In this thread we talked about print quality, how good a print is, what influences a print and that even a simple photo can change the look, which is for our photos most of the time unintentional. So some people decided to try to really compare their prints (yet to be decided who that really is ). For this to work I created this thread for defining the terms. We have to define several things to get comparable results, I want to start with the basics before we go any further. The basics are: Filament Model The next step would be: Slicer Slicer settings And if we all printed something we have to talk about how to photograph what we have done... To start I vote for the following: Colorfabb XT dark grey or light grey For the model I have no idea what is good, a benchy takes too long to print IMHO, I print "twin test cylinders" I had since I read this guide from here. I normally test infill, extrusion multiplier stringing with it. I print them without infill and bottom/top surfaces, sometimes with but I am open to suggestions...
I would like to point out that there is a "standardised" photo rig for the benchy at thingaverse... I put a link when I find it. http://www.thingiverse.com/thing:1085472 If we assume that all has some sort of smartphone then with this we get the same framing. So if we make something that print fast but still fit in this rig we should have something half way comparable. I think that colour control should be removed from the specification. But lighting (placement and type can be good)
I don't agree. For me color has a huge influence on the ability to make a photo that shows reality. Compare the two cylinders from my other thread. They are printed with the same parameters and looking at the in real life the black is as "bad" or as "good" whatever you think they are as the white but the white shows it more on the photographs...
White always looks worse on photos, I think that's the colour more than anything. For print quality my default is the 3d hubs Marvin, it has overhang and curves and prints in less than 20 minutes so I find its a good indicator of what the quality will be. Here is colorfabb pla/pha at various layer heights that I was testing
Hmmm didn't think of the Marvin. Yeah that's good. Let's collect some more opinions then we work out the terms...
For the color I still think what I said the gray tones could be good. Not so hard as white and not so forgiving as black maybe...
If you expect all be able to have a greycard, calibrated screen, etc. You are in for a very hard work to standardise this... have been working on my colour control for my photography and it is not easy to get it right. So I would go for black and white and focus on the lighting...
Well you can't get it perfect. The photo taking is the problem I agree. But to try is better than doing nothing. We gotta test it. Let's start at the beginning. I think we all can benefit from this even if we just try...
Marvin tests slow printing and details and fan. A cylinder test layer regularity - the Z. A cube tests corners and ringing - if it's a big enough cube to pick up some speed. We could make a hybrid object with Marvin standing on a 5mm high 2cm cube (don't need much height to see the corner effects) That is situated in the middle of a 10mm high 3cm (diameter) cylinder (a bit more height because looking for Z effects) Print the first 15mm at 0.25mm layer height (problems show up more clearly) And marvin at 0.1mm. (because he's a detail-test!) A low infill % throughout. For example... @GrodanB love the standardised environment thing on Thingiverse....! Perhaps one more barrier to entry, though.
That benchy photo stand is cool... Maybe an option... @R Design love the idea of the Marvin on a pedestal and the different layer settings, maybe we can also combine with a cylinder... We could modify the benchy photo thing for that Marvin... But I think not the model is our main issue and if it just shows a specific problem. If we can standardize the rest as much as possible then we can use different models for different effects we wanna compare...
Ditto, so my favorites still XT dark or light grey, didn't print with both yet so no preference there...
You should standardize the temperature as well. PLA @ 185 will look very differently than PLA @ 210. For one, there's the glossyness (gloss reveals the mistakes and issues immediately), but there's also the viscosity. The hotter, the more liquid, the more liquid, the more succeptible to layer unevenness and ringing. Also, disagree on black hiding issues. It depends on what you're looking for. Glossy black is very revealing for ringing, corner blobs and top surface quality. White hides those, but shows others. Silver can be a nice compromise. Incorrect, ”picking up speed” has nothing to do with ringing and corners, you can test those with a tiny cube as well. They're dependent on jerk and acceleration, and those are constant with travel, so size is irrelevant, as long as acceleration isn't nerfed. Here's my test cube for ringing and blobbing: http://www.thingiverse.com/thing:277394 Example clean print: https://lh3.googleusercontent.com/C...tW0Q9qKxlBEo-sUoxftEO5n48PyFLIaX7WBvgjiuQ5R1M Example print exhibiting issues: http://thingiverse-production-new.s.../bc/d5/47/4c/c0/_7280628_preview_featured.jpg
Sounds true. I think I confused one thing with another: when printing small things S3d by default slows things down (unless you are printing a lot of them) which in my experience removes artifacts. (I really do have the impression that slowing down from 50mm/s and slowing down from 5mm/s is not the same, even if acceleration settings are the same in both cases.) It would probably be wrong to turn off the 15 sec layer rule (minimum time for a layer before S3d will slow the speed down in an attempt to achieve it) for fear of creating heat/melt related artifacts. In which case doesn't that mean that cubes should be a certain "minimum" size (for a given speed)?
Ah, true, KISSlicer has a similar setting (minimum layer time). And yeah, it's used to prevent overheating of small parts. I think it's safe to disable that when testing, cubes, even hollow ones, have plenty of surface area to dissipate the heat, but even when on, it shouldn't influence things much, unless we're talking extreeeeme slowdown.