz / 29.97 Hz Electronic Interlaced 5.7 MHz [4.4 MHz], 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio Mode 2 - 343/30i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines : Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 1.70 [1.37] 26.5 MHz 720 / 316 10,280 Hz / 29.97 Hz Electronic Interlaced 4.4 MHz [3.5 MHz], 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio Mode 3 - 819/25i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 0.59 27.0 MHz 576 / 738 20,455 Hz / 24.975 Hz Electronic Interlaced 6.5 MHz, 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio 28 NTSC Mode Multi-Standard Output Specifications Mode 4 - 625/25i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 0.94 [0.75] 27.0 MHz 720 / 576 15,610 Hz / 24.975 Hz Electronic Interlaced 6.5 MHz [5.25 MHz], 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio Mode 5 - 405/25i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 1.43 [1.14] 26.7 MHz 720 / 378 10,115 Hz / 24.975 Hz Electronic Interlaced 4.5 MHz [3.4 MHz], 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio Mode 6 - 240/24p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 6:5 / 2.29 27.8 MHz 648 (cropped 720) / 236 (cropped 240) 5,754 Hz / 23.976 Hz Electronic Progressive 2.0 MHz [1.6 MHz], 1Vpp into 75 ohms, composite sync, 70/60 video/sync ratio Mode 7 - 120/24p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 6:5 / 2.33 26.2 MHz 324 (cropped 360) / 116 (cropped 120) 2,877 Hz / 23.976 Hz Electronic Progressive 0.50 MHz, 1Vpp into 75 ohms, composite sync, 70/60 video/sync ratio 29 NTSC Mode Multi-Standard Output Specifications Mode 8 - 60/20p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 6:5 / 2.22 21.5 MHz 160 (cropped 180) / 60 1,199 Hz / 19.98 Hz Mechanical Progressive Left to Right, Top to Bottom 59.94 Hz ref output / 60 Hz ref input 96 KHz, 2Vpp into 20K, sync through left audio Mode 9 - 48/20p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines : Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 6:5 / 2.78 26.9 MHz 160 (cropped 180) / 48 959 Hz / 19.98 Hz Mechanical Progressive Left to Right, Top to Bottom 59.94 Hz ref output / 60 Hz ref input 77 KHz, 2Vpp into 20K, sync through left audio Mode 10 - 48/15p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 6:5 / 2.78 23.9 MHz 160 (cropped 180) / 48 719 Hz / 14.985 Hz Mechanical Progressive Left to Right, Top to Bottom 59.94 Hz ref output / 60 Hz ref input 57 KHz, 2Vpp into 20K, sync through left audio Mode 11 - 45/15i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 1:1 / 2.84 22.1 MHz 128 (cropped 180) / 45 (cropped 48) 675 Hz / 14.985 Hz Mechanical Triple Interlaced Right to Left, Top to Bottom 59.94 Hz ref output / 60 Hz ref input 43 KHz, 2Vpp into 20K, sync through left audio 30 NTSC Mode Multi-Standard Output Specifications Mode 12 - 24/15p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 6:5 / 2.78 20.3 MHz 80 (cropped 90) / 24 360 Hz / 14.985 Hz Mechanical Progressive Left to Right, Top to Bottom 59.94 Hz ref output / 60 Hz ref input 14 KHz, 2Vpp into 20K, sync through left audio Mode 13 - 32/12.5p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 2:3 / 0.40 25.3MHz 32 (cropped 60) / 120 12.488 Hz / 399.6 Hz Mechanical Progressive Bottom to Top, Right to Left 399.6 Hz ref output 26 KHz, 1Vpp into 75 ohms, composite sync Mode 14 - 30/12.5p Baird: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 3:7 / 0.58 24.2 MHz 30 (cropped 90) / 120 12.488 Hz / 374.6 Hz Mechanical Progressive Bottom to Top, Right to Left 374.6 Hz ref output 24 KHz, 2Vpp into 20K, sync through left audio Mode 15 - 30/12.5p TeKaDe: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 4:3 / 2.25 24.2 MHz 90 / 30 374.6 Hz / 12.488 Hz Mechanical Progressive Left to Right, Top to Bottom 374.6 Hz ref output 18 KHz, 2Vpp into 20K, sync through left audio 31 NTSC Mode Multi-Standard Output Specifications Note: numbers in brackets [] are effective parameters. 32 PAL/SECAM Mode Multi-Standard Output Specifications PAL/SECAM Multi-Standard Output Specifications Mode 0 - 525/30i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 1.13 [0.90] 27.0 MHz 720 / 480 15,750 Hz / 30 Hz Electronic Interlaced 6.5 MHz [5.25 MHz], 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio Mode 1 - 441/30i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 1.35 [1.08] 27.1 MHz 720 / 400 13.230 Hz / 30 Hz Electronic Interlaced 6.5 MHz [4.4 MHz], 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio Mode 2 - 819/25i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 0.59 27.0 MHz 576 / 738 20,475 Hz / 25 Hz Electronic Interlaced 6.5 MHz, 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio Mode 3 - 625/25i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 0.94 [0.75] 27.0 MHz 720 / 576 15,625 Hz / 25 Hz Electronic Interlaced 6.5 MHz [5.25 MHz], 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio 33 PAL/SECAM Mode Multi-Standard Output Specifications Mode 4 - 441/25i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 1.41 [1.13] 28.1 MHz 720 / 382 11,025 Hz / 25 Hz Electronic Interlaced 4.7 MHz [3.7 MHz], 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio Mode 5 - 405/25i: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 1.43 [1.14] 26.9 MHz 720 / 378 10,125 Hz / 25 Hz Electronic Interlaced 4.5 MHz [3.4 MHz], 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio Mode 6 - 240/25p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 1.96 26.9 MHz 576 / 220 6,000 Hz / 25 Hz Electronic Progressive 1.9 MHz, 1Vpp into 75 ohms, composite sync, 50/50 video/sync ratio Mode 7 - 180/25p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 2.45 25.6 MHz 576 / 176 4,500 Hz / 25 Hz Electronic Progressive 1.4 MHz, 1Vpp into 75 ohms, composite sync, 50/50 video/sync ratio 34 PAL/SECAM Mode Multi-Standard Output Specifications Mode 8 - 120/25p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 4:3 / 2.20 25.3 MHz 352 / 120 3,000 Hz / 25 Hz Mechanical Progressive Left to Right, Top to Bottom 3000 Hz ref output / 50 Hz ref input 528 KHz, 2Vpp into 20K, sync through left audio Mode 9 - 96/25p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 4:3 / 2.75 20.3 MHz 352 / 96 2,400 Hz / 25 Hz Mechanical Progressive Left to Right, Top to Bottom 2400 Hz ref output / 50 Hz ref input 422 KHz, 2Vpp into 20K, sync through left audio Mode 10 - 90/25p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 4:3 / 2.93 25.3 MHz 352 / 90 2,250 Hz / 25 Hz Mechanical Progressive Left to Right, Top to Bottom 2,250 Hz ref output / 50 Hz ref input 396 KHz, 2Vpp into 20K, sync through left audio Mode 11 - 60/25p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 4:3 / 2.20 25.3 MHz 176 / 60 1,500 Hz / 25 Hz Mechanical Progressive Left to Right, Top to Bottom 1,500 Hz ref output / 50 Hz ref input 132 KHz, 2Vpp into 20K, sync through left audio 35 PAL/SECAM Mode Multi-Standard Output Specifications Mode 12 - 50/25p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 4:3 / 2.64 24.64 MHz 176 / 50 1,250 Hz / 25 Hz Mechanical Progressive Left to Right, Top to Bottom 1,250 Hz ref output / 50 Hz ref input 110 KHz, 2Vpp into 20K, sync through left audio Mode 13 - 32/12.5p: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 2:3 / 0.33 26.2 MHz 32 (cropped 60) / 144 12.5 Hz / 400 Hz Mechanical Progressive Bottom to Top, Right to Left 31 KHz, 1Vpp into 75 ohms, composite sync 70/30 video/sync ratio Mode 14 - 30/12.5p Baird: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 3:7 / 0.49 21.7 MHz 30 (cropped 90) / 144 12.5 Hz / 375 Hz Mechanical Progressive Bottom to Top, Right to Left 375 Hz ref output 28 KHz, 2Vpp into 20K, sync through left audio Mode 15 - 30/12.5p TeKaDe: Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines: Horizontal / Vertical Frequency: Scan Type: Sync Frequency: Video Characteristics: 4:3 / 2.25 24.2 MHz 90 / 30 375 Hz / 12.5 Hz Mechanical Progressive Left to Right, Top to Bottom 375 Hz ref output 18 KHz, 2Vpp into 20K, sync through left audio 36 PAL/SECAM Mode Multi-Standard Output Specifications Note: numbers in brackets [] are effective parameters. 37 CBS Color Output Technical Specifications CBS Color Output Technical Specifications Image / Pixel Aspect Ratios: Pixel Clock: Active Pixels / Lines : Horizontal / Vertical Frequency: Scan Type: Video Characteristics: 4:3 / 0.74 24.8 MHz 360 / 364 29,131 Hz / 71.928 Hz Electronic Interlaced, Field Sequential Color 6.2 MHz [5.25 MHz], 1Vpp into 75 ohms, composite sync, 70/30 video/sync ratio Note: numbers in brackets [] are effective parameters. 38 Firmware Revision History Firmware Revision History Revision 4.0, June 7, 2007: 1) All mechanical standards now support external reference. Revision 3.9, November 4, 2006: 1) Improved 120/24p and 240/24p standards to better match original specifications. Revision 3.8, November 11, 2005: 1) Increased maximum pixel frequency to 34.5MHz to facilitate 3 line interpolation on the 441/30i standard. Revision 3.7, August 30, 2005: 1) Corrected error in RAM CE generation logic that could cause the RAM not to be enabled during storing to the Image Flash. This error did not affect operation in any way, only that when trying to store an image into the Image Flash, there was a 50% chance that the image would not get stored. Revision 3.6, February 22, 2005: 1) Added dynamic power down control to the RAM chips to lower overall power consumption of the converter. Revision 3.5, December 10, 2004: 1) Changed many of the standards for higher output resolution. Due to the sampling nature of digital methods, the number of pixels along the line in the direction of scanning must be made higher relative to the number of lines to better approximate what was capable in a fully analog system. While a certain number of pixels are required to achieve a given level of bandwidth, this does not take into account pixel straddling effects that are not present in fully analog systems. The “Kell Factor” was designed to describe this effect, and is generaly accepted that given a certain number of sampling points, approximately 0.7 times this number is the actual resolution. While this is true, it only represents the perception limit of 39 Firmware Revision History resolution, and not what an analog system was capable of. To better approximate an analog system, the converter now uses a Kell Factor of not more than 0.4 in the scan direction relative to a 0.7 Kell Factor for the number of lines on any stanadrd that can allow for this. Revision 3.4, November 29, 2004: 1) Corrected internal parameters for Baird 30 line and TeKaDe 30 line standards. Revision 3.3, November 20, 2004: 1) Corrected internal parameter for generation of mechanical synchronization sine wave. Revision 3.2, November 11, 2004: 1) Added option to enable/disable equalization pulses in standards that did not originally use them. This allows the user to choose between better interlacing, or the original waveform. Revision 3.1, November 10, 2004: 1) Added “Zoom” functionality to cut a 4:3 size image out of the center of a 16:9 source and fill the screen. This function is available only for electronic standards. Revision 3.0, November 1, 2004: 1) Initial production release. Note: Firmware revision level can be found on the label on the bottom of the unit. 40