Conversion Reversal

[nickpike 0]

Does anybody know how to reverse the following formula. Basically, I want to undo this conversion. It's in this C++ speak that I don't understand. If anyone can do this, I'll probably then ask for the XML equivalent reversal. If I get this to work, I'll then share the info. NOTE, I have added a space after the 8 to stop a smiley from showing Conversion: Y = 100 * (1000000 + ((x >> 20) & 0xF)*100000) + ((x >> 16) & 0xF)*10000) + ((x >> 12) & 0xF)*1000) + ((x >> 8 ) & 0xF)*100) + ((x >> 4) & 0xF)*10) + ((x) & 0xF))) cheers, nick

[Guest VulcanB2]

Hi,You can't reverse it as there is an AND operator - the result would be ambiguous.e.g.:11reversed could be:00 && 11-- or --10 && 01That's the problem. :( You could start branching into probability of which would be the most likely answer, but I think that is way beyond the scope of FS.Best regards,Vulcan.

[mgh 89]

I replied to this in another forum:It isn't syntatically correct - the numbers of opening and closing brackets aren't equal.Ignoring that and assuming x is unsigned long or a DWORD, it appears to extract packed BCD(?) digits from the lower nibbles (4 bits) of x and build a positive integer in the range 100 - 200 million and assign it to Y. Although it is possible reverse the expression to calculate the packed BCD value of (Y/100) - 1000000 and put it in the lower 20 bits of x, it isn't possible to determine what the remaining upper bits of x should be set to from the information available.The ANDing has nothing do with ambiguity, given that 0xF sets all 4 bits (1111). ((x >> 20) & 0xF), for example, shifts right by 20 bits then extracts the lower 4 bits unambiguously.Anyway, the ANDing is bitwise (&) not logical (&&) and the examples you give both AND to 00 not 11.00 & 11 = 0010 & 01 = 00

[nickpike 0]

Thanks guys, I'll knock it on the head,cheers,nick

[Guest JeanLuc_]

it seems to be BCD convertion. Given the numeric ranges, it looks like frequency packed in BCD converted to its numeral form.For example: say frequency is 136.50 I would assume from your code it would be BCD packed in the following way (shown in hexadecimal)0x00365000This is a typical 16.16 fix point format (the 2 MSB are integer frequency and the 2 LSB are fractional frequency)The code then extract each individual BCD digit (1 then 3 then 6 etc...)Y = 100 * (10000000 + (3 * 100000) + (6 * 10000) + (5 * 1000) + (0 * 100) + (0 * 10) + 0)this would give: 100 * 1365000 = 136500000It is often used the way around:(1.0 / 1000.0) * 1365000 = 136.5Hope this helps!

[nickpike 0]

Many thanks. Your logic is spot-on. It takes a frequency and then applies the conversion (but shouldn't have done). I think it converts the frequecy to a BCD number.If I understand this properly, the proper frequency digits are recognisable in the result.I know that, say, a frequency of 113.90 becomes 230604000.Can you shed any more light on this. If your thinking is correct, how would you convert the number back to a frequency?Cheers and thanks for the help,nick

[phjvh 0]

Hi,Would be interesting!May be finally we could use: (@c:FlightPlanWaypointFrequency...?Jan"Beatus ille qui procul negotiis..."

[nickpike 0]

Yes correct. BTW, this variable is not used in the GPS and so is not supported by MSFS.cheers,nick