Hello people,
Everybody knows when the new version of control panel coming out?
Thx

Hi Robert,
You can save and reload your EmoKey mappings - just give them sensible names. With Consumer Control Panel, look under the EmoKey menu. With SDK EmoKey,look under the Application menu item. EmoKey Mappings are stored as *.ekm files, typically in C:/Documents and Settings/"your name" by default, but you can move them wherever you want.

I guess there's no reason not to be able to extract good training data. Don't forget though that a good action is distinct from your Neutral data in the same profile, which may differ from your current lot. We will have to do some research on this... Also, an Undo Last Training option has been on my desirable features list for about 2 years Thanks for your support guys Now I can beat up the programmers with some extra ammo

I actually agree to a point Alex said as I found 1 sys to be trained eerie like and others vary to degrees. I would at least like to be able to export the good training parts into some of the few last comps I haven't tested on. Sorry only so much time per day, and training on each comp sys is a task in it's self. Only compounded with the fact I still only go for 2 hours before 30 min break at least, I was hoping to go longer by now, but I'm only human and don't like discomfort for prolonged periods. I.E. I get a weird type of head ache more liken to muscle ache, and I don't want to over do it. On a similar note we need under user profiles a place for preset key maps and trigger times for various games or app.'s, and it needs to be sharable with friends. I.E. all the setup of key map to key with the 1-3ms triggers and 30ms hold or keep pushed type presets and not the trained individual actions.
P.S. a good use friendly way with option for usb drive export or website export Ect. for sharing, and helping the types that want more instant gratification after they train. + some gamers will want to try different setups, as I have found to be as I fumble threw games and try to find that sweet warm place that my mind interfaces with. Also on a side not I noticed using one set for to long and my mind wants to adapt, so that I loose some control of other sets. So far it looks like epoc and OCZ hit my sweet spot as neurosky has no native keymaper, and when I'm awake and alert the 2 do some sweet stuff.

We need to be able to train Cog actions independant of profiles.
So we could put reliably trained actions together into a user defined profile.

I have a couple of profiles that work very well with some actions, and other profiles that work better with other actions. It would be useful to be able to put all the best trained into one profile.

Especially as those action get refined and changed over time as the user gets better at focussing on them. I have older profiles that worked very well, but now my trigger thoughts have been refined so some actions in teh old profiles don't work. It is a lot of effort to clear and fully retrain those action. It would be much better if I could choose more recent versions fr om a list.

Also this would make it much easier for configuring for specific games. Especially if actions have been trained with specific thoughts relating to a specific game.
A different game may cause those thoughts to be naturally different due to physical differences related to the same type of action.
In which case you would have to create and train a profile for different games.
Far better if you could create a custom profile from existing actions.

We need to be able to deduct overlapping Cog actions that cause false triggers

I have had some success in deducting false triggers from unreliable actions wh ere they overlap.
To do this I experimented with training the nuetral in one profile with the similar action in a second profile (two control panels).
To clarify; an example:
I think left in one profile, it triggers push in the second.
I train neutral in the second profile (push profile) while thinking Left.
This reduces or stops the push being triggered when I think left.

So the nuetral in the second profile becomes an amalgamation of the normal random or quiet nuetral signal and the added left signal. So when I trigger Left, it sees this as part of nuetral, and doesn't false trigger 'push'.
I hope you understand what mean.

It would seem that the training can not only be used to detect and refine the thought pattern recognition. But it could also be used to compare and remove one trained action signal from another.

So if I trained 'Left', but for a few split seconds the signal was the same as 'Push'. The push signal would be included in the left signal as noise.
If I then trained 'Push' (a perfect training)
Whenever I think 'Left', There is a chance it may false trigger 'Push'.

So Left could be trained in a negative way to remove the push noise. Yes?

What I am getting at is, we need a magic button that when pressed, will compare the trained trigger signals. If a hint of one signal is found in another, it could be removed. Similar to dark frame subtraction in astrophotography.
You have a signal that has been trained, you then have a second signal that sometimes cause the first to be false triggered.
So while training increase S/N ratio by dilution of other signals. Deduction of the second signal will remove that noise. or somesuch. It's far too late in the night


LOL, pick the bones out of that

Quote

chad wrote: Is there going to be a new SDK version at some point,(or consumer version, however you all want to do it) that allows a little better control of training, such as being able to pick out 'bad' training sessions and delete them without having to also discard the good ones? maybe compare the similarity between the sessions, and allow you to discard the outliers (maybe got distracted during a training, or perhaps your activation itself is changing a little over time, and it would be good to discard some of the early ones) I think some kind of tool to examine the individual trainings in testbench would be great too, as that might allow the research users to come up with good concrete info for the 'Bootcamp' and help out everyone else too.

All great points, but in the interim can you at least add an 'undo last training session' button, so that we dont have to clear an action if we train one bad one.

Thanks