Dear rns team,
when select effect column with pattern length 64 lines (default), starting with value 0900
and ending with 09ff, then ctrl+I (or interpolate linear) causes at line 20 and above invalid (shifted, non linear) interpolation.
when interpolate for same pattern lenght from 00 to 0x80, everything seems to be ok.
dont know whether it is WAD or error
Adam
I can’t duplicate, what did you do before that? something in the advanced edit?
there are different shortcuts that do lineary interpolation (ctrl-i -> selection, shift-i -> track, ctrl+shift -i -> column)
If you want a perfectly scaled linear ramp between 00 and 100h in 64 lines, you have to interpolate between 00 and FC.
This looks a bit strange, but remember that the max is not FF but 100 and you do set the interpolation end at the last line of the pattern, not the first line of the next pattern. For a pattern of 128 use FE, for a pattern of 32 use F8 and so on…
hm, hm,…strange that U cant replicate my prob…
I use rns 2.1.9, workflow is:
open rns (appears default 64 lines pattern)
put 0900 into first row in first track effect col
put 09ff into last row in first track effect col
select whole column
then right mouse, column interpolate linear.
all should be 00,04,08,0c,00,04,08,0c…
but since row 20 it shows 50,55,59,5d…
Eduard,
can you please explain more detailed what do you mean ?
By reading help, found this:
http://tutorials.renoise.com/Renoise/EffectCommands
…
09xx - Trigger sample offset, 00 is sample start, ff is sample end . Offset positions are related to the total sample size in Renoise mode. The larger the sample, the more inaccurate stepping will be (Using Amiga/FT2 mode, ff means the 255th block of 256 bytes). (Sample instrument related only)
…
anyway guys, this is not serious problem,…I was just a bit confused when playing around with samples why my sample sound that bad and with shifts, then I found why… I repaired column manually and everything is ok.
A.
FF is not the samples end, but the last 1/256 slice of the sample. I don’t know how I could explain this to you other than this, so simply use this as guideline please:
If you want to interpolate the 09 effect linearly for a whole pattern, then use:
00 to FF for a 256 lines pattern
00 to FE for a 128 lines pattern
00 to FC for a 64 lines pattern
00 to F8 for a 32 lines pattern
it’s not smooth because the amount of lines you’re interpolating is not equally divisible by your value each new command is increasing by.
you’ll need to have 09FC on the last line to get a smooth interpolation because you’re interpolating 64 lines.
the maths is line count (minus 1) multiplied by offset jump value
in your case http://www.google.com/search?q=63+*+4+in+hex
the more you track it’ll come naturally.
guys,
shame on me 
tactic was right !!!
here is summary:
my sample is always same lenght as pattern, in this case 64 row lenght in “original” speed as it was recorded by record command with sync start&stop “pattern” option.
sample is dividable to 256 slices.
255 = 0xFF is NOT end of sample but start of last slice.
now 256 is is dividable into 64 rows by 4, so I have virtually 4 slices per 1 row.
last row in pattern must have !!!90FC!!! as tactic suggested
(4 slice duration play = 1 row play duration, in described case = 09FC is last 4-th slice)
What was confusing is that, rns breaks the linear interpolation at line 20
as I am not that experienced and tried with samples I thought that it is rns error BUT it was my fault.
when I changed end to 09FC and repeat interpolation is fine.
sorry for this annoyance and thanks for your help
adam