MechaLynx

@RobertGützkow I will, thanks for your time. If I may ask, is the hdri you used available somewhere so I can use it as a control test so that the results are comparable? It's ok if not, I just thought I'd ask, no rush ofc. Thanks

@RobertGützkow It would make sense to assume such a workflow but the question was how to render to images faster. The problem is, if they're rendering hdri-like images, then EXR will work out fine for them. If not, then the reasons behind their slow renders to PNG vs video may be different (such as disk I/O for multiple files, too high a compression on PNG etc.). Given the vague question and the specific answer, many might come under the impression the results are true for all cases. On the displacement, my claim isn't that it's strictly better, but contextually more optimal.

@RobertGützkow (sorry have to comment again to be clear) The reason we tested EXR32zip vs PNG16 is because both are lossless and we want to see if 32bit is worth it. The context is "what's the smallest file with enough precision". Though he didn't test EXR16zip, since the only difference is half the bitwidth, it's a reasonable assumption that it would be approximately half size. That's still over 2x PNG16 100% and even 15%. He tests 15% due to compression time. EXR16/32 on Pxr24 is smaller but lossy, which may be unacceptable in some cases. PNG16 is safer, small enough and precise enough.

@RobertGützkow I am not the downvoter, I decided on a neutral vote.

@RobertGützkow True, context-free, my comparisons are unfair. However, my point was that we don't know the OP's context and your answer is accurate for a specific context (HDRIs), as well as there being (non-trivial) contexts in which PNG is more optimal (e.g. optimal precision/size in disp-maps). I appreciate your tl;dr and I agree, if OP is in that workflow, EXR is needed. PNG is expectedly bad with unpatterned photographic data (i.e. HDRIs) - in PNG's design context, JPEG is available for that instead.

@RobertGützkow (sorry for the extra comment but I have to be clear after you added yours) There are use cases such as displacement maps (I've personally tested and compared PNG16 w/ OpenEXR32 and the differences are barely noticeable while exr32 is 5x times larger - Aidy Burrows agrees: creativeshrimp.com/normal-map-blender-tutorial.html), rendering sRGB output from sRGB input in the VSE (if all your inputs are video clips) as well as PBR maps in many cases where you may not need 8K+ resolution. I may in fact run the tests myself and if I do I'll post here. Thanks for the response.

@RobertGützkow I didn't disagree with the choice of having PNG at 100% when comparing file size but in the speed comparison. If I'm rendering an HDRI then I'm not going with PNG for obvious reasons. However, the OP doesn't state what they're using it for. PNG, depending on the render, can still provide small file sizes at sufficient depth/precision while still being fast at 15% compression rather than 100%. OP didn't state their compression level either. PNG, to my knowledge, was designed for general/web use. In a way, you're comparing PNG at its worst and OpenEXR at its best.

At the risk of disagreeing in a lion's den of anti-PNG experts, while I do appreciate this work I have to say I find it questionable that in a test of speed PNG is only tested at 100% compression (the test script has 100% compression by default) whereas (at least) OpenEXR has no settings for how far compression goes. Also, while I don't think anyone can disagree OpenEXR is the way to go for hdri images (which appears to be what is being tested), I disagree it's the format of choice for speed/file size in all cases. Some other types of renders work sufficiently on PNG16 at a far smaller size.

@CortAmmon Indeed, Dan Djurdjevic has written on this effect in martial arts. Joe could end up having an effect similar to mentalism on those around him, possibly convincing even attackers that they will drop dead if they hurt him and have it happen, even if it's just them believing it too hard. Could have further implications, such as unintentionally starting cults and religions when he says things.

This makes me immediately think Joe is an SCP. I think I've read more than one SCP with a similar effect. Not that this necessarily helps much to answer the question but browsing the SCP wiki for SCPs with similar effects might give you some inspiration based on how their effects work and how they're tested. Description: SCP-2344 is a male caucasian approximately ██ years old. Any statement SCP-2344 makes is immediately and unconditionally taken to be true by anyone who can hear and understand it.

@Rob You haven't proven you're worth a question but you have proven you're unwilling to share information. Why would anyone want to ask you a question and what makes you think any of the above was one to begin with?

@Rob If you both acknowledge misinformation and admit you will do nothing about it and find it adequate and acceptable to be condescending to others as if it is their duty to both be as knowledgeable as you and agree with you, then I find nothing wrong with considering your comments utterly useless and self-serving. The comments are quite adequate for a quick elaboration and links can supplement - this is obvious to anyone with common sense.

@Rob That doesn't add anything to the discussion, it's just condescending. Care to elaborate?

@Chris Most common in what way? Most any kind of software that uses numerical values for colors allows for RGB, even if they don't allow another system. There may be exceptions of course, but I think it's hard to argue that RGB isn't the most common primary color triplet. Personally, I haven't yet come across any kind of software that prominently used RGY and it's the first time I've heard of it to be honest. I don't see how it would be preferable from a UX perspective.

At the 1yr threshold

I wouldn't say it's overly optimistic to expect graduation soon, within a few months at most

Considering what stats I've seen for sites that have already graduated and the time they spent in beta, I would say graduation is almost overdue with WB

@HDE226868 Ha! I almost made a meta post about this xD

Forgot to add that for nomadic trading, the distances between stops can still be forbidding - how do you know the goods you pick up in one star system aren't useless by the time you get to one where they have value? It could be decades and comms are too slow to tell you whether the goods will be needed without some near-perfect sociological understanding that provides you with accurate predictions of civilization states. And even then, war, destruction and acquisition through other trade or plain luck can thwart your incredibly costly mission.

I almost posted this as an answer but it really doesn't qualify as one, thus I'm using comments. Essentially, by making FTL comms and travel impossible, you've made both comms and trade at interstellar distances so inefficient and ineffective yet so incredibly expensive that it is almost never going to be worth it. In order for a choice to be made, competing choices need to be inferior in some way or that choice needs to be special - the choice of interstellar trade in your case is always inferior to local information gathering and local production, thus it has no reason to exist.

I think you've worked this premise into a logically un-solvable situation. Pretty much every answer boils down to either trading information or nomadic trading. It's already been mentioned that the round-trip time for comms is so long, that information may well be value-less by the time it's traded (you need to know what the other side is offering, but when you do, you know what to look for and thus get the info yourself). Nomadic trading on the other hand is circumventing the lack of FTL - the goods travel slowly, but the trade happens locally, speed is irrelevant.

"Most effective long-term zombie killing weapon" - the first thing that came to my mind is "the cure" :P that 'long term' factor shifts the perspective radically.

Bye and thanks :D

@BasileStarynkevitch hmm I might - I just think it's tangential and not the core issue. I'll add some details, but this chat session will be available anyway.

from my experience on SE, it'll just be left here as it is :P

Ok

Our discussion gave me as much as I can get I think

do you think I should delete the question or leave it?

btw

anyway, I don't want to hold you up with what is essentially a theoretical discussion - I'll check that link you gave me and use it as a lead to look into this more. I'll probably study how openGL does it as well and clarify my api needs to see if it boils down to possible or impossible problem.

@BasileStarynkevitch but i have a strategy for that

@BasileStarynkevitch well, the latter is also true :P

@BasileStarynkevitch true - it isn't that i'm worried people will complain, it's more that someone might hijack it and make the world explode

@BasileStarynkevitch unfortunately i'm kind of constrained to the web platform.

@BasileStarynkevitch I might just disable external plugins until I'm comfortable with the security, but it may be too late if it then needs extensive refactoring

@BasileStarynkevitch For web documents, of course. That's what it's designed for. But when it comes to chrome level code, you can make something as big as Firefox if you want - that's too powerful. My app will include networking code (hence the possibility to implement a bot-net or ddos someone), database code (hence the possibility to farm information remotely) and DOM manipulation with user information logging. Javascript wasn't built to handle this kind of security.

and may have up to 500.000 users

mine is oriented towards non-developers

But that's oriented towards developers

In the same market, there are others who created similar yet much smaller extensions that have exploited this lack of caution

@BasileStarynkevitch I'm not worried I'll get sued or anything, I'm worried it will generate an ecosystem of malicious code propagated to users with the promise of convenience.

I probably need to work out exactly what the api is, before deciding

but I think the message + api filter may work - in a specific enough setting, i think I can avoid accidental turning-completeness

@BasileStarynkevitch lol it may boil down to that, I might not be able to protect things enough

that's a great link btw, thanks

the target user base

and the user base is not nearly cautious enough

@BasileStarynkevitch That's what I think I can't do practically, there may be too much activity to handle

since it will be released only through secure channels (mozilla's AMO, chrome store) it shouldn't be an issue if the core is secure

I will, I just want the release version most people will use to be secure enough