Replies

  • Not only do I look forward to the version 5.5 release, I'm totally excited about a version 6! Didn't even know one was in the works. Having loved every upgrade since SA version 1, I expect it to not disapppoint.

    Thanks for the prompt reply to my 32/64-bit question.

    --Joel

    • Version 6 is ultimately about breaking out the functional blocks and associated components of what is going on inside of this huge beast of a program known as Studio Artist.

      Having gone through one 'apple knows best rewrite everything from scratch for our latest and greatest (at least until next month) api.  And watching the result lead to features being lost, entire accompanying complementary programs being lost (erased from history is probably a more accurate description, like those institutions so lovingly described in novels like '1984' or 'Brave New World', institutions devoted to rewriting history as circumstances dictated), endless effort devoted to the redoing of things that already worked perfectly fine the first time they were written.  Because when apple says jump, developers are supposed to say yes sir, how high sir.

      Now if what you need to rewrite from scratch is pretty simple, then maybe its not such a big deal.  Although even then, there's this assumption on apple's part that developers are some kind of robotic slaves that will do whatever they want them to do, whenever asked.  And why get 2 or 3 apple engineers to make it easy to do it when we can just ask every single one of them to do it by themselves.  Yes, my apple developer PTSD disorder is rearing it's ugly head. You have been warned.

      And apple has thrown down the gauntlet again with this kind of thing in a series of moves over the last few years.  The big obvious 'much talked about in the press' one being the Intel x86 to ARM chip architecture change for cpu cores.  Coming in osx Big Sur, osx 11. 

      The cynical viewpoint would be that apple finally turned their mac product line into big clunky iphones that are missing the phone part (like the ipad, that big iphone you can't actually make a call on on some reason.  But it's the same os, running on the same chip set.  And this new ARM based mac could be destined to be much the same.

      My even more cynical viewpoint is that apple vies all of these devices a credit card readers, end of story.  Their point is to get people to charge purchases to a credit card (preferable an apple credit card), and preferably a occurring change that happens every month (easy to turn on, hard to turn off).

      ...

      But the whole take a lead pipe and wack everyone using 32 bit apis in the knees with that big heavy lead pipe so they just fall down on the pavement in pain and can't get up again, no one rushing to their aid, they just lay there and no longer function.  Apple essentially did that on purpose, just to be mean. An arbitrary mean marketing decision.  Fuck you developers.  Fuck you customers.

      Let's be 100% clear on this. They didn't have to do it that way.  They could have cared enough for their loyal existing developers, and by association, all of their loving and devoted customers of those developer's products.   They could have cared enough to make their desired transition smooth, move new code developments up the food chain while still insuring that all old software ever created to run on the platform could still be run on the most modern computers, in sandboxed emulation environments.  Because customers care about old data, about old but still very useful programs that would never be re-written if forced to do so.

      And we could have continued that whole group hug thing that was the cult of the mac, a bicycle for the mind, devoted to graphical presentation of ideas, tie in to spatial memory as an aid in organizing abstractions and turning them into documents, documents stored in a file system that had a whole graphical representation for working with it, smart people made that decision for lots of good reasons. And that graphical representation of information extended all the way from how you interacted with it in a file system, to what it looked like when printed on a page, and it looked gorgeous.  And indeed apple was very closely tied into the whole digital printing industry and market in those olden days (20% of apple revenue at one historical period was from selling printers, not personal computers).

      But we don't live in that world.

      So what is happening, and how are we addressing it?

      1: We have the Intel to ARM hardware transition going on for cpus used in apple computers (mac).

      And apple is providing Rosetta 2 to aid in this transition, so that customers can continue to run their existing x86 compiled 64bit software in emulation via Rosetta 2. Rosetta 1 being the equivalent developed when apple switched from Motorola chips to Intel chips. 

      Just like then, they will provide it for 2 years, because if they don't it might impact people purchasing their new ARM mac computers. After about 2 years, then they will arbitrarily cut it off, because they love to erase the past and pretend it didn't exist.

      2: We have the 'pretend 32 bit apis and applications never existed', 'wipe them all from the face of the earth transition', which is well under way.

      Because apparently the company with the largest capitalization on the plane can't be bothered to put a few engineers to keep them running in emulation sandboxes for anyone who needs to use them.  

      So we dropped all of those apis.

      3: We have a whole other thing we need to deal with that you may or may not be aware of depending on how technical you are.  Apple made a big point of supporting OpenGL as a graphics standard for 3D for many years.  And then as GPUs started to be used a lot for just general computation instead of just graphics, openGL Shader language was there to muck with to kind of do that sort of thing.

      Apple also announced with great flourish their own standard called  OpenCL for using GPUs for general purpose computing. Passed it off to an official standards committee.

      Then apple decided to create a proprietary closed off non standard based GPU universe called Metal. So current apple Open GL support is destined the fate of quicktime api, beware all who depend on it. And lets be honest, the entire digital art community was heavily invested in it, so apple saying fuck the OpenGL standard was like saying fuck you to every single person on the planet who does digital computer art.  Just saying what apple really thinks about you people as customers.

      Apple basically said to all of these fine creative people, We love you so much as customers digital artists, that we're going to insist that you rewrite every single piece of code you use to be based on an apple proprietary api that will not run anywhere else.  And that works better now on iphones than macs.  Have fun.

      ...

      Sorry: That was way more apple PTSD then i had intended it to be, my apologies, but it's worth understanding what is going on behind the scenes that needs to be addressed and dealt with for the software you love and use to continue to run in the immediate and longer term future.

      ...

      So, How are we dealing with all of these issue in Studio Artist 6.

      1:  In the short term we have Rosetta 2 to fall back on for the Intel x86 to ARM transition. But we're very committed to doing direct compiles for the new ARM machines as they become available (shooting for fat binaries).  We specifically choose low level libraries that are doing the very hard work to do that.  And do that well.

      2:  We have removed the quicktime dependencies that ended up biting us in Studio Artist 5.  And i have to be honest, when ever possible we will use cross platform software apis rather than apple proprietary ones in the future.

      Now why would we want to do that?

      I can tell so many war stores.  But let's just focus on OpenCL.  Proposed by apple themselves.  Good thing we didn't devote tremendous engineering efforts porting a ton of our stuff to OpenCL, because apple is wacking everyone who used it with a different lead pipe in the knees (or may it's the groin this time), knocking them down on to the sidewalk (where they lie helpless, because apple isn't supporting their own OpenCL standard). 

      So much for doing what apple tells you to do. Or what about that half baked not fully feature functional quicktime 64 bit replacement we were supposed to use, and then they killed it off too, screwing everyone who bothered to follow their instructions to use it.  Or who took them seriously when they said they would never abandon the carbon api library calls.

      3: Apple saying fuck you to OpenGL and everyone who depends on it is a big consideration to deal with. Huge ramifications if your application framework makes heavy use of it at low levels. Now people might make arguments that OpenGL was getting long in the teeth, had issues that needed to be addressed, and apple should have jumped onto supporting Vulkan, which was supposed to be the open standard to fix those issues.

      But apple loves proprietary apis these days that only run on apple hardware and lock developers in to that platform.  So that wasn't going to happen.

      Did we really expect anything different?

      For us, again the best strategy is to highly isolate our code from all of this.  Whenever possible.  Let the low level standards battle it out hidden away in some closet no one ever goes in.  We want our code to live on a much higher level.  Whatever knife fight is going on in that closet, we don't want to even know about it, let alone have to deal with it. 

      So if apple decides to screw over everyone again just to be jerks, and you know they will, you want to have a software api structure that hides all of that pain from us (and hence from you as well), and you also want a very large army of 3rd party open source soldiers that are going to be committed to dealing with the interfacing issues to fix the issues needed to be addressed with behind the scenes.  So you as an application developer dont have to worry about it.

      Again, i think we have positioned ourselves with a very good approach that is going to do this for us.  Isolate us from all this crap.  Which was certainly not the case with everything written that interfaced to quicktime.  Far from it.

      So, what have we learned other than john has apple developer PTSD?

      You got a little bit of insight into all of the crazy shit we've had to be thinking about, so that what we write will continue to run for you 10 years from now.

      So, a big component of the Studio Artist 6 design is to make sure we have a solid underpinning that is going to keep running.

      But we also want to start breaking out the functionality into modules.  Modules that could be combined together in different new ways.  That make it easier to try out experiments and hopefully get more limited functionality applications out to people to play with. That still can mesh together into a larger framework of how they can talk to one another, or the file formats they work with.

      I have a lot more to say about that, but we should save it for another post.

      What we will also save for another post is how deep learning neural net magic is going to change everything we do.  Everything everybody does in digital imaging.  And i do mean change.

    • Hi John.

      Thanks for the deep dive into the dungeon of software development, a place I don't and can't venture into.

      I have two questions:

      1. I run SA v5 on both a new Win 10 powerhouse which I put together myself and an aging Mac Pro (mid-2010) that I am looking to replace with some new Mac (probably not the too-expensive new Mac Pro, but maybe a new iMac with the new Apple processor if that happens). My question is should we be looking to Windows rather than Mac for SA?

      2. I am very curious about your statement: "What we will also save for another post is how deep learning neural net magic is going to change everything we do.  Everything everybody does in digital imaging.  And i do mean change." Can you tell us more about that.

      Thanks.

      --Joel

    • You can think of neural nets from a high level as 'software 2.0'.  What the people who coined that phrase mean is that traditional 'software 1.0' is all about hand writing a program, which is a series of steps that a human programmer has to define.

      'Software 2.0' is working with models that learn from data.  So you put together data that corresponds to what you are trying to do, and then the 'software 2.0' system essentially writes the software for you.  And our definition of what 'software' is has changed, because these neural net systems are really about learning transformations (at least that is how i think of them).

      So say you want to transform a  horse into a zebra.  How would you do that.

      A human programer would maybe try to use computer vision models to fin a horse in an image, and then maybe you'd try to write a texturing algorithm, that you would fit to a mask.  You can see that figuring out all of the details of this would very quickly get fiddly an very very hard. 

      The software 2.0 approach is to build a database of images of horses, and a database of zebras, and then let a neural net system analyze them, and learn a transformation from the one to the other.

      Wow, that sounds like magic, you might say. Yeah, exactly.  It's like we woke up and are actually living in Harry Potter's world of magic.

      So from the perspective of digital artists, this is incredibly exciting.

      So this is what i mean by 'everything in digital imaging is going to radically change'.  In new and potentially very exciting ways.

      How you take these new algorithms, these different kinds of neural net systems, and then turn them into something that is actually a creative and exploratory tool for digital artists is a very open and un-answered question.

      So that's the questions we hope to try to address in some form in future advances of both Studio Artist, and other potential new projects here.

      It's the reason why i'm so interested in trying to grok an artist's perspective.  How people might want to work with these systems.  How can you incorporate the human into the loop.  So artist and computer are working together to create a piece of art.  That has always been the Studio Artist philosophy.  So you can think of neural nets as a magic steroid we want to pump into our software to give to amazing new capabilities.

    • Thanks for the explanation. Software 2.0 sounds really exciting. I will be more than curious to see how you incorporate it into future SA releases.

    • 1: Which mac to buy is a real interesting question.  Obviously if you need an imac today, you have to buy the imac they sell today, which is intel based.  If you need to start using it to get work done the day after today, then it's the decision you have to take. 

      If you're trying to buy a machine that you want to use for years to come, and you can hold on a little bit longer with your older machine, maybe you should wait until the risc chip imac comes out.  

      Once the dust settles on the whole intel to risc transition on the mac, it's going to be interesting.  How you take a neural net model and then package it into a useful form for actual software application on a desktop or smart phone device is currently a giant mess.  But that may be changing based on some recent research i've been doing (since we need to deal with exactly this question).

      2: I've been babbling a little bit about a few things like style transfer in some other very recent posts here, so feel free to check them out for a few more glimpses. I'll try and post some more on this topic soon.

  • It's easy to tell, because Studio Artist V5 is only available as a 32 bit application build on windows.  And that will be the case forever for SA V5, SA V4.

    Studio Artist V5.5 will be available as a 64 bit build for windows. And it won't require quicktime for windows to be installed, since it uses our new cross platform 64 bit video framework, for movie file io, as well as for for live video capture.

    We had planned on releasing SA V5.5 in Sept.  So yes, we are late getting it out the door.  Unlike our country's president, i take full responsibility for my fuckups. So i do sincerely apologize for the delay in getting SA V5.5 out the door (both for mac and for windows customers).  Believe me, we all want to see it released sooner rather than later.

    Anyone buying Studio Artist V5 at this point will get an automatic upgrade to V5.5 when it is released.

    Let's just say that there are a lot of different amazing things going on behind the scenes here development-wise, because we're finishing up SA V5.5 to deal with apple dropping quicktime api and 32 bit app support on the mac, and windows customers have been wanting a 64 it build (since windows sucks when it comes to 32 bit virtual memory  allocation (less 2GB is all you get)). 

    And dealing with those issues for existing customers is great.  We need to do it. But it pales in comparison to the amazing things going on with the real work we're doing here, which is very much focused on the future.  Studio Artist V6 and what that will develop into.  We're been re-architecting Studio Artist 6 from the ground up, so that all of the core libraries it is built on will continue to run for the next 10 years.  We feel good about the decisions we made. We had to wait for some of them to become available, and it was worth the wait to be using them now,

    We're also doing a pretty heavy rethink on what the whole notion of a digital art program even means. The world is changing very quickly.  What you can think about doing with images on a computer is changing very quickly. We're very heavily invested in working to bring our customers the latest and greatest developments in deep learning neural networks.

    Certainly we are not the only company doing this. Every digital graphics program would be insane not to be doing so.  And how they end up using this new technology is very much going to be a function of how those existing companies work today, the kind of design and workflow decisions they make.  I often liken the results to 'design by committee', where a group of people try to decide on the 3 most generic features they can provide for the most people to adjust in the most limited way possible. 

    Studio Artist has a very different take on digital art then other digital imaging programs that are more about just photo manipulation, or really graphical layout when you are talking about the most famous one.  So we're trying to bring our customers access to cutting edge state of the art neural network technology in very Studio Artist centric ways.  We want to give the individual artist as much control as possible.  Not restrict your creativity.

    So that's what we're trying to do.

  • when i had 2 versions in my folder one of them said 64 bit. when i upgraded my new mac the older version doesn't work anymore. search the forum with your question. i have a feeling this has been discussed before but i just looked for the mac answers. i can't make videos on this computer unless i do image sequences of partition an older os.

    • So the Studio Artist V5 release folder has 2 different application files in it, one named Studio Artist 5 (which is the 32 bit application build), and the other named Studio Artist 5 64 bit.

      Catalina will mark the 32 bit one as not being runnable.

      You can always get a new release installer from your personal myaccount at Synthetik.com if you are a V5 customer.  in case you need to reinstall that Studio Artist 5 folder.

This reply was deleted.