forums.ps2dev.org

chameleon_789 · Joined: 01 Nov 2005 Posts: 9

I've just bought an IRdA keyboard to play around, and possibly write drivers for the PSP..

I couldn't get the IRdA keyboard program that was posted here to work, so I tried Vanya Sergeev's IRdA raw input prog (modifed to press O after every keypress which added some 0's to the file making it easier to read)... but, I can't see any sort of pattern which I'd be able to use to work out what key is being pressed.

Seems like there's 4 apparently random hex digits, followed by C1FFFF, repeated 3 or 4 times depending on how long the key's held. These are the results I got from repeatedly pressing A:

seventoes · Joined: 02 Oct 2005 Posts: 79

Theres a serial port echo program included with luaplayer, try that out maybe? I dont know anything about Ir, but i dont see a patern in those numbers either.

chameleon_789 · Joined: 01 Nov 2005 Posts: 9

ty, could help a lot.. i'm taking a look now. So far, no luck.. i've seen BA6C C1FFFF AF1F C1FFFF 0909 C1FFFF for B quite a few times, but I've had the same code showing up using S.. not sure what's going on there..

Dr. Vegetable · Posted: Sun Dec 11, 2005 9:31 am Post subject:

Out of curiosity, what keyboard are you using? I ask because I have a Belkin IR PDA keyboard that I had been thinking of doing this with. I had spent about 15 minutes playing with AgoIn's keyboard sample, but got flaky results. I never bothered to try to decode the Belkin output, but have meant to circle back at some point. (Too many projects going on at once!)

Make sure the keyboard is close to the IR port on the PSP, with straight line-of-sight. My IRda keyboard had to be very carefully positioned in order to get anything at all. I think most low-power IRda devices have a range of about 12 inches.

Shine · Joined: 03 Dec 2004 Posts: 728 Location: Germany

Perhaps it is the IrDA Control protocol. You can search Google for the specification.

chameleon_789 · Joined: 01 Nov 2005 Posts: 9

I'm using this keyboard (although the one I got was branded under a different name, and came with more drivers than the online .zip). It's got two IR ports, one which is adjustable so positioning isn't a problem, the only problem is the cradle doesn't like the PSP :)

Adesso SK-6688

Decoding 'by ear' is more difficult than I imagined, so at the moment I'm trying to decompress the symbian drivers to see if there's any way of getting the keycodes from there.

edit: thanks shine, I read a coule of specifications but I think the irda control protocol is built in, the psp just transmits/recieves data packets from it.

Dr. Vegetable · Posted: Sun Dec 11, 2005 6:01 pm Post subject: IrDA Monitor Program

I am not familiar with Vanya Sergeev's program, but I have modified Shine's IrDA sample to operate as a simple hex display of raw IrDA data. You can download the program (with source) from here:

IrDAmon.zip

Using this program, I am getting fairly predictable results with my Belkin keyboard. I get 4 bytes per key-down and another 4 per key-up. It seems that the keyboard is sending raw scan codes, as the values sent do not seem to be affected by holding down the Shift key, etc.

Next I will map out the scan codes to see if I can make a simple text input program for this keyboard. Hopefully, we can eventually incorporate support for all of the commonly-available PDA keyboards in any driver that might get written.

As this is my first published homebrew application, I welcome any feedback.

UPDATE:
Ok, I'm on a sleepless roll. I mapped out the scan codes for the Belkin keyboard, and wrote a simple typing program. This is just a quick-and-dirty application that doesn't support backspace, delete, or arrow keys, and simply echoes keystrokes to the screen. It does track the state of the Shift and Caps Lock keys, although it doesn't handle Shifted Caps Lock combinations correctly at all. The table in this program contains the codes for all keys and buttons on the keyboard, although some are not supported (or even scanned for) by this program.

The keyboard is made by Belkin, Model No: F8U1500-E, and is available at Staples for about $60. It is a fairly decent design, the keyboard folds in half and includes a PDA stand that does a nice job holding the PSP. The stand has a metal reflector that can (and must!) be carefully positioned to mirror the IR from the keyboard into the PSP's IrDA port. Once you get it adjusted, the comms are reasonably reliable.

You can download the application (with source) at:

IrDAtype.zip

I have not yet attempted to incorporate this into AgoIn's Keyboard Demo application, although the basic concept is the same. Although I haven't looked at his source code in awhile, I wouldn't be surprised to find out that my code is very similar to his. I seem to recall that somebody made a library out of his program, so I'll see if I can find that and add what I have learned.

I did notice that at times, the PSP would start receiving a stream of random crap, depending on how I positioned the keyboard and PSP. I think there is some form of IR interference around here, either off my desk lamp, or perhaps off my laptop computer. I am pretty sure that my laptop doesn't have an IrDA port, but sometimes those things are easy to overlook!

Again, any comments or suggestions are welcome.

EDIT:
Apologies to Vanya Sergeev, who wrote the original IrDA program upon which Shine based his program, upon which I based my program...

Also not sure why I thought AgoIn wrote the keyboard demo. The demo I was referring to was actually written by Tim Underwood.

EDIT #2:
I found a stupid mistake in the original program that limited the rate of input from the IrDA serial port to one byte per vsync, which limited the program to receiving about 15 keystrokes per second, max. (Keyboard input would slowly trickle in if you typed faster.) The link above now points to the updated project that processes all available IrDA bytes, for immediate keyboard response.

chameleon_789 · Joined: 01 Nov 2005 Posts: 9

Nice code :) Being a beginner, I've learnt quite alot from looking through it. Using your IrDAMon, I was able to out some more about the scan codes from the SK-6688 (enough to map the A-Z 0-9 keys) :

-Each code is 2 bytes long.
-For each key, there are 16 unique keydown and 16 u.nique keyup codes, making a grand total of 32(!) unique codes for each key (the codes can be in any order, a single keydown code can be follwed by any of the keyup codes).
-While the key is held, a keyhold code is transmitted, which can be any one of 16 codes (not bound to a specific key).
-In between codes the 3 byte 'endcode' 0xC1FFFF is sent.

chameleon_789 · Joined: 01 Nov 2005 Posts: 9

On another note... I actually emailed the makers of the keyboard asking for the specifications of the data transmissions. I got a reply saying they are "the most secret & important property" and "We are not allowed to release them as we need to protect our software"... ironic I worked it out like 2 hours before I got their reply :D

If a beginner can work it out in the time it takes to send that reply (literally a couple of hours of my spare time over two days), why not just send the codes? lol..

Dr. Vegetable · Posted: Tue Dec 13, 2005 10:50 am Post subject:

Great! Glad to hear things are starting to make sense.

That is kinda messed up that they use 16 different codes for each keydown. I suspect that there is some obscure pattern to the set of codes used for a key, and if you can decode it, you'll save a lot of brute-force effort. If you have a chance to post some examples of all sixteen codes for a few keys, I'll take a stab at figuring it out.

I made this post in the PSPSDK forum to ask if anyone has thought about creating a standard library to easily add IrDA keyboard support to homebrew apps, but no response so far. I am thinking that it might make sense to see if we can add keyboard support to the p-sprint keyboard emulator library. Maybe the library could check for a keymap.dat file in a standard location on the memory stick, and use the specified IrDA keyboard, if any, instead of (or in addition to) the PSP button combinations.

Anyway, keep on plugging, and good luck!

dot_blank · Joined: 28 Sep 2005 Posts: 498 Location: Brasil

this is becoming interesting ;)
i look forward to see if people
around the world can get this list
of irda keyboards and what they signals
are ....and making a keyboard driver
is very much needed and very interesting work

Dr.V: great source :) ...i have similar keyboard :D
you can find great prices for some wireless
keyboards at this link http://wireless-keyboard.com/c/1497/Home-Video/wireless-keyboard/index.html
works just fine and i like the legos!!
especially this silly guy right here
http://drvegetable.com/images/throttle.jpg

;) cheers
_________________
10011011 00101010 11010111 10001001 10111010

Dr. Vegetable · Posted: Tue Dec 13, 2005 1:25 pm Post subject:

Good links! (I assume these all lead to eBay listings?)

These links appear to be the same keyboard that I am using:
Belkin Wireless Infrared Keyboard ($18.99 starting bid)
BELKIN Wireless PDA Keyboard F8U1500 Works w/Most PDAs ($9.99 starting bid)
Belkin Wireless PDA Keyboard ($15.95, reserve met)

I actually got mine on sale at Staples for $35, but the normal price is just under $60. There are a few other IrDA keyboards on this list that would be nice to support as well.

One thing that would need to be thought out a bit is support for international character sets. My sample program only handles 7-bit ASCII. With a little bit of work, it should really be able to read a configuration file to map scan codes into unicode characters for true worldwide appeal.

Glad you noticed my LEGO stuff. The same things that draw me to hacking the PSP were what originally attracted me to the LEGO hobby. I am hoping at some point to do a few projects that incorporate both interests.

EDIT:
Actual price for the Belkin keyboard at Staples is just under $60, not $70 as originally reported.

chameleon_789 · Joined: 01 Nov 2005 Posts: 9

As requested, here are the hex codes for a couple of keys.

I couldn't see a connection betweeen them myself, but there must be some sort of pattern, it doesn't make sense that they'd store multiple codes for each key. There are also some anomalies to the normal formula, B seems to only have 15 codes for keydown, and some keyup codes are >0x8000.

pspvideo9 · Joined: 15 Mar 2005 Posts: 19

We're curious, do you see any lag when printing characters to the screen from your IR Keyboard at high wpm rates?

Dr. Vegetable · Posted: Thu Dec 15, 2005 8:09 am Post subject:

On the Belkin keyboard, yes. If I type a bunch of keys, the data seems to trickle in. I haven't measured the maximum rate, but that could be added to my sample program pretty easily, I am sure.

Speaking of the sample program, it is possible that I am doing something foolish that would artificially limit the rate of input.

pspvideo9 · Joined: 15 Mar 2005 Posts: 19

We are seeing such lag on our ThinkOutside IR Keyboard as well.

The odd thing is, for us, the lag isn't in the keyboard. We can quickly type out a sentence, then move the PSP away from the IR keyboard sensor, and the _all_ the data ends up being displayed but only at a "trickle".

From our investigation, the sceIoRead is causing some unknown delay, because clearly all the IR data is getting to the PSP in near-real time.

Do you get the same such results with your keyboard?

Dr. Vegetable · Posted: Thu Dec 15, 2005 8:34 am Post subject:

Yes, I get the same results. I can type a bunch of stuff (or run my finger across the QWERTY... row) and then move the keyboard out of line-of-sight. The correct characters continue to trickle in.

Are you using my program, or something else? I ask because this was a quickly thrown together app, and it could be written incorrectly. It sounds like you have isolated the trickle to the firmware, though...

pspvideo9 · Joined: 15 Mar 2005 Posts: 19

I was hoping it wasn't the firmware, looks like thats the culprit. It seems like there is some sort of internal buffering going on, which is totally odd since the PSP reads the IR data in real time.

Nope, I'm not using your program. I'm using my own code to read in characters from my ThinkOutside IR keyboard.

Dr. Vegetable · Posted: Thu Dec 15, 2005 8:54 pm Post subject:

Ok, thanks for the added info. I'm usually more relieved to find that the bug is not in my code. :) Are you planning to publish the keycodes so that we can incorporate them into the upcoming library?

chameleon_789: I haven't had a chance to play with the keycodes you posted yet, but will try to do that today. But I am surprised that any hardware manufacturer would guard their protocol so closely. After all, getting people to support their hardware on other platforms can only help their sales. Why they would think any of their competitors would want to copy such an overly complicated protocol is beyond me. Maybe if you explain to them what you are doing and how it could benefit them, they might loosen up a little? If that doesn't work, ask to speak to a manager until you get someone who doesn't have their head up their ass...

Shatterdome · Joined: 22 Nov 2005 Posts: 11

I'm not sure if the is an sceIoPeak function ?

I know when reading keys from the psp if you use sceCtrlRead it is slow, but if you use sceCtrlPeak it reads keypresses real time...

dot_blank · Joined: 28 Sep 2005 Posts: 498 Location: Brasil

sceCtrlPeekBufferPositve is best for use
in inputs that require realtime input ...ie..at the
rate of 60hz ....i think thats pretty fast enough ;)
_________________
10011011 00101010 11010111 10001001 10111010

Dr. Vegetable · Posted: Tue Dec 20, 2005 10:23 pm Post subject:

Good tips! I will try this out when I have a moment.

I haven't had much time to work on this project the past several days, but (as previously mentioned) I would like to incorporate this into p-sprint. I will try to support any keymaps that are documented, and hope to help chamaleon_789 decode that oddball keymap as well.

I'll have more time to work on this next week.

user303 · Joined: 15 Dec 2005 Posts: 2

I have converted the codes to binary and tried to see if I could discover some sort of pattern but all I came up with was the obvious.

Dr. Vegetable · Posted: Wed Dec 21, 2005 6:14 am Post subject:

Playing with various bitwise logical operators for about 15 minutes, I am starting to see some patterns emerge from this data. The only clear pattern so far is that, if you take the keydown codes for any given letter and XOR them together, every fourth entry will result in a zero.

ironlung · Joined: 21 Jul 2005 Posts: 13

thanks to everyone in this thread - it seems like progress has been made in the right direction! I still have my Adesso KB in hopes of someday putting it to good use on the PSP.

btw- I'm startled at the prices going UP - I bought mine for $12 or $15 new at geeks.com - now I can't even find my KB on Froogle...

I am glad that Dr. V made a post in the dusty, cobwebbed IR Keyboard thread from 6 months ago so I could see what's goin' on.

[finds comfortable spot to lurk]

Dr. Vegetable · Posted: Wed Dec 21, 2005 12:03 pm Post subject:

Yeah, if I'd been paying closer attention, I would have realized that chameleon_789 posted a link to that dusty old thread in his first message of this one. While I'm cross-linking threads, here are two others that are relevant to this topic:

http://forums.ps2dev.org/viewtopic.php?t=2685
http://forums.ps2dev.org/viewtopic.php?t=2604

Anyway, I am on vacation from now into the new year, so hopefully I'll be able to devote some energy to the multitude of PSP projects that are piled on my plate.

Arwin · Joined: 12 Jul 2005 Posts: 426

Ok, Dr. Veggie, I'm on standby when you pick this up. I'm about ready to do some more work on p-sprint, after having worked my arse off for a project at work.

Dr. Vegetable · Posted: Sun Feb 12, 2006 12:58 pm Post subject:

I've been working on integrating my IrDA keyboard code into Arwin's p-sprint library, and making decent progress so far. I am trying to make it easy for people to add support for other keyboards. For now this will require writing a little C code to parse the keyboard's input stream. Eventually it might be possible for a user to train the system to recognize almost any keyboard, but not yet.

In the course of doing this, I took another look at my IrDAtype sample program and realized that I made a rookie programming mistake that was causing the slow keyboard response that I reported earlier.