gessman Loc: Colorado

That's not very fair. You don't have to turn control over to anything if you choose not to. Arsenal has an automatic feature that supposedly uses Artificial Intelligence in selecting your settings for your shots based on its analysis of the scene you point it towards if you choose to use it. If not, it has manual override and you can choose all the settings on your smartphone or tablet via a closed circuit wifi channel and retain total control of all your settings just as though you were holding the camera in your hands. Only difference is that you don't need to touch it as you shoot thereby reducing blur from shake as you shoot. It's just a remote control like any other remote control but with some features you can't get on other remotes. If you think your camera won't work unless you're holding it and turning the knobs with your fingers then Arsenal is probably not a good fit for you. If you like the idea of being able to put your camera on a tripod and control every aspect of it from a little distance away, this will do it and probably more efficiently than any other such product on the market none of which offers all the features that Arsenal does. Happy shooting!

dsmeltz Loc: Philadelphia

Fine. If you choose not to use the AI, what does it offer that justifies the $185.00 price?

gessman Loc: Colorado

There's several things I perceive to be advantageous to me but maybe not you. I guess we each have to decide what's good for us. I'm getting kinda old, 82, and live in Colorado, perhaps you noticed, and I'm not in the best of health, being a nearly fifty year diabetic with a lot of side-effects. I do a lot of my shooting up in the mountains. Sometimes it gets cold. If my existing health issues weren't adequate to justify my methods, I got my fingers and toes frostbit when I was younger and cannot stand cold but yet I like to shoot snow scenes and critters in the snow. While there may be nothing that can't be worked around, Arsenal facilitates how I want to shoot. I have a couple of other similar devices such as Case Air and use some remote control apps but Arsenal does some things that nothing else I have does.

Just one example, if I choose to shoot a sequence to focus stack, Arsenal does the stacking onsite so I can review the result of the stacking prior to final processing and make sure I have what I want before I pack up and head for home, assured that I won't have to trek back up later to re-shoot. Further, I can set my camera on a tripod out in the snow with a remote control 360 degree motorized tripod head and I can retreat to my car and shoot without the pain associated with the frostbite, using a 10" tablet for all the settings, transfer the stacked images to my tablet to review. Now, I realize that there are those who would criticize my methods as not being "real photography" but I really don't care what anyone thinks - I'm able to get the shots I want and not just sit home drawn up in a knot watching "the toob."

Do you recall the earlier urging of the pros back in film days when the old adage was, "film is the cheapest element of the equation so when you travel to a destination for a shoot, don't hesitate to use extra film and bracket or otherwise get a variety of shots so you have choices that'll save you from having to go back and shoot it over?" Arsenal takes that to a new level. Nowadays some cameras do that for you but mine do not so I'm really just buying a cheap upgrade that gives me what my camera makers left out when my cameras were new. Besides that, when I "knuckle down" my images are always better when I'm not physically touching the camera when it comes time to shoot.

P.S. I'm a "gadget guy" and like to experiment. 'nuff said!

BebuLamar

I don't know if it works as advertised but although I don't care for the AI I think the remote control aspect is great but I doubt that they could make it work.

gessman Loc: Colorado

It works, at least those parts of the software they've wrapped up. Part of the delay in the total release is there are some aspects they're still working on, one of which is what they refer to as a "grail time release," which is to be able to cover a period from dawn to dark or vice versa. There's a couple of other things like it only works with smartphones at present but will work with tablets once it's fully released. The Case Air I have does about the same thing but only for Canon and Nikon. I wanted the Arsenal because it supports my Sony a6500 right out of the starting gate and there's no attempt to even add Sony or any other brand to the Case Air. Case Air makes no pretense of using AI but otherwise, it does about the same thing as Arsenal and does a good job of it and it works with tablets. A nice little unit really.

I just looked on ebay to see what's happening with Arsenal. They have four up right now and looking at "solds," there have been 80 sold, lowest price and only one at $100 on up to just over $200. The four that are up for bid now asking $175 to $218, all new. They are closing in on having sold 72k since November 1st. It's moving along very nicely. There were two people in uhh who said they bought the Arsenal but have no use for it and will sell. I've contacted both of them and offered to buy theirs. One is due back from a trip on Feb 13th and we'll do a deal then and the other one hasn't responded to my PM, yet. We'll see what happens.

dsmeltz Loc: Philadelphia

OK. Those reasons make sense. I would have given it more consideration, if they were no stressing the AI so much. I may give it another look.

gessman Loc: Colorado

AI is the next "hula hoop." Everybody's gonna claim they're using it. I'm not so sure about how AI is being applied in this instance of these few programs that are claiming to be utilizing it. I think it would be pretty tough from what I recall about it when I experimented for a short stint programming in AI back in the early '80s with a language version called Prolog. It was laboriously tedious and over bloated then, as I recall, and very difficult to get it to do much of anything so I gave up on it. Spent more time setting up the "learning" database than programming. I saw a few days ago that much of the AI that's actually out there has been written in Prolog. Someone must have found some shortcuts that weren't obvious to me back then.

I mentioned earlier in another thread that the on-board database of excellent pro images that Arsenal "checks against" to come up with accurate settings is not new. The method may be but the idea isn't.
Nikon put such a database in a camera they named "FA" back in the early '80s but they didn't refer to it as AI. I picked up a very early copy of that camera and still have it. It was supposed to alleviate a lot of missed shots by looking at a scene using the total viewing screen and comparing it against those 100,000 images stored in that database. It worked pretty well if you were in a situation where lighting was goofy and had you fooled but it was far from foolproof. I relied on it more heavily than I should have a couple of times and got burned. I'd like to reverse engineer or decompile Photolemur and see just what they've done and if it matches up to what AI started out to be. I see that Skylum bought out Photolemur so they can learn from that team and incorporate AI into Aurora HDR and Luminar. We'll see how long it takes them to attach such a claim onto those two programs or if they'll come out with something entirely new.

Harry0 Loc: Gardena, Cal

Hmmm ... I have a few Arsenal Russian cameras. Any relation?

dsmeltz Loc: Philadelphia

Interesting. I also played with AI in the 80's for public decision makers. But I used a project weighting questionnaire combined with 3 point estimation that fed into a linear programming model with a feedback loop. It was fun. The difference was that the model's output was fed back to the primary decision maker and compared to the decision maker's expectations of the output. The goal was not to make the decision for the official but to focus their attention on the parts of the problem that could not be addressed mathematically. Rather than providing the answer the model produced, the system raised the question of "Here are the points where your expectations do not align with what the model is producing. What is it about these aspects that the model is failing to see? How important are these issue to you?"

The reason I used "project" weighting is that the possible combinations of projects under a budget cap are significant. A decision maker cannot evaluate all possible combinations and potential benefit outcomes by hand. And using LINDO was because, hey it was the 80's, we did not have DEA or any of the more sophisticated approaches available today.

AI is most useful when it marshals the users unique understanding on the effort at hand. It should focus the human element, not remove it. My concern with the AI approach in Arsenal is insufficient feedback leading to excellent but not exceptional results along with little learning or growth.

Harry0 Loc: Gardena, Cal

So. At what point is this tech to be built in on prosumer cameras?
Alreay a major portion of my shots are in Auto-ISO, lite Vivid, centered focus, with calibrated IS lenses. Point and shot, seemingly. And a little remote for tripod use.

gessman Loc: Colorado

So how successful do you feel your efforts were? It sounds like your effort was considerably more structured than what I did. Feeling adventurous and for my own edification, my scenarios were each contrived, probably less demanding than real life scenarios under pressure but still enlightening to a degree.

My comments about Photolemur may not hold water later but for now seem to be born out of a reasonable set of expectations. As things currently stand, I cannot see where Photolemur is fulfilling my idea of what AI is all about. When I first got Photolemur, which was as soon as I saw it after release, I put it to a test. I gave it a decent Raw file and saw what it would do with that which turned out to be to punch up saturation, contrast, exposure, a little sharpening, denoise, brightness, etc., kinda like I'd do with a few simple edits to bring a Raw file up to snuff. Then, I gave it a poorly underexposed Raw file to work with and as it turned out, it applied about the same amount of basic processing as with the better exposure which left the bad image seriously lacking in appearance. I cannot be sure how well my expectations matched the creators of Photolemur but I found the experience disappointing.

As I understand AI, Photolemur should have recognized, in its analysis, that it wasn't working with a well exposed image and put in a little more effort to bring it up to standards. As things currently stand it isn't observable that Photolemur has "learned" anything at all after several months although in all fairness, I haven't been nearly as busy as at times in the past so perhaps it needs more "feed" in order to indicate that it's learning. It'll be interesting to see if it picks up any since we're nearing Spring and I'll likely be shooting more often. As it currently stands, tapping the "auto enhance" button in almost any other software seems to do as about as much as Photolemur unless there's something I'm missing.

gessman Loc: Colorado

I'm probably not the right person to be replying to you but I'll just say that some of these features are in some of the cameras currently being produced, to some degree or another. As to whether the whole concept of AI will go on to be incorporated into camera bodies is anyone's guess or if it will ever even be deemed to be desirable. Like the rest of us, hang on 'cause we're gonna find out sooner or later and then it'll be up to each of us to decide to what extent we want to buy into it or just keep on shootin' like ol' Ansel did.

dsmeltz Loc: Philadelphia

The success was academically great, but practically more difficult. I was not able to continue in this since I was moving on to a different job. The problem is in the nature of political decision makers who, often, have difficulty understanding the difference between a median and a mean let alone understanding a variance off a utility curve.

gessman Loc: Colorado

Sounds like you had a good experience but you're making my head hurt now. LOL. It's never fun when you bust your hump and create a masterpiece only to have a DEU defeat everything. I saw an
ad on tv earlier today - Microsoft is now producing apps that use AI. I didn't figure it'd take long for that to happen. Adobe will probably be next.

dsmeltz Loc: Philadelphia

My general skepticism regarding AI is the over-reliance of the models on the models and the input quants without sufficient qualitative user input. However, with EVF's it could be possible to develop photographic AI that moves you into the ballpark and suggests other possible parameters to adjust. In landscapes it might even suggest changing POV. But the final adjustments must be placed on the photographer to avoid just accepting the AI. I mean for many who just want really, really good snap shots, it will still be useful. However, if you want to go beyond mediocre, the AI can only be one input in your process. But if all it is doing is selecting settings based on analyzed shots that "someone" thought were good, it is AI to achieve average.