Image conversion in BASIC code

zef · **Inscription:** 24 Oct 2018, 21:54 **Messages:** 9

Hello everyone, I have created an online tool for image conversion to the basic code 128/512, this is the link:: http://retrohead.appodesign.com/

adnz · **Inscription:** 06 Avr 2010, 01:59 **Messages:** 32

Hi,
interesting tool, good :good:

Samuel Devulder · **Posté:** 11 Jan 2019, 11:59

If you wish, you could use the conversion algorithms that I designed for GrafX2 and which are available in the Zip file of >>this message<<.

From what I can see, this tool generates a list of basic instruction to draw the picture via PSET and LINE instructions. Is it fast to display ?

I think that there might be a bug in the produced BASIC code for pictures in 320x200 16colors. Adjacent pixels of consecutive colors should alternate positive and negative values to switch between background and foreground colors, otherwise you'd get color clashes (the last plot color will bleed onto the previous ones.) Eg I mean things like:

Code:

PSET (226,191),0
PSET (227,191),3
PSET (228,191),0
LINE (229,191)-(231,191),3
PSET (232,191),9
PSET (233,191),6
PSET (234,191),9
PSET (235,191),6

shoud be coded as

Code:

PSET (226,191),-1
PSET (227,191),3
PSET (228,191),-1
LINE (229,191)-(231,191),3
PSET (232,191),-10
PSET (233,191),6
PSET (234,191),-10
PSET (235,191),6

for 320x200x16 color mode.

I also noticed an issue with the SAVEP line:

Code:

24659 DIM IM%(3000): SAVEP"www",IM%(3000)

It misses a GET(0,0)-(40,25),IM%(3000) instruction before the SAVEP instruction otherwise the saved file will not contain the picture.

Samuel Devulder · **Posté:** 11 Jan 2019, 14:17

(a bit of improvement)

With all these PSET/LINE instructions, the resulting basic file is quite huge on the boat picture of the link above. I think that packing the data in a series of <length>,<color> DATA would reduce the size a lot. Eg:

Code:

X=0: Y=0: REM start of display
READ LEN,COL
IF LEN=1 THEN PSET(X,Y),COL ELSE LINE(X,Y)-(X+LEN-1),COL: REM it is possible to only use LINE if one wishes
X=X+LEN
IF X>=320 THEN X=0:Y=Y+1
IF Y<200 THEN 1010
GOTO 1060: REM end of display

DATA ...: REM lots of data describing the image
DATA 1,-1,1,3,1,-1,3,3,1,-10,1,6,1-10,1,6
DATA ...: REM etc...

One can even imagine an even smaller encoding. Let N be a number. It can be split into C and L such that N=C+L*16. Using that we can encode both the length and the color into a single number. Notice that the color sign isn't alternating this time, but this can be retrieved easily with the "XOR" trick below

Code:

X=0: Y=0: C=0

READ N
L = N@16: REM extract length
C = (N AND 15) XOR (C>=0): REM extract (alternating sign) color
LINE(X,Y)-(X+L,Y),C
X=X+1+L: IF X>=320 THEN X=0:Y=Y+1
IF Y<200 THEN 1010

GOTO 1070

DATA ...
DATA 0,3,0: REM draw consecutive pixels of color -1,3,-1
DATA 35: REM 35 = 2*16 + 3 meaning then draw a line of 2+1=3 pixels wide with color 3
DATA 9,6,9,6: REM continue with consecutive pixels of color -10,6,-10,6

zef · **Inscription:** 24 Oct 2018, 21:54 **Messages:** 9

Hello Samuel, thanks for your directions, I had in anticipation of solving the problem of color / background in mode 0. I was trying your code, but I get a Syntax error in the line 1010 READ LEN, COL.

zef · **Inscription:** 24 Oct 2018, 21:54 **Messages:** 9

Ok, LEN is a BASIC function

Samuel Devulder · **Posté:** 11 Jan 2019, 20:19

yeah, I wrote the basic code directly on the page without checking the syntax. But the basis of the decoding algorithm is okay. BTW: better try the 2nd version with the L,C variable (and the @16 operation which is an integer division by 16): it'll be 50% more compact than the version with COLOR,LEN pair.

Notice: if you do

Code:

DEFINT A-Z

in the beginning of the code, all variables will be of type "integer" and operations will execute faster than with default variable type (32bits floats in the case of Thomson basic).

Logicielsmoto.com

Image conversion in BASIC code

Qui est en ligne