BC: When you examine their code, you will find the details on how this is
VP: Well, now you got me curious. So, how is it done?
BC: <big explanation>
VP: OK, so you don't know how it is done
I'm lost. I've analysed __loads__ of code in that sort of way over the
last 20 years - I can assure you it is a very workable strategy - they may
well be dozens of other strategies also, I gusees that one is just the way
_my_ brain works.
"you just guess from the release notes :-)"
Had to start somewhere :)
The journey moved on to tracking down all the MAS docs - Linus wrote a
codec_upload plugin; I obtained access to MAS tech support and some demo
codecs with the assurance that it was okay to use my access for the benefit
of Open-Source stuff; Someone else found some lecture slides; another
produced an instruction set; etc. etc.
Sadly, in the end, I could not fulfil the requirement of the signature of
the CEO of Micronas, so as far as Rockbox was concerned, the project was
Recently (check this archive), one of the Brothers Stenberg (don't recall
which - _think_ it was Bjorn) said that they would now accept MAS
codecs. So if you've got what it takes (time/dedication/experience/etc)
the information is there to play with - pitch shift and pcm support are
only the start of what you can do with the MAS core.
Perhaps that is more in line with the answer you were hoping for? Ie. Not
how to hack a lump of code from one firmware to another, but how to program
At 17:23 22/02/2005, you wrote:
>>Possibly the biggest clue is this snippet from the m3po version history
>>[numbered by me for clarity]:
>>2.10 >> V2.11
>>1. party lock - eject cds while lock complete
>>2. pitch and speed of playback was wrong
>>3. minor shuffle bug fixed
>>4. more files on one CD supported
>>When I read that I get:
>>3. this list includes all changes to firmware
>>4. a buffer/malloc/bss got bigger somewhere
>>2. the pitch/speed code was tweaked
>>1. & 3. two other bits of code also changed
>OK, so you don't know how it is done, you just guess from the release
>>Ever play "spot the difference" on a pub quiz machine? ...One of the
>>differences directly effects the pitch/speed change ;)
>Look at the MAS datasheet, there is an internal register that is used to tell
>the MAS what frequency value the external crystal has, if you set that one
>wrong you get a higher/lower pitch and speed. My guess is that this value
>was slightly off and was corrected. Unfortunately the value range of this
>does not allow for the large range of the speed change that the M3PO features.
>>My next hit is a mind-numbingly simple moment of genuis called FTEXT, an
>>ancient DOS util which isolates ascii strings in unencrypted binary files
>>- you would not be surprised at how far you can get by simply tracing
>>things from the point they print "speed" on the screen.
>it is called "strings" if you use Linux/Unix, maybe ancient but still included
>in every distro.
>>Grab yourself a good hex editor [and please tell me what it is - LOL - I
>>have NEVER found a _good_ hex editor (or calculator) for windows] and
>>"hit the bits" as they say.
>Editor: WinHex, for the calculator I prefer real HW.
Received on Tue Feb 22 19:25:01 2005