but i was wondering if there was a way to find smaller size versions or ones that can be more easily converted
Also, there are a few svg wizards who are able to create advance svgs under 15kb, which takes a lot knowledge about the structure of svgs.
it took a second, but your points explanation worked....
I put the precision number down and it got to under 15kb and uploaded. Now i have to experiment and see what i can get away with then
I feel compelled to interject here just to make sure
@brutus13 is fully understanding things - and I've ended up typing out a lengthy post which I suggest anyone who finds themselves frequenting this thread reads, to help them understand a little of how to do it for themselves, and why sometimes you just get nowhere with requests.
** and yes some of it is an over simplification, but it's the principles I'm trying to convey...
There's two ways of creating an SVG from an image you already have.
1) Auto trace
2) Re-draw it yourself
Firstly
Auto trace:
An auto trace is when a program looks at a
raster image (i.e. JPG/PNG etc.) and tries to automatically re create it as
a vector image (SVG)
- A vector image assigns co-ordinates for lines that make up the image, and fills and styles to apply to those lines.
- A raster image is simply information about what colour each pixel in the image should be.
They are two fundamentally different things and you need to remember that going from JPG to SVG is not like going from JPG to PNG (for example).
A
raster file is generally compressed itself, but in very simple terms you can imagine that it doesn't really matter what the pixels are, the file has a value for each of them anyway, and the more there are the bigger the file size will be.
A 10×10 image will have 100 pixels (and might be 1kb (for example), and a 1000×1000 image will be a million pixels (i.e. 1 MegaPixel), and that might be 100kb. If you scale the image up or down, it changes the number of pixels and therefore the file size.
A
vector file on the other hand is basically just a co-ordinate list. This list is written in XML, which is human readable, you can literally read the list of co-ordinates in the file. Doubling the image size might be as simple as changing a 1 to a 2, rather than doubling the amount of data the file needs to store.
- Imagine you want to draw a diagonal line from one corner of a square to another using standard X,Y co-ordinates - in the format 'startX startY, finishX finishY'.
- You could express this as 0 0, 1 1 (i.e. you start at zero on X and Y axis and end one unit across and one unit down). A set of co-ordinates can also be called a node.
- In this example the coordinates are only one digit, so you have 4 digits (and a comma and spaces but we'll forget about that)... But, imagine the auto trace app has decided it was actually 0.0010 -0.002, 0.99183 1.0003 (which it'll do because of anti-aliasing in the original image) - this is now 28 digits.
Because every digit in the SVG file adds 1 byte, you can see that in our imaginary diagonal line, there's 4 bytes, and in the auto trace there's 28 bytes, you can
begin to see why auto-traced files can be much bigger than they need to be. If your image was made up of 100 of these diagonal lines, you should have 400 bytes, but you end up with 2.8kb in the above example.
SVGOMG etc, simply rounds these numbers down/up as per the slider. That slider is just rewriting the XML as '1' instead of '1.00003' for example, this is the major part of how it saves file size.
So when you auto trace, the app you're using looks at every pixel and tries to figure out what groups of pixels in the raster image it can draw lines around to make a vector image. Because there's thousands of pixels in most images, or more, the app will end up creating thousands of lines to draw around all the pixels, and each of these lines are defined by nodes. Given the auto trace will try and draw around groups of pixels that are a similar colour, imagine how many lines it has to try and make for the subtle fade in skin tone across someone's face etc., or something that is blurred, or has lots of different colours in a small space.
So...
- SVG file size is mostly 'number of nodes' × 'number of digits per node'... so considering all of the above, an auto traced image can have a lot of nodes, and each one might have lots of digits.
- You can reduce the number of nodes when you auto trace by going for a lower quality trace - but this can give you a very poor looking decal.
- Using SVGOMG will reduce the number of digits contained in each node. Often you can reduce it quite a lot with little to no perceptible loss of quality, then at some point it drops off a cliff.
It may only take a couple minutes to create an SVG from an image using auto trace, even if you have to drop it in SVGOMG on the way, it can be done in a few clicks. You can split images up as well, so if you get a 250kb auto trace of a photo, it's possible to slice it up into tiles that are 15kb each. This is tedious, but if you want to put a photo style picture on the car, it's likely the best way.
...
Secondly
Re-draw it yourself:
If you have access to a graphics utility like Adobe Illustrator, Inkscape, Affinity or CorelDraw, you can attempt to remake the image from scratch.
This generally requires only very simple tools within the application.
Line, Arc, Polyline, Rectangle, Ellipse to draw lines
Fill colour, line colour to make the above look how you want (gradient fills included in that)
Shaping tools such as
Union, Intersect, Back minus front, front minus back and
combine
Text of course
Then there's labour saving features such as
fillet or
contour, which can just save you time.
This really all there is to it. Any decal, in principle, can be created with just a few of these tools.
I made this video of me remaking a simple logo as a decal to to give an idea of a typical process.... It's neither that quick or that amazing, but what it should demonstrate is how simple the tools being used are, and that there's no artistic or creative talent required.... it ain't like learnin' blender.
You'll typically paste the image into the application and just draw over it by hand.
- If you were doing a simple company logo, and you have lots of fonts on your system... this might be a five minute job, it might be an hour.
- If you are doing something detailed, like a character illustration, it can take hours or days, or even weeks.
The benefit of going through the redrawing process, is that it allows you to draw things more efficiently than an auto trace will - you will likely use
significantly less nodes when drawing for yourself... and consequently end up with much, much smaller file sizes, likely of better quality, but at the expense of your time. Detailed looking Illustrated characters as one or two decals will almost definitely have been redrawn by hand, and you can often draw things up in such a fashion that you can adjust different elements of the character when on the car.
...
In conclusion... complex, good looking decals just don't happen without a lot of effort, there is no magic 'save as' or 'convertor' that will turn a complex image into a good looking useable 15kb SVG with a couple of clicks. It's not necessarily about the physical size or file size of an image that makes it difficult to produce as an SVG, but once you done a few auto traces and drawn a few decals from scratch you quickly learn to see why some requests are ignored, and why some are completed - and just because you've seen other really good decals on the showcase, it doesn't mean everything's possible, and almost certainly not just from a few clicks in a convertor.
