Home Trainers: A challenger appears.

Marcus · Post by **Marcus** » Wed Jan 05, 2011 8:10 pm

Nate,

try this with a scatt file

http://www.centershot.com/downloads/scattsdk.zip

As is. No guarantee. Use scattexp.vbs on a scatt file. exports data to excel

Marcus

paulo · Post by **paulo** » Wed Jan 05, 2011 8:55 pm

NateG, I find the data analyzes idea of great value, to creat a model of your own perfect shot, and not so perfect.
Would audible cues tuned to your analyzed shooting data be a good thing to have implemented, or just a distraction.
The element of stopping a shot has been my biggest hurdle to surpass, trying to make it part of a trainer would be of great value to me, certainly as a function you could turn on/off.
It certainly is great to have space and time, but to go beyond that is the goal for all regarding an investment of this kind.

David Levene · Post by **David Levene** » Thu Jan 06, 2011 2:01 am

NateG wrote:One feature I wish my SCATT had--enough that I could be swayed to buy something in the $500 range-- is better data export.

Have you checked this out thoroughly?

I have seen full data dumps of all of the XY coordinates, 200 sets per second of control time.

I know this was originally done by someone hacking the Scatt software but thought that Scatt now had a procedure for doing it.

If your local importer cannot help then you could try contacting Scatt direct.

paulo · Post by **paulo** » Thu Jan 06, 2011 9:19 am

Also like the heart bit and trigger finger monitors idea of the scatt?
That is for the second phase.

Is the Z axis of importance at all?
Can it be incorporated, or it would need infrared for distance to target registry?
This is certainly third phase.

The botom of this page http://www.scatt.com/software.htm seems to make you believe scatt provides some data, would that be enough for "perfect shot" analysis/modeling ?!

A shooting pattern tool, would also be interesting after having data from several matches/training sessions, something that points out peaks and valleys in performance, with an attempt at the why's, and a more stable performance in the future, at least at my level it still is important.

David Levene · Post by **David Levene** » Thu Jan 06, 2011 2:18 pm

David Levene wrote:
NateG wrote:One feature I wish my SCATT had--enough that I could be swayed to buy something in the $500 range-- is better data export.
Have you checked this out thoroughly?

I have seen full data dumps of all of the XY coordinates, 200 sets per second of control time.

I know this was originally done by someone hacking the Scatt software but thought that Scatt now had a procedure for doing it.

Try looking at the "How can I get source information (for example, position data of aiming points) out of shooting files?" section approximately halfway down this FAQ page.

NateG · Post by **NateG** » Thu Jan 06, 2011 6:36 pm

Marcus, (et al.)

Thanks for the information about the SCATT export stuff. I'll have to check it out later, when I'm at my laptop with SCATT on it.

AAlex · Post by **AAlex** » Thu Jan 06, 2011 9:51 pm

paulo,
Hardware you suggested are below the minimum spec.
I'm researching other options that look promising.
WRT additional features - let's take it one step at a time; those are off the radar at this time.

NateG,
The use of good camera with powerful lens was to facilitate having the camera next to the computer, without the need to setup the camera downrange.
The alternative is to use cheaper camera (if I find an acceptable solution) positioned next to the target with 10m USB cord.
Placing a mirror downrange and having the beam-target at the firing line is another alternative, and it doubles the effective distance.

Use of IR laser has more cons than pros:
1) Would require modding a camera, since by default most of them come with IR filters.
2) Working with IR lasers without laser-safety goggles is unwise.
3) They are more expensive and options are limited. IR laser sights are outrageously expensive.
4) There's no real advantage WRT possible distance.
5) There's no real reason not to use the dirt-cheap red laser, as you could only see the beam spot with the non-shooting eye some distance below the target bull. Using a blinder or a partial blinder would make it a total non-issue.

Brian M · Post by **Brian M** » Thu Jan 06, 2011 10:01 pm

AAlex wrote: The alternative is to use cheaper camera (if I find an acceptable solution) positioned next to the target with 10m USB cord.

USB cables are limited to 5meter for "effective use", you can put a powered booster in there to get 10M, but then you have yet Another power supply needed right in the middle of the run. Everyone can make their own decision, but I'd rather have the camera on my shooting table.

AAlex · Post by **AAlex** » Thu Jan 06, 2011 10:12 pm

Brian,

That depends on the application. Camera is mostly one-way data flow and latency is of little concern. I hooked up my camera with this cable with no issues.

Having said that, I prefer camera next to the shooter option as well.[/url]

AAlex · Post by **AAlex** » Wed Jan 12, 2011 9:06 pm

PS3Eye camera performs adequately, bringing down the total costs close to absolute zero.

If anyone is interesting in beta-testing, please PM me; having any technical background is a plus.

You'll need the following:

1) PS3Eye camera, e.g. http://www.amazon.com/PlayStation-Eye-3/dp/B000VTQ3LU
PS3Eye works on Windows using 3rd-party driver: http://codelaboratories.com/get/cl-eye-driver/
You'll need AMCap utility to set up the camera and adjust its settings appropriately http://noeld.com/programs.asp?cat=video
Note: other cameras might work as well, but unrecommended.

2) 10m USB extension cable, e.g. http://www.amazon.com/Meters-Active-Ext ... 002VG39BI/
The cable is needed to set up the camera next to the target.

3) Constant-on laser pointer (<5mw), e.g. http://www.executivegiftshoppe.com/ls11-sw.html
or http://bestofferbuy.com/LXGD-Gun-Kit-Re ... 13765.html

4) A tripod, or other means to mount the camera. It should be located perpendicularly to the target, 20-50cm from it.

5) 0.5" gaffer tape to attach the camera to the tripod (camera does not have a tripod mount), and also to attach the laser to the pistol.
(As a side note, gaffer tape is awesome)

6) You'll need a relatively modern and/or fast Windows XP/7 PC (e.g. Atom-based netbook is too slow)

7) A built-in or external microphone (webcam's microphone should not be used).

8) Patience

Note: beta-testing means providing feedback.

silentfury214 · Post by **silentfury214** » Wed Jan 12, 2011 9:21 pm

Is it only for pistol?

AAlex · Post by **AAlex** » Wed Jan 12, 2011 9:24 pm

Theoretically, should work for rifle too, but I don't have the means to test myself.

silentfury214 · Post by **silentfury214** » Wed Jan 12, 2011 9:30 pm

I shoot sporter class for air rifle, but can still shoot pretty well. Would you need something like that for testing?

Brian M · Post by **Brian M** » Wed Jan 12, 2011 9:53 pm

I'd be interested in helping out (15 years IT experience, though more in the project/planning side of things plus I've used RIKA)... money is tight though, and even $60 is pushing it.

Is that 2nd laser link the one you have on the pistol in the first set of photos?

Billy33 · Post by **Billy33** » Thu Jan 13, 2011 12:25 am

Alex, looks great so far. Are you aware that the PSEye has a 320*240@120Hz mode? Would that give you more trace detail, or does the drop in rez affect accuracy too much?

Soupy44 · Post by **Soupy44** » Thu Jan 13, 2011 12:58 am

How accurately have you been able to measure the location of the laser at 10m?

AAlex · Post by **AAlex** » Thu Jan 13, 2011 7:45 pm

silentfury214: sure! if you're up to it - why not.

BrianM: understood; let me know when/if you're up to it. The laser in the pic is LaserLyte FSL-3. It is tiny, but the battery life is abysmal. The laser I referred to in the earlier post is 80x20mm in size, not heavy. I would prefer something that is more compact than that, but still use a real battery (AA/AAA/CR123A, etc), not button cells; however, I was unable to find such product.

Billy33: Yes, I'm aware of 120fps mode, and I think 320x240@120 would work a bit better than 640x480@60. However, since the camera is unbuffered, it is not trivial to take advantage of that mode as by default Windows does scheduling at 15ms time slots, which results in FPS cap of 66fps regardless of the camera's capabilities. I'll possibly deal with this in the future.

Soupy44: That's a good question, except keep in mind that the distance that determines this is the distance between the camera and the target (i.e. 20-30cm for PS3Eye, or 7m with my other camera that I had placed next to the computer). The distance from the shooter to the target does not affect the precision of the acquisition.

As to what the actual spatial resolution is - I didn't measure it after I verified that the spatial resolution is better than 0.1 target-points. In the sample screengrab the field of view of the camera is 300mm horizontally, and that gives 0.5mm/pixel resolution off the bat. However, in actuality the spatial resolution is much better than that, as the brightness of the individual pixels comprising the beam shape is considered. Reducing the field of view of the camera by half by placing it closer to the target will double the spatial resolution.

A more general question is: is the spatial and temporal resolution sufficient for the task at hand? Is capturing 320x240@120 better than 640x480@60? How about 1024x768@10? Intuitively, say, if you want to capture a trajectory of a fly that keeps changing direction really fast, and you sample at 0.5s intervals, if you connect resulting sample points it will not give a good representation of the trajectory, and you'd be better off sacrificing spatial accuracy and instead capture at a faster rate. On the other hand, if you capture a trajectory of a tiny slow-moving bug, you might want to prefer higher spatial resolution over temporal.

Determining whether one spatio-temporal sampling combination is better than the other and if it is good enough is non trivial and requires local-convexity-defect-analysis of the trajectory. Intuitively, if you sample a squiggly line that crosses itself many times using sparsely-spaced points, the resulting poly-line will have many "jaggies", indicating undersampling. On the other hand, if you sample a curve densely, the poly-line will look smooth.

Eyeballing the data, the cameras I used are "good enough". It looks like it will be difficult to predict the precise location of the bullet hole, so you can't use it as virtual electronic target, but that's a limitation of all current trainer systems. However, even with the consumer-grade camera the performance is sufficient to have a faithful representation of one's shot process and errors.

silentfury214 · Post by **silentfury214** » Thu Jan 13, 2011 10:51 pm

Going over your computer requirements, I was wondering if the Dell Inspiron 6000 would work. It is not the newest laptop in the world, but I just have a Mac. Also, if that doesn't work, would it be able to run off of Windows from my Mac?

Soupy44 · Post by **Soupy44** » Fri Jan 14, 2011 3:19 pm

The reason I ask is because air pistol targets are much larger than rifle. Tenths for air rifle are .4mm. I'm about to start practicing air pistol a bit but rifle is my main discipline, so I'll vouch for it and see what you system can do for rifle.

Billy33 · Post by **Billy33** » Fri Jan 14, 2011 6:04 pm

Alex: Sounds like 0.5mm accuracy should be fine (IIRC that's the thickness of the scoring lines on the target). Plus 120Hz would prob increase CPU reqs even more. One option might be to switch to the red "zoom" setting with narrower field of view as opposed to the blue setting?

I would be happy to beta test, tho I would be limited to 7m distance (more work for you!). Just a standard microphone hooked up at the firing line to hear the trigger click?