iJenAPE. Be one with AVS.

dude, sweet (again) 🙂

Originally posted by UnConeD
What I meant with the passthrough thingy was simply to have an option to skip the motion detection altogether and just use it as a webcam plug-in. It shouldn't be too hard to implement and could be very useful. AFAIK there isn't a good one (there was one a while ago, but I heard it kept crashing).

hey man,
I'm sorry but we're not looking to do anything like that at the moment, in a nut shell, because we are working on the core maths engine all the time, and we don't have the time or need to create a better feed source for what we do right now, and thus no time to impement in the APE. Sorry (again)

We found the APE that put the feed on in the AVS, and it worked, but we couldn't see a use for it? Why'd you think it would be so useful?

We are going to give you the option of turning the camera off, so our engine process the motion going on underneath it in the AVS 🙂
So then everyone can use iJenAPE 😁

That is now a confirmed feature for v2.0 😎

If anyone out there has any other ideas or features etc, we'd love to hear them, and if we think we can put them in and they fit the 'philosophy' of what we are trying to do here, we'll add it to the list.

chilm.

Originally posted by Limpet
You could reinvent that game that came out a long time ago for the nintendo (or was it the snes?). It was basicaly an interactive 3d pong type game. Dont know if this is possible with your software though.

Hi Limpet,
That game sounds interesting, can you give us anymore of an idea what it was? 🙂 or how it played / looked etc.

I like the idea (personally).
chilm.

Originally posted by SYFi
We found the APE that put the feed on in the AVS, and it worked, but we couldn't see a use for it? Why'd you think it would be so useful?

What is so useful about the original design? To be honest I can't see much potential in this APE... at least not yet, the framerate is really poor on lower spec systems and the motion detection is also really poor (at least with my cam). What 'uses' does motion detection have in AVS anyway? Its really just something which may or may not be visually appealing depending on the individual, just like <insert any rendering APE here>.

I think that the coolest things about having an unprocessed feed are:

1.) It should be easy to add since your ape already gets the camera feed to detect the motion.
2.) If you were VJ'ing at a party you could blend in camera shots of the crowd which seems like something that would increase the usefulness of AVS in the area of live VJ'ing.
3.) You could use the image to feed avs... e.g. feeding fractal DM's, generating a texture, applying distortion to the image like a water bump for instance.
4.) You could make your own motion detection in AVS using this live feed hence rendering the original function (at least partly) redundant.

I think you should consider adding this feature, it will make your APE into something a lot more useful with minimal effort.

Originally posted by jheriko
What is so useful about the original design? To be honest I can't see much potential in this APE... at least not yet, the framerate is really poor on lower spec systems and the motion detection is also really poor (at least with my cam). What 'uses' does motion detection have in AVS anyway? Its really just something which may or may not be visually appealing depending on the individual, just like <insert any rendering APE here>.

I think that the coolest things about having an unprocessed feed are:

1.) It should be easy to add since your ape already gets the camera feed to detect the motion.
2.) If you were VJ'ing at a party you could blend in camera shots of the crowd which seems like something that would increase the usefulness of AVS in the area of live VJ'ing.
3.) You could use the image to feed avs... e.g. feeding fractal DM's, generating a texture, applying distortion to the image like a water bump for instance.
4.) You could make your own motion detection in AVS using this live feed hence rendering the original function (at least partly) redundant.

I think you should consider adding this feature, it will make your APE into something a lot more useful with minimal effort.

Hi jheriko,
I think we've come across in the wrong way about what we are doing with this APE. It was an idea we came up with a couple of weeks ago, and took 2 weeks to get to state it's in now (mainly due to windows interface coding). This is not our 'main' thing by any stretch. It is a mini-project to show partly how easy it is to integrate this generic motion detection engine inside any application. If you go look at our website, under the section where you can get the APE there is a section on 'Future Projects'. THAT's what we're up to 🙂

The configuration of the engine inside the APE has been set at some predefined values, which we believed would be suitable for most medium to high end systems. Many of the complex visualisations we've looked at run very slowly on low-end machines too.

The camera source is always going to be a problem, we know this, and the core engine can only perform as well as the quality of the data you provide it with. We're looking at high-spec composite cameras, and the like, for developing security systems etc.

The library is in beta development stages still too, and is currently to the greater part un-optimised, and still runs (in our opinion) rather well on mid to high spec machines 🙂

iJenAPE is just a cool toy, just like any other visualisation 🙂
It's just another way of looking at the world (man).

chilm.

Hm? I don't get it. You're doing complicated maths on the feed, yet you can't provide a simple option that simply copies the original captured image into the AVS framebuffer?

The use for a simple, live feed is endless. And if you don't realize that, you should download some of the more technical preset packs...

Originally posted by UnConeD
Hm? I don't get it. You're doing complicated maths on the feed, yet you can't provide a simple option that simply copies the original captured image into the AVS framebuffer?

The use for a simple, live feed is endless. And if you don't realize that, you should download some of the more technical preset packs...

... from previous post
"... we're not looking to do anything like that at the moment, in a nut shell, because we are working on the core maths engine all the time, and we don't have the time or need to create a better feed source for what we do right now, and thus no time to impement in the APE ..."

chilm.

I've seen your website, you seem to have unrealistic expectations for this technology.. especially since there are existing technologies (i.e. laser grids, radar, sonar) which are much better suited to many of these tasks due to their accuracy and reliability.

The only real benefit of using motion detection for security, 3d object identification etc is the cost, a webcam is far cheaper than lasers. The bluescreen++ seems like a practical idea though but you should give careful thought to the application of this technology to security systems and control systems, after all we live in a world of constant motion which will interfere with such systems. In case you didn't know there already exist affordable fingerprint scanners... they are very reliable in my experience.

Just out of curiosity what kind of method are you using to determine movements?

Originally posted by jheriko
I've seen your website, you seem to have unrealistic expectations for this technology.. especially since there are existing technologies (i.e. laser grids, radar, sonar) which are much better suited to many of these tasks due to their accuracy and reliability.

The only real benefit of using motion detection for security, 3d object identification etc is the cost, a webcam is far cheaper than lasers. The bluescreen++ seems like a practical idea though but you should give careful thought to the application of this technology to security systems and control systems, after all we live in a world of constant motion which will interfere with such systems. In case you didn't know there already exist affordable fingerprint scanners... they are very reliable in my experience.

Just out of curiosity what kind of method are you using to determine movements?

Hi jheriko,
We understand that what we purpose as concepts on our site may not be 'the best' way of solving some of the issues, but that they are 'do able'; plain and simple. More than anything we are trying to get across the sheer versatility of what we will ultimately be able to provide.

The existing technologies you mention are all great, and the points you bring up about cost are true, but you cannot say that we "have unrealistic expectations for this technology" as you have only seen the APE and do not really know what is 'going on' here with regards to our motion tech r&d 😉

The technology and the library we are going to release, gives any developer the ability to incorporate motion tracking software into their application. The whole thing is under heavy development, and this is only a beta release (and hence, will have limited features to begin with), but will provide the basic building blocks, as it were.

As for our motion detection methods ... seriously? 😉

chilm.

Originally posted by SYFi

The existing technologies you mention are all great, and the points you bring up about cost are true, but you cannot say that we "have unrealistic expectations for this technology" as you have only seen the APE and do not really know what is 'going on' here with regards to our motion tech r&d 😉

I might not understand the motion tech r&d but I do understand the basic laws of physics and how the world generally is... things are always moving so you will need a way to filter out small movements, since your motion tech only analyses an image (no mass, volume etc data) it will have to filter out small objects, small movements or a combination of both in some clever way. The system will never be perfect (in the strictest mathematical sense, like nothing is ever going to be perfectly 1 metre in length) which means that for control systems there will be some (even if rare) glitches when used in the outside world, as for security systems the whole idea is useful for catching idiots and animals but real criminals are smart and they will learn how to move slowly or trick your system into thinking that no movement has occured.. perhaps by slowly filling a room with mist, there are doubtless many ways that a motion tech security system could be rendered useless. Sometimes specific knowledge or experience can blind you when it comes to basic logical deduction, though the same is true in reverse that simple principles may not lead to the truth... I just find it highly unlikely that a feasible and useful motion detection system can be created with the limited processing power we currently have available.

Originally posted by SYFi
As for our motion detection methods ... seriously? 😉

I was really just wondering if it was anything much more than pixelold-pixelnew>tolerance and what sort of advanced maths is being used to filter your motion data (however you get it).

Don't take offense from my questioning... I just can't except anything without first picking at every flaw i can find. I think its a mathematician thing... it certainly is a useful trait for finding holes in proofs.

Originally posted by jheriko
"... since your motion tech only analyses an image (no mass, volume etc data) it will have to filter out small objects, small movements or a combination of both in some clever way ..."
"... as for security systems the whole idea is useful for catching idiots and animals but real criminals are smart and they will learn how to move slowly or trick your system into thinking that no movement has occured.. perhaps by slowly filling a room with mist, there are doubtless many ways that a motion tech security system could be rendered useless ..."
"... I was really just wondering if it was anything much more than pixelold-pixelnew>tolerance and what sort of advanced maths is being used to filter your motion data (however you get it) ..."

I'm sorry jheriko, but you're bouncing happily over the boundaries of what I can and can't openly discuss 🙂 What I can say is that with regards to the securty system, we don't see it as a replacement, fully automated system (yet😉), but more as an aid to existing security personnel. One guard could watch many more monitors, only reacting to motion notification, but as a back up in all cases.
We're not developing AI here, well, not for a while anyways 😉 and we believe that all the ideas we put forward on our site are feasible, and if nothing else an aid to existing systems 🙂

Originally posted by jheriko
Don't take offense from my questioning... I just can't except anything without first picking at every flaw i can find. I think its a mathematician thing... it certainly is a useful trait for finding holes in proofs.

Heh, your posts made me smile 😉 no offence taken. I like the way your brain works ...

chilm.

i found a error on your site.
in the archives there is a error involving the orginal Ijen ape download link.

Why are you reviving 3-year-old threads!!!

yea.. just make a new one and link to this thread if really needed