|
|||||||||
| For recording school events, we generally
use Sony TR913E camcorders as cameras only, connected to
a vision mixer. These camcorders were bought for the
school by the Parents' Association in July 1999 just as
the market for domestic mini-DV camcorders was beginning
to take off. Hitherto, the top of the affordable end of
the domestic market had been occupied by the so-called
"high-band" analogue formats - Hi-8 and Super-VHS(C).
However, the camcorder manufacturers were at this time
engaged in a market repositioning exercise to put the
newer and superior digital technology at the top of the
domestic heap and price range. In practical terms, this meant cutting the prices on the older analogue products. Good news as far as it goes, but in this market repositioning exercise various useful features and facilities were removed from the analogue products so as to lower the manufacturing cost. The Sony TR913E's did not fare too badly from this point of view - features which one might expect to disappear such as stereo sound, and microphone and headphone sockets, were in fact retained. What did unfortunately disappear was the ability to manually set the white balance. This resulted in the school's TR913E's being cast aside in favour of older camcorders with manual white balance adjustment for recording an event at which the lighting scheme was both imaginative and variable. What makes the whole issue so irritating is that the manual white balance adjustment capability is believed to be contained within the operating system of many of these Sony analogue camcorders but deliberately not enabled, so as to make the higher priced digital camcorders (which would presumably offer the facility) more attractive. The matter of the non-appearance of manual white balance has been subsequently repeated by Sony at the bottom end of the digital market, in their range of so-called "Digital-8" camcorders introduced in 1999 and updated in later years. The prominent feature of these camcorders is that they would record in full mini-DV quality but on the much cheaper Video-8 and Hi-8 cassettes, and very good they were too. However, manual white balance adjustment was omitted from the specification, and in addition. the lowest priced camcorders in the range were not "input enabled". This means that although they could output their recordings through their digital terminal (usually known as "Firewire" but also called "iLink" by Sony) straight into a computer for off-line editing, they would not accept the edited material from the computer via the digital terminal for re-recording back onto tape. There is some logic in this from a camcorder manufacturer's point of view, because such an "input enabled" camcorder would be regarded as a VCR rather than a camera by the idiots in Brussels and the product would attract 10% more input duty, so the EC is the villain of this particular piece. However, many of these cheaper camcorders can have this latent but disabled input facility enabled, either by specialist companies or devices which can be bought for the purpose (see adverts and recommendations in "Camcorder User"). In the case of the Digital-8 models mentioned, as well as the digital input being enabled, the analogue sockets are also enabled for input. This means that the camcorder can be used as a digital VCR for analogue signals originating elsewhere - such as a vision mixer desk which uses a couple of cameras as sources. This is in fact how the backup master recording of the aforementioned event was made. The reader may reasonably ask how this diversion into the realms of input-enabling and bizarrre EC tax policies advances the argument concerning the disappearance of manual white balance. The answer is simple - the devices which enable the inputs on these Sony Digital-8 camcorders also enable manual white balance adjustment. The P.E. Department at the school has such an input-enabled camcorder (a TRV320E), and the manual white balance works too. This enabling of the white balance is not apparently an inevitable consequence of enabling the inputs. The specialists who "hacked" the Sony camcorder control system also found out how to enable the white balance, and they did just that for the benefit of their customers. So thanks to the specialists for their efforts, and no thanks to Sony for hiding the facility in the first place. We had long suspected that the control system in the school's TR913E Hi-8 camcorders also held the hidden capability to allow the user to set the white balance according to Sony's usual 4 step system - Auto; Hold; Daylight; and Tungsten (indoor) light. Some research on the Internet led us to Martin Smola's site, which listed control codes for the first generation Sony Digital-8 camcorders. These just happened to come onto the UK market at approximately the same time as the TR913E, so we speculated that there may be similarities in the control systems for the two ranges. We had available a suitable parallel port to LANC cable and a copy of DV-in Lite 1.7, so we tried out on the TR913Es the memory changes for manual white balance and viewfinder readout of aperture, which were listed on Mr Smola's site for 1st generation Digital-8 camcorders. Both the modifications worked! However, as we did not know how to update the checksum, we were unable to store the values to NV-RAM, consequently the camcorders forgot the new information when the batteries were removed (although they did at least work after power reconnection). If we had made the changes and stored them to the NV-RAM without updating the checksum at the same time, the camcorders would only have worked up to the point of disconnection of the power. When the power was reconnected, the start-up diagnostic routine would have detected a checksum error, and the camcorders would have gone expensively flat on their faces and not worked at all. Clearly the key to this matter is in knowing how to update the checksum. However, even if the method for updating the checksum is not known, it is still possible to make the necessary changes provided that one compensates for them by applying offsetting changes elsewhere. The objective is to ensure that the value of the finally resulting checksum is identical to the value before the changes were made. We are very grateful to Ernst Parth of Adaptive Telecom (Austria) for informing us of this technique, which is described in more detail in the following paragraphs. On early models of Sony Digital-8 camcorders, the checksum is derived by performing "XOR" operations on memory values in approximately the range 0D:14 to 0D:2E. Memory page 0D is where the control values for determining the features and facilites of the camcorder are held. Although we found that the analogue Hi-8 TR913E camcorders responded similarly to the memory changes listed for the 1st generation Digital-8s, we could not safely make the assumption that the method for updating the checksum would be identical for both the analogue and the digital product ranges, and in any case we were not familiar with the method for either product range. Herr Parth was able to advise us that as changes were made to the memory values in the control range, the camcorder kept a continuously updated record of the resulting checksum at location 02:F0. What it did not do was transfer this value to the location where it would be used for checksum evaluation on start-up. We deduced that if the dynamically revised value at 02:F0 was found to be different from the expected vaue stored elsewhere, the checksum test would fail and the camcorder would not start. The trick, then, is to make the changes necessary to enable the required features, while keeping a record of what has been done in respect of bit manipulation. Once the changes had been made, offsetting changes would be made to an (ideally unused) memory location to bring the calculated checksum (at 02:F0) back to the original value. With no difference between the checksum at 02:F0 and that held in the evaluation location, the camcorder should not be able to detect that a change had been made, and would then presumably start normally. In respect of the changes detailed below which we made to both of the school's TR913E camcorders, this did indeed turn out to be the case. We found that 0D:1B originally held the value 00, so we used this location for entering our compensating values. The following procedure was adopted:
The second camcorder was then modified, and it too initially started up in 16:9 FULL mode. However, it too performed correctly after switching off this feature. To test whether our modifications would cause the camcorder(s) to always switch on in 16:9 FULL mode after a complete power-down, we removed the main batteries and left them unpowered for more than 24 hours. On replacing the batteries, both camcorders started up in normal mode. We then removed both the main batteries and the lithium button cells, and left the camcorders unpowered for a further 24 hours. Again, on power-up, they started up in normal mode. It would therefore appear that starting up in 16:9 FULL mode after making these memory modifications is a one-time quirk of the system, and does not subsequently cause a problem. However, removing the lithium cell causes a camcorder to forget not only the date and the time, but also all the user preferences set in the menu system, resulting in having to spend time resetting everything. For this reason, lithium cell removal is not recommended as a technique for system tests, unless there appears to be no alternative. For reference, the complete memory dump of a Sony TR913E (UK PAL model) may be viewed here. Page last updated: 20 November, 2004 |
|||||||||
| This web page is optimised for viewing in
800 x 600 resolution & 16 bit colour, or better. The
use of lower resolutions is not recommended. Web page design © Video Pages (2004). |