Well, I got it to a usable state so I figured I should release it.
Please post any comments and criticism you have, I want to hear it!
Also, post any cool images you come up with.
Download:
[url]http://www.mediafire.com/?mzjbe0t03hm[/url]
(Poodle's request): I wrote it in C#, using VS2010.
I can release source code if you like, I just haven't gotten around to uploading it all yet.
Edit: Source is uploaded
[url]http://www.mediafire.com/?ztyyhzynjow[/url]
Edit: My girlfriend had a go, came up with this (click for full-size)
[img_thumb]http://www.cubeupload.com/files/4d1c00untitled.png[/img_thumb]
I thought it was quite cute :v:
What language? Any information?
Of course, sorry.
Updated the OP.
If you type a scale that isn't an integer it complains about it not being between 0.1 and 4.0.
Maybe it's just me tho... :v:
Really good, Here's one I really like:
[img_thumb]http://www.cubeupload.com/files/7bce00blu.png[/img_thumb]
Edit: This one too:
[img_thumb]http://www.cubeupload.com/files/259000guitar.png[/img_thumb]
Pretty sweet, good work!
Since you didn't bother with smoothing, I implemented it.
My Program.cs:
[cpp]*using System;
using System.IO;
using System.Windows.Forms;
using System.Drawing;
namespace ASCII_Generator
{
static class Program
{
/// <summary>
/// The main entry point for the application.
/// </summary>
[STAThread]
static void Main()
{
Setup mySetup = new Setup();
if(mySetup.ShowDialog() != DialogResult.OK)
return;
string shades = mySetup.CharSet;
float ixscale = 1 / mySetup.XScale, iyscale = 1 / mySetup.YScale;
OpenFileDialog ofd = new OpenFileDialog();
ofd.Title = "Select input file";
ofd.Filter = "Supported image files|*.bmp;*.gif;*.jpg;*.jpeg;*.png;*.ico";
if (ofd.ShowDialog() == DialogResult.OK)
{
// Choose a save location and setup the file stream
SaveFileDialog sfd = new SaveFileDialog();
sfd.Filter = "Text files|*.txt|All files|*.*";
if(sfd.ShowDialog() != DialogResult.OK)
return;
FileStream fs = new FileStream(sfd.FileName, FileMode.Create);
StreamWriter sw = new StreamWriter(fs);
Bitmap image = new Bitmap(ofd.FileName);
// Iterate over each pixel
for (float fy = 0; fy < (float)image.Height; fy += iyscale)
{
int y = Convert.ToInt32(fy), ycount;
float yfract, iyfract;
if(y >= image.Height)
{
y = image.Height-1;
ycount = (int)Math.Floor(fy) - y;
yfract = 0;
iyfract = 1;
}
else
{
ycount = (int)Math.Ceiling(iyscale);
yfract = (float)Math.Ceiling(fy) - fy;
iyfract = 1 - yfract;
}
for (float fx = 0; fx < (float)image.Width; fx += ixscale)
{
int x = Convert.ToInt32(fx), xcount;
float ul, ur, ll, lr; //upper-left, upper-right, lower-left, lower-right
if(x >= image.Width)
{
x = image.Width-1;
xcount = (int)Math.Floor(fx) - x;
ul = iyfract;
ur = yfract;
ll = 0;
lr = 0;
}
else
{
xcount = (int)Math.Ceiling(ixscale);
float xfract = (float)Math.Ceiling(fx) - fx, ixfract = 1 - xfract;
ul = ixfract * iyfract;
ur = ixfract * yfract;
ll = xfract * iyfract;
lr = xfract * yfract;
}
float avShade = 0;
for(int smoothy = y; smoothy < Math.Min(fy + iyscale, image.Height - 1); ++smoothy)
for(int smoothx = x; smoothx < Math.Min(fx + ixscale, image.Width - 1); ++smoothx)
{
avShade += image.GetPixel(smoothx, smoothy).GetBrightness() * ul;
avShade += image.GetPixel(smoothx + 1, smoothy).GetBrightness() * ur;
avShade += image.GetPixel(smoothx, smoothy + 1).GetBrightness() * ll;
avShade += image.GetPixel(smoothx + 1, smoothy + 1).GetBrightness() * lr;
}
int count = ((ul == 0) ? 0 : 1) + ((ur == 0) ? 0 : 1) + ((ll == 0) ? 0 : 1) + ((lr == 0) ? 0 : 1); //how many did actually contribute?
avShade /= count * xcount * ycount;
avShade = shades.Length - avShade * (shades.Length - 1) - 1;
sw.Write(shades[Convert.ToInt32(avShade)]);
}
sw.WriteLine();
}
sw.Flush();
sw.Close();
}
}
}
}[/cpp]
I also replaced your spaces with tabs, removed the temporary string and instead write directly to the file as well as removing the temporary string for the filename.
I read on this forums that it's faster to access pixel data with unsafe code, so you could look into that to make it faster.
I probably should but a bunch of comments on my algorithm, I don't think it's too easy to follow without them. I might add some after I slept; if there are any questions about it I'd be pleased to answer :)
The difference of befores version and my smoothing version:
no smoothing/before:
[img_thumb]http://anyhub.net/file/1before.png[/img_thumb]
smoothing/after:
[img_thumb]http://anyhub.net/file/1after.png[/img_thumb]
reference picture:
[img_thumb]http://a778.ac-images.myspacecdn.com/images01/78/l_b0370fea0387d4937c33a183e8880869.png[/img_thumb]
[QUOTE=ZeekyHBomb;22033590]Since you didn't bother with smoothing, I implemented it.
My Program.cs:
*code*
I also replaced your spaces with tabs, removed the temporary string and instead write directly to the file as well as removing the temporary string for the filename.
I read on this forums that it's faster to access pixel data with unsafe code, so you could look into that to make it faster.
I probably should but a bunch of comments on my algorithm, I don't think it's too easy to follow without them. I might add some after I slept; if there are any questions about it I'd be pleased to answer :)
The difference of befores version and my smoothing version:
no smoothing/before:
[img_thumb]http://anyhub.net/file/1before.png[/img_thumb]
smoothing/after:
[img_thumb]http://anyhub.net/file/1after.png[/img_thumb]
reference picture:
[img_thumb]http://a778.ac-images.myspacecdn.com/images01/78/l_b0370fea0387d4937c33a183e8880869.png[/img_thumb][/QUOTE]
Images aren't loading for me.
Also, I know unsafe code is faster for accessing pixels, but this thing is so fast anyway, I didn't think it was worth it :P
Thanks for the replies!
[QUOTE=Chris220;22038448]Images aren't loading for me.[/QUOTE]
Works for me.
XD Thanks for updating the post, this is really cool XD
[QUOTE=turb_;22038664]Works for me.[/QUOTE]
Yeah I see them now.
[editline]12:11PM[/editline]
[QUOTE=Dj-J3;22030386]If you type a scale that isn't an integer it complains about it not being between 0.1 and 4.0.
Maybe it's just me tho... :v:[/QUOTE]
That's odd... Works fine for me :O
What OS are you on? And which .NET framework?
[QUOTE=Chris220;22039912]Yeah I see them now.
[editline]12:11PM[/editline]
That's odd... Works fine for me :O
What OS are you on? And which .NET framework?[/QUOTE]
Windows 7, .NET Framework 3
[editline]02:18PM[/editline]
[QUOTE=Chris220;22039912]That's odd... Works fine for me :O
What OS are you on? And which .NET framework?[/QUOTE]
Oh wait, it works if i type "," and not ".".
:v:
[QUOTE=Dj-J3;22040571]Oh wait, it works if i type "," and not ".".
:v:[/QUOTE]
Ah, that's a Windows setting, whether you use , or . for decimal point notation. I'm not sure how you change it :P
I can use both in Linux :buddy:
Though if I use the , then it will just be parsed as 5 o.O
I guess I broke something when I installed LXDE upon Openbox.
[editline]04:38PM[/editline]
Alright, so here's a fixed and improved version of my smoothing asciiart generator.
[url]http://anyhub.net/file/asciiart.zip[/url]
Includes binary compiled with Mono 2.4.4 (should run on .NET 3(.5?) or higher as well) and source-code.
Additions apart from the smoothing are individual x and y scaling, a bigger default charset and radio-buttons to tell the program if the charset is ordered from dark to bright or the other way.
[editline]04:38PM[/editline]
Aww, no bump :(
[editline]04:42PM[/editline]
I also wanted to profile my code, but MonoDevelop doesn't include a profiler ._.
SharpDevelop <3 Too bad there's no official Linux port.
GetPixel vs unsafe code:
unsafe version:
[code]z33ky@mobileblahbuntu:~/Projects/smoothing_asciiart/smoothing_asciiart/bin/Release$ time mono ./smoothing_asciiart.exe
real 0m0.671s
user 0m0.540s
sys 0m0.090s
z33ky@mobileblahbuntu:~/Projects/smoothing_asciiart/smoothing_asciiart/bin/Release$ time mono ./smoothing_asciiart.exe
real 0m0.647s
user 0m0.600s
sys 0m0.030s
z33ky@mobileblahbuntu:~/Projects/smoothing_asciiart/smoothing_asciiart/bin/Release$ time mono ./smoothing_asciiart.exe
real 0m0.667s
user 0m0.590s
sys 0m0.050s
z33ky@mobileblahbuntu:~/Projects/smoothing_asciiart/smoothing_asciiart/bin/Release$ time mono ./smoothing_asciiart.exe
real 0m0.650s
user 0m0.580s
sys 0m0.050s
z33ky@mobileblahbuntu:~/Projects/smoothing_asciiart/smoothing_asciiart/bin/Release$ [/code]
GetPixel version:
[code]z33ky@mobileblahbuntu:~/Projects/smoothing_asciiart/smoothing_asciiart/bin/Release$ time mono ./smoothing_asciiart.exe
real 0m0.654s
user 0m0.580s
sys 0m0.070s
z33ky@mobileblahbuntu:~/Projects/smoothing_asciiart/smoothing_asciiart/bin/Release$ time mono ./smoothing_asciiart.exe
real 0m0.664s
user 0m0.530s
sys 0m0.110s
z33ky@mobileblahbuntu:~/Projects/smoothing_asciiart/smoothing_asciiart/bin/Release$ time mono ./smoothing_asciiart.exe
real 0m0.697s
user 0m0.660s
sys 0m0.020s
z33ky@mobileblahbuntu:~/Projects/smoothing_asciiart/smoothing_asciiart/bin/Release$ time mono ./smoothing_asciiart.exe
real 0m0.686s
user 0m0.590s
sys 0m0.070s
z33ky@mobileblahbuntu:~/Projects/smoothing_asciiart/smoothing_asciiart/bin/Release$ [/code]
Conclusion: Almost the same; difference negligible.
Unsafe code might just make a difference when writing to the pixels.
And a link to a version with more accurate brightness calculation and minor improvements (detecting if the file could not be opened, yay): [url]http://anyhub.net/file/2asciiart.zip[/url]
Again includes binary and source.
Edit: Fixed rounding-issue that would cause an IndexOutOfRangeException. Link updated.
Actually quite cool.
---------------------------
---------------------------
Error: Scale must be a numerical value between 0.1 and 4.0
---------------------------
OK
---------------------------
Sigh, its a valid nunber. runnig Vista without UAC
[editline]08:42PM[/editline]
Figured it out, had to use 1,0 instead of the default settting: 1.0
It just picks a symbol for the greyscale of a pixel, yes?
It should actually find edges and shit. That'd be awesome.
Why would you make it do that though?
It'll look like hand-created ASCII-art (Better)?
Ahhh right yeah I get what you mean now. Also that mean you could change settings to thicken lines etc.
Wow, this is really cool. Thanks.
[QUOTE=Tobba;22045990]---------------------------
---------------------------
Error: Scale must be a numerical value between 0.1 and 4.0
---------------------------
OK
---------------------------
Sigh, its a valid nunber. runnig Vista without UAC
[editline]08:42PM[/editline]
Figured it out, had to use 1,0 instead of the default settting: 1.0[/QUOTE]
Hmm yeah, someone else had that issue. I wish I could fix it on my end :/
Sorry :)
[editline]09:49PM[/editline]
[QUOTE=Mindtwistah;22047694]Wow, this is really cool. Thanks.[/QUOTE]
Thank ye :)
And, you're welcome
Maybe you can somehow detect which one the system needs and then just replace the , or . with the appropriate one.
Just replace commas with decimal points.
Well I for example need points. Dj-J3 and Tobba on the other hand need commas for it to work.
[editline]11:34PM[/editline]
Actually a point [i]should[/i] be the one working fine: [url]http://msdn.microsoft.com/en-us/library/system.globalization.numberformatinfo.numberdecimalseparator.aspx[/url]
I guess you could try to explicitly tell it the format and see if that works.
Edit: For future use: It's probably be better to use NumberFormatInfo.InvariantInfo instead of cloning NumberFormatInfo.CurrentInfo. It's not in the code of the zip-archive from down there, that small change is not work re-uploading it ;)
[editline]11:41PM[/editline]
Works.
[url]http://anyhub.net/file/3asciiart.zip[/url]
Fixed accidentally (and partially) broken smoothing and throwing of an ArgumentException when the scale cannot add up to the image resolution (e.g. xscale is 2 and the image has a width of 3; 1 * 2 < 3 > 2 * 2).
[url]http://anyhub.net/file/5asciiart.zip[/url]
Sorry, you need to Log In to post a reply to this thread.
