OK, cool.
And I find no frustrum clipping sort of annoying, don't know about others though.
And I find no frustrum clipping sort of annoying, don't know about others though.
flyingfisch wrote:
OK, cool.
And I find no frustrum clipping sort of annoying, don't know about others though.
I implemented a lame sort of frustrum culling/clipping; it just culls every line with an endpoint behind the eye plane rather than calculating the actual frustrum volume and clipping at the edges of the frustrum. Works well enough, including for your 1.5^X example. I also implemented the [EXIT] key from all the menus as requested. And I find no frustrum clipping sort of annoying, don't know about others though.
KermMartian wrote:
flyingfisch wrote:
OK, cool.
And I find no frustrum clipping sort of annoying, don't know about others though.
I implemented a lame sort of frustrum culling/clipping; it just culls every line with an endpoint behind the eye plane rather than calculating the actual frustrum volume and clipping at the edges of the frustrum. Works well enough, including for your 1.5^X example. I also implemented the [EXIT] key from all the menus as requested.And I find no frustrum clipping sort of annoying, don't know about others though.
Nice
So it looks about ready for release?
Yup, looks just about ready. I haven't seen any solid suggestions or comments on Omnimaga, I'll wait to see what Ashbad has to say about it, and pending another other comments while I'm packing it up and preparing a video, 1.0 is ready.
I'm sorry (and glad) to say that while I had suggestions and problems, they have all been mentioned and fixed already; awesome work Kerm. I won't be using it for math classes this year, but I'm sure many others will be!
Ashbad wrote:
I'm sorry (and glad) to say that while I had suggestions and problems, they have all been mentioned and fixed already; awesome work Kerm. I won't be using it for math classes this year, but I'm sure many others will be!
Many thanks! That's about all that I needed to hear, so I'm going to start packing this up for its 1.0 release now.
Hey, I found one more little problem! 
In trace mode, the bounding box is drawn over the trace numbers. It should be drawn below.
Great job!
In trace mode, the bounding box is drawn over the trace numbers. It should be drawn below.
Great job!
RC2 is great. The bugs seem to be gone, and it's easier to use
One quick last-minute suggestion: having a sample equation in the Z= editor on first use. DS's Graph3 uses Z=(X3Y–Y3X)/390.
One quick last-minute suggestion: having a sample equation in the Z= editor on first use. DS's Graph3 uses Z=(X3Y–Y3X)/390.
Thanks for both of those last-minute suggestions, both of which I have implemented. The video is published, the file is about to head into the archives, and the news article will be a short write!
That's strange... My graph screen is now pale red/pale green instead of black/white. I assumed it was a new feature to lower the contrast until I saw the video D:
Deep Thought wrote:
That's strange... My graph screen is now pale red/pale green instead of black/white. I assumed it was a new feature to lower the contrast until I saw the video D:
Err? Please try with the latest version and let me know what happens. The issue might be that you were using the old state save file; try deleting @MainMem/@Graph3DP/G3DP. Deep Thought wrote:
Yeah, that was it. Thanks!
Neat-o. This happened when you switched from Release Candidate 1 to 2, correct? If so, I can rest easy; otherwise, I'll worry that this is going to pop up again? 
I have a suggestion to make. With generic 3d code it is easy to render stereo pair of a 3d plot. On a computer it does not make sense since screen is far away, but with a calculator it might work.
Anyway, I made several stereo images (on a PC) and sized them to 384x216 to view on Prizm with imgviewer to give a preview what it would feel like.
Anyway, I made several stereo images (on a PC) and sized them to 384x216 to view on Prizm with imgviewer to give a preview what it would feel like.
And here is some quick and dirty lua code to rotate a cube in stereo.
Just as an illustration of a concept.
arrows=rotate cube
+/- = farther/closer (affects perspective)
*// = larger/smaller (does not affect perspective)
sin/cos = chnage eye base
exe = switch left and right (for cross-eye vs parallel eye)
I think it has its potential. For tricky 3d surfaces or illustrations of stereometric theorems.
Code:
Just as an illustration of a concept.
arrows=rotate cube
+/- = farther/closer (affects perspective)
*// = larger/smaller (does not affect perspective)
sin/cos = chnage eye base
exe = switch left and right (for cross-eye vs parallel eye)
I think it has its potential. For tricky 3d surfaces or illustrations of stereometric theorems.
Code:
-- Rotate 3d object in stereo
-- (C) nsg, 2013
function mat_mul(m1,m2)
-- resulting matrix has as many rows as m1 and as many columns as m2
local res={}
for r=1,#m1 do
local row={}
assert(#m1[r]==#m2)
-- print("mul r=",r)
for c=1,#m2[1] do
local s=0
for k=1,#m2 do
s=s+m1[r][k]*m2[k][c]
end
table.insert(row,s)
end
table.insert(res,row)
end
return res
end
function mat_transpose(mat)
local res={}
for i=1,#mat[1] do
table.insert(res,{})
end
for r=1,#mat do
for c=1,#mat[r] do
table.insert(res[c],mat[r][c])
end
end
return res
end
function mat_make_rotation(axis,angle)
local mat={}
for r=1,3 do
table.insert(mat,{})
for c=1,3 do
local v=0
if r==axis and c== axis then v=1
else v=0
end
table.insert(mat[r],v)
end
end
local c=math.cos(angle)
local s=math.sin(angle)
mat[(axis+0)%3+1][(axis+0)%3+1]=c
mat[(axis+0)%3+1][(axis+1)%3+1]=s
mat[(axis+1)%3+1][(axis+0)%3+1]=-s
mat[(axis+1)%3+1][(axis+1)%3+1]=c
return mat
end
function make_cube(side)
if nil==side then
side=1
end
local function coord(i,mask)
return (0==bit32.band(i,mask) and side or -side)
end
local vertex={}
local edge={}
for i=0,7 do
table.insert(vertex,{coord(i,1),coord(i,2),coord(i,4)})
local bit=1
for b=1,3 do
local next=bit32.bxor(i,bit)
if next<i then table.insert(edge,{i+1,next+1}) end
bit=bit*2
end
end
return vertex,edge
end
function mat_dump(m,name)
if nil==name then name="?" end
print(string.format("matrix %s with %d rows",name,#m))
for r=1,#m do
io.write(string.format("%3d:",r))
for c=1,#m[r] do
io.write(string.format(" %3d",m[r][c]))
end
io.write("\n")
end
end
function dumpargs(...)
print(...)
end
function nothing() end
if nil==zmg then
zmg={
makeColor=function() return 999 end,
drawLine=dumpargs,
keyMenu=function()
io.write("key=");
local k=io.read();
return tonumber(k)
end,
drawRectFill=dumpargs,
fastCopy=dumpargs,
}
else
mat_dump=nothing
end
m,e=make_cube(10)
mat_dump(m,"vertexes")
mat_dump(e,"edges")
SCREENX,SCREENY=384,216
BLACK=zmg.makeColor(0,0,0)
xrot=5.585
yrot=0.436
SCL=265
AWAY=57
OX=SCREENX/2
OY=SCREENY/2
function screenfrom3d(v)
local z=SCL/(v[3]+AWAY)
return OX+v[1]*z, OY+v[2]*z
end
function V_drawbody(v,e)
for i=1,#e do
local x1,y1=screenfrom3d(v[e[i][1]])
local x2,y2=screenfrom3d(v[e[i][2]])
zmg.drawLine(x1,y1,x2,y2,clr)
end
end
function V_display()
local rot=mat_transpose(mat_mul(mat_make_rotation(1,xrot),mat_make_rotation(2,yrot)))
mat_dump(rot,"rot")
mat_dump(m,"m")
local v=mat_mul(m,rot)
mat_dump(v,"translated vertexes")
clr=zmg.makeColor(255,255,255)
zmg.drawRectFill(0,0,SCREENX,SCREENY,BLACK)
local vt=mat_mul(v,EYE1)
OX=(flip and 3 or 1)*SCREENX/4
V_drawbody(vt,e)
vt=mat_mul(v,EYE2)
OX=(flip and 1 or 3)*SCREENX/4
V_drawbody(vt,e)
zmg.fastCopy()
end
PI2=math.pi*2
dx=5*PI2/360
dy=dx
flip=false
function makeeyes(eyebase)
EYEBASE=eyebase
EYE1=mat_make_rotation(2,eyebase)
EYE2=mat_make_rotation(2,-eyebase)
end
makeeyes(4*PI2/360)
K_EXIT= 47
K_LEFT= 38
K_DOWN= 37
K_UP= 28
K_RIGHT= 27
K_SIN= 46
K_COS= 36
K_DEL= 44 -- [DEL]
K_MUL= 43 -- [*]
K_PLUS= 42 -- [+]
K_NEG= 41 -- [(-)]
K_DIV= 33 -- [/]
K_MINUS= 32 -- [-]
K_EXE= 31 -- [EXE]
function playstereo()
quit=false
repeat
V_display()
local k=zmg.keyMenu()
if k==K_EXIT then quit=true
elseif k==K_LEFT then yrot=yrot+dy
elseif k==K_RIGHT then yrot=yrot-dy
elseif k==K_UP then xrot=xrot-dx
elseif k==K_DOWN then xrot=xrot+dx
elseif k==K_PLUS then AWAY=AWAY-1
elseif k==K_MINUS then AWAY=AWAY+1
elseif k==K_MUL then SCL=SCL*1.1
elseif k==K_DIV then SCL=SCL/1.1
elseif k==K_SIN then makeeyes(EYEBASE-0.001)
elseif k==K_COS then makeeyes(EYEBASE+0.001)
elseif k==K_EXE then flip=not flip
end
xrot=xrot % PI2
yrot=yrot % PI2
until quit
print("yrot=",yrot)
print("xrot=",xrot)
print("AWAY=",AWAY)
print("SCL=",SCL)
print("EYEBASE=",EYEBASE)
print("flip=",flip)
end
A=playstereo -- to easy typing in prizm
playstereo()
nsg wrote:
I have a suggestion to make. With generic 3d code it is easy to render stereo pair of a 3d plot. On a computer it does not make sense since screen is far away, but with a calculator it might work.
Anyway, I made several stereo images (on a PC) and sized them to 384x216 to view on Prizm with imgviewer to give a preview what it would feel like.
Anyway, I made several stereo images (on a PC) and sized them to 384x216 to view on Prizm with imgviewer to give a preview what it would feel like.
If I ever have a bunch of money and time to waste, I would get a 3D monitor and glasses kit and make gCAS3 have support for 3D stereo graphing. Also, making these image to make me go cross-eye was a nice touch
ok... I have found a bug now that i am actually using this for math.
First, when tracing Z1=X, when X=1, Z=-.999..., when X=2, Z=-1.999...
This is obviously incorrect.
Also, maybe it is too late, but what about having it calculate the points on the graph once, store it to an array, and then display the 3d array the same way you would with any 3d model? Wouldn't this allow faster spinning?
First, when tracing Z1=X, when X=1, Z=-.999..., when X=2, Z=-1.999...
This is obviously incorrect.
Also, maybe it is too late, but what about having it calculate the points on the graph once, store it to an array, and then display the 3d array the same way you would with any 3d model? Wouldn't this allow faster spinning?
flyingfisch wrote:
ok... I have found a bug now that i am actually using this for math.
First, when tracing Z1=X, when X=1, Z=-.999..., when X=2, Z=-1.999...
This is obviously incorrect.
That's an artifact of the floating-point format. I (and by "I" I mean "vprintf") should probably be rounding instead of truncating.
First, when tracing Z1=X, when X=1, Z=-.999..., when X=2, Z=-1.999...
This is obviously incorrect.
Quote:
Also, maybe it is too late, but what about having it calculate the points on the graph once, store it to an array, and then display the 3d array the same way you would with any 3d model? Wouldn't this allow faster spinning?
That's what it already does. The computations it needs to do during rotation are the transformations for scale and rotation. Register to Join the Conversation
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
Advertisement