Hey, I'm nc513 and here's a sneak peek
at a fascinating work in progress by SLC, who is known as, for example, the wizard
behind TapEx, which is a brilliant piece of software that you can use for analyzing
and tidying up .TAP tape image files, etc. Right. So our test file will be the game
Arkanoid, cracked and exomized by StatMat. We've already prepared a .TAP tape image
file using a program SLC has written in C#. So we'll start by writing this tape image
to a good old cassette tape by using the 1541 Ultimate 2+ cartridge with its tape adapter. There's nothing special about this C64. It
just happens to have a few screws loose, just like Z... Uh, nevermind. What I meant
to say is that inside my breadbin case is a completely unmodified 250469 model motherboard. And now we'll just wait for maybe 10 more seconds. It's done. So let's rewind to where we
started and then I'll remove all extra stuff from the computer, so that we're
left with a more basic traditional setup. When I'm done rearranging, there will
be nothing else connected to the C64, except the datassette and two joysticks,
and the power supply, of course. I guess it could be worthwhile to mention
that while recording a tape using extreme bitrates like this, you of course need to use a
device with accurate timings for streaming the pulse data to the datassette, and the 1541
Ultimate 2+ seems to do a very good job. All right, I'll hit Shift
and Run/Stop and off we go! The file size is about 27.3 kilobytes and when
the screen starts flashing, it means the C64 has started turbo loading those 28005 bytes into
RAM. So let's see how long that phase will take, or if the game will load at all. At the point where the C64 hopefully will
show a black screen for about a second, we know that the loading process has
finished and the datassette has stopped. It's executing! It's
decompressing the exomized data. And... Well, that's... that's an amazing new
world record, almost guaranteed. This is the highest bitrate ever achieved when recording
and reading back heavily compressed data with a Commodore datassette. As you can see, I'm jumping
up and down with excitement. This is so amazing. SLC's estimation of the bitrate
achieved here is 13415 bits per second which is, in my opinion, actually a quite
modest claim, as I believe he chose not to deduct the turbo synchronization phase from the
total turbo duration. Note that I say modest even though it's over three and a half times
faster than Turbo-250, ABC Turbo or Turbo III. You could in fact argue that the actual bitrate
was even higher, if we choose to exclude the synchronization phase and then use, for example,
Luigi's excellent DC2N5-LC tape dumping device for measuring the total duration of the turbo
part exactly corresponding to the 28005 bytes we needed to load. After we set this new
world record, I dumped this very same recording to a .DMP file that I then converted to .TAP
and then I just made the following calculations. As you can see, I arrived at the number
13727 bits per second and regardless of which calculation method you prefer, this
is most probably a new world record, by far. As a final note, I could mention that
this same experiment failed miserably on my black 1531 datassette. I'm not quite
sure why, but it might have to do with the circuit design and the components Commodore used
for that model. SLC and I have been testing a bunch of different strategies during the last
couple of weeks, and for this experiment we made sure the .TAP file and the recording datassette
were a really good match. That's all. Bye!
