This is from my friend and colleague Robert Childs regarding the dispute. There is a lot here, so please be patient with yourself, and when you have time read with a open mind....
Billy Mitchell Donkey Kong Dispute my tech analysis
NORMALLY I do not get involved in disputes and arguments especially one as silly as this, however, as a witness to the event and technician who performed the game/equipment set-up for Billy Mitchell, I can attest Billy did not submit his record scores playing on MAME. It all boils down to this. When you hook up a 1980s era Donkey Kong board to a 2000s era RGB to NTSC converter board and then run it to a 1970s technology VCR for recording God only knows what it will do to the signal. Billy is my childhood friend, however, I am not writing this post because Billy is my friend but rather it is the truth.
The following is a LAYMAN’S explanation of the hook-up used for each of Billy Mitchell’s scores
Also below is posted a TECHNICAL explanation for those who are so inclined.
In the comparison between Weibe’s video and Billy’s in the side by side, the differences are apparent. The purpose here is to explain the difference. Those unfamiliar with the explained method of hook-up to record have raised an unnecessary firestorm of innuendo and accusation. Said firestorm could easily have been avoided if Billy Mitchell had first been asked instead of first being accused. This explanation and the more detailed TECHNICAL explanation below will answer the claims of those unfamiliar in this method of recording.
In Wiebe’s video, he used a camera on a tripod angled slightly behind him. This showed on the recording the screen as Weibe actually saw it. This is the common way of filming a score submission.
In Billy’s video recording, he did not use MAME. Billy used a RGB to NTSC Video Converter Board to record his footage. I still have the original receipt for this board which I can post if asked . He paid $200 for the board.
The way the video converter board works is that it TRANSLATES the entirety of the RGB type signal, which is what originates from the DK video game board, into NTSC when that signal is run through the converter. The NTSC signal is what VCRs and TVs of all types used back in the day. The entire reason Billy used a Video Converter was for direct feed after MY expensive camera was once destroyed when it was knocked over. Cameras were nearly $1000 in those days versus a safely out-of-the-way $200 video converter board.
It is important to understand that the visual Billy saw on his DK monitor is EXACTLY the same type RGB display that Weibe saw on his monitor (and ultimately was recorded by his camera).The converted video offers NO advantage to the player as the player never sees it during gameplay. The converter signal is sent solely to the VCR.
The Converter hook-up works like this:
The video converter is connected directly to the video output of the DK inverter board. The inverter receives the video from the DK board itself NOT the Sanyo monitor that the player sees from his view. The converter board does its job translating RGB to NTSC which goes to the VCR and is recorded as the finalized media. There is NO possibility of an advantage to the player using this method of recording.
The TRANSLATION of RGB to NTSC changes the video’s appearance and characteristics of the signal (as in a CGA monitor) to a “more MAME looking signal outputâ€. However, if you want to make the argument this recorded video signal APPEARS more MAME looking, maybe it does. Looking MAME is not MAME, nor a violation that even approaches “cheatingâ€.
Why would any player, nearly eight years ago, desire to put a MAME computer in the back of a Nintendo cabinet and connect the output to a Sanyo monitor? No one playing MAME had approached the 1 million mark at that time. The accusation of using MAME doesn’t make sense. It would be too much work. Anyone who could reach 1 Million on MAME could undoubtedly reach MORE on an arcade machine. The alleged deception does not make sense. The risk would not present enough gain over the possibility of being discovered.
Even if …
multiple eyewitness testimony is discounted (which it should not be),
Even if …
TG would dismiss the video footage that was in their possession which is TG’s responsibility to produce since they initially certified the record (if they dismiss THIS record because they can’t “find the tape†then how many thousands of other records/scores MUST they AUTOMATICALLY dismiss upon any challenge?). TG has a responsibility to honor tapes that were once in their custody, regardless of ownership changes, especially if they cannot “find†a tape that was in their charge to protect in the “archivesâ€,
Even if …
the prescribed method of verification at that time was put aside (which would be “changing the rules after the acceptance of proofâ€),
THE FACT REMAINS that the hook-up described here produces these results and that alone is enough to prove, along with Billy’s ability in other high scores, that cheating was not involved.
$5000 CHALLENGE
If ANYONE can disprove what I have stated within this explanation, not in opinion, but in ACTUAL proof I will donate $5000 in their name to the charity of their choice . We also will test/prove any discrepancies in opinion in my shop on video. A contract will be provided if this option is sought.
Regards
Robert Childs
TECHNICAL EXPLANATION
A TG member has made an accusation that Billy Mitchell used MAME instead of a legitimate and original Donkey Kong board. He assumed that MAME was used based on comparisons of three sources that were not equal. That TG member used Gleed’s direct video feed recorded at 60fps, and used his own camera phone that recorded at 60fps, and compared it to Billy’s direct feed video tape (30fps, non TBC (please read for more information
http://www.digitalfaq.com/…/video-r…/2251-tbc-time-base.html)), which was encoded to digital format, to make an assumption that cheating was involved since Billy’s video does not behave as his standard. That TG member wrongly concludes that due to the rotation of the image and its alignment to the left instead of the right, that MAME must have been used. However his argument is weak for he either leaves out nor considers that the Sanyo monitor used is known to have capacitor problems in the vertical circuit (namely c407 (10uF, 160v) which often causes a fold to the bottom half of a Horizontally mounted monitor.
Nor does he consider other plausible explanations such that the majority of these games from that era were notorious for image burn on the face of picture tubes and that most likely the picture tube was changed and the donor tube’s yoke was either installed 180 degrees the wrong way and the header connector was flipped to correct the orientation, or the yoke may not have been wedged in correctly and the picture was off center, or that the yoke was that of the donor monitor and was close enough to work with that monitor chassis. (Please read
http://www.junknet.net/tube-swap-diy-part1)
That TG member made a comparison to his own 60 fps video shot with a high definition Samsung S8 to Gleed’s direct feed captured at 60 fps (which is assumed to be a direct digital recording in mp4), to a direct DK board to video recorded on a 30 fps VHS tape which was then encoded into an unknown video format. That TG member makes no mention of the technology that is used to convert digital TTL RGB to analog low level composite, nor does he mention the inferior quality and losses of a VHS recording and how that might be problematic to his comparison.
The main issue here to consider is that TG member compares video recorded in high definition to a video that was converted from TTL RGB to low level composite video. Let’s explain the evidence and why it’s important to keep Billy’s direct feed and type of recorded media in mind.
Billy’s direct feed was born out of a necessity to avoid people tripping over a camera on a tripod pointed at the screen and I was also tired of hearing “Rob watch the cameraâ€. In the last few years cameras have become cheaper and more lightweight, but the situation a dozen or so years ago was not the same. Direct feed was a viable option. It's a very simple process to wire and does not take much skill. To wire a DK machine to an external display you would use an RGB to composite signal converter.
In the very early 2000s, Two Bit Score (a company that sold coin op video accessories), sold such a device. It cost $199.00. This was a simple device where you took standard positive digital TTL RGB video signal and a composite negative sync from a standard 15.7 kHz video game and converted them to a composite low level analog video signal that a TV or VCR could display. The quality of this reproduction while acceptable, never truly reproduced the original (please read this link to see note of this known issue.
http://www.jrok.com/hardware/RGB-FAQ.html).
To make a connection to this RGB adapter from a Classic Nintendo coin op you needed to invert the TTL output from the negative signal to a positive signal. Please pay attention to the TTL acronym. TTL means Transistor-Transistor Logic (current driven switching), which was the architecture of integrated circuits of the time (1980’s). It was inefficient, slow, and bulky. The negative signal that was generated from the board was fed into a buffer transistor and then to a driver transistor that drove the individual R, G, or B guns on the neck board of the monitor (Please search out and see schematic online as I don’t have time to post I have a family to get back to today).
In order to successfully output the correct colors to an RGB to composite signal converter, you need to invert the colors. Billy’s machine utilized Nintendo’s built in video inverter board, mounted stock above the horizontal flyback transformer, to do that. Normally this board is wired in “pass though†by default. With the movement of the output wire to this board to the inverted pin header, the games image displayed white as black and vice versa. This board is a very simple device composed of a couple potentiometers, a few resistors, a couple capacitors, some diodes for the power supply, and some 2N3904 NPN transistors. In short, the video was inverted from one state to its opposite state with two transistors: one for buffering, the other for inversion. Just by simply using this board you introduced some delay due to the slow switching nature of transistors. Now I bring this information to you for this reason: it is possible to invert the video signal by either rewriting the color PROMs on a stock board or by utilizing a CMOS 74HC IC with a few resistors and get far superior results. For those who do not know CMOS (Complementary Metal Oxide Semiconductor) is much faster than TTL. Since it is possible to invert the original DKs by other means, which scheme did Gleed use for his direct feed is not stated nor is it considered.
The adapter, used to make the video (from Two Bit Score), had the IC’s number scratched off, so I can’t give you exact information on its construction. This company was notorious for removing identifying information from their products to prevent people from copying them. However a quick google search will produce a popular IC, AD775, which is probably what their board was built on (Please see data sheet for more information-http://www.analog.com/…/technical-doc…/data-sheets/AD725.pdf).
Here is the important part. The simple explanation is that by inserting an RGB signal into this device you introduce loss, delay and latency. This is a by-product of the AD775 process. If you take a look at the IC’s block diagram you’ll see the many processes that are utilized in order to convert the signal to a low level video signal.
The simplest explanations of the AD725 conversion process are as follows. If you simply follow each signal line in the function block diagram(get online) you will see that each RGB color signal is fixed to an equal level, inserted into an RGB (lumence matrix) to YUV (basically YCbCr colors), encoding matrix. From there the Y signal is put in to a 3 pole low pass pre-filter and a composite sync is added, the U and V signals are put to a 4 pole low pass filter to remove the harmonics of the switching modulation and clamped down to a fixed level and then color burst vectors are added to the U and V signals. The U and V signals are used to modulate a pair of quadrature clocks (sine and cosine) at one-fourth the reference frequency input 3.579 545 MHz for NTSC. In order to be time-aligned with the filtered chrominance signal path, the luma signal must be delayed before it is output. The AD725 uses a sampled delay line to achieve this delay.
A layman explanation of the inner works of the AD725 IC would simply be: Signals are inserted, balanced, inverted, and filtered (where noise is removed). In the case of the Y signal it is sampled and delayed in order to synchronize colors then again filter. In the case of U and V signal, the combined signals are summed and then again filtered. The entire Y U and V signal is then summed to produce the single composite output.
SUMMARY
To surmise what the converter board does in its simplest form: You put in a signal, you balance the signals, you invert them, you filter them, you delay, you filter them again, you add timing signals, and you sum them again. This entire process distorts and adulterates the original input thus no comparison can be made by recording the original RGB signal at 60fps and then comparing that to that same signal that had been converted with the AD725 and then recorded to a VCR (which has 3 MHz of video bandwidth and horizontal resolution which is 240 lines per picture height).
Simple put: THE INFORMATION IS JUST NOT THERE! You are comparing two unequal sources!
That concludes both the Layman and Technical explanation. Please take time to examine the intricacy and accuracy. I have a DK machine along with a two-bit score converter board,and VCR if you would like to see Billy play another game on this setup and see the results this can be arranged in my store in Fort Lauderdale .
http://www.digitalfaq.com/forum/video-restore/2251-tbc-time-base.htmlThanks for Reading
Robert Childs
Arcade Game Sales
Fort Lauderdale, Florida
