Old Man Mike

After seeing the Blackdragan video in an earlier forum thread, I've been trying hard to come up with a camera mount that matches his performance. Although Blackdragan chose not to the reveal mount or even the camera, I will be providing the details of my approach. But first, I would like some feedback to see if the group thinks I've come close to his level of performance. The following is a 30mb unprocessed section of the 4 min video (the full length size is over 400mb!). This is in the camera's native format (MOV) which will play on the Quicktime player that I hope most of you already have linked to IE. I fly without TI enabled and I admit that my flying is not as good as Blackdragan's since I tend to rock the DF to keep my bearings. I look forward to any feedback.

[link=http://www.af9y.com/TFMod3.mov]Custom2f[/link]


Thanks,

Mike

BB_DF

Hey Mike,
You're really on the right track! The video doesn't have any jitter or bobble - it just follows the movement of the bird and you will always have good sections where you're straight and level. I'm sure everyone would love to see how you did it, but if you want to keep it close to the vest, that's understandable.

The thing about the DF is that it's light and pretty unstable in any kind of disturbance. You can tell because it's an absolute joy to fly when it's still, but a struggle in a gusty wind. To demonstrate the other end of the spectrum, here's an electric helicopter with a 6 foot rotor, a motor the size of your fist, and a battery the size of your bank account. German, of course. It is so smooth, it doesn't need any kind of camera dampening at all. This guy's got more guts than me flying that thing over all that water!

http://www.youtube.com/watch?v=yPLTroBd2YE

Cheers,
Bruce

Sky High

Old Man Mike that's some nice looking video! I can't see why Ti wouldn't work well from what I've seen of the layout of your property. The pitching would be significantly diminished using Ti and make your video even better. I'm convinced that onboard recording is the only way to maintain resolution and color representation. The downlink would still be useful, though not necessary, for framing.

BB_DF [link=http://www.helifreak.com/viewtopic.php?t=38607]here[/link] is video of a Bergen turbine flying out over water. [X(]

Old Man Mike

Here is the construction details as text associated with each of the pictures:

1) A new baseplate was constructed out of thin aluminum. The large hole is to allow access to the camera controls. Three rubberbands are used on each side to hold the camera mount. Three are used so that the mount is held somewhat loose to allow movement that will be shown later.

2) The primary camera housing is carved out of a toy foam stick called Plantet FOAM. The Pansonic DMC-FX07 camera slides into the cut slot. Additional holes provide for the lens, top controls access and back LCD viewing. The extra glued piece is to keep the assembly from tipping over when resting on the floor.

3) This is the back side of the foam housing showing a portion of the LCD. The white plastic piece at the top was an attempt to control the camera shutter button. (This did not work and I ended up using a rubber band around the camera when taking continuous still shots)

4) Camera lens extends within the protected housing when the camera is turned on.

5) Here is the MOST important aspect of the mount. I use a thin sheet of plastic that sits between the foam camera housing and another lay of softer foam. The thin plastic sheet allows the softer foam above the camera to pivot back and forth so that yaw movement is dampened for the camera. This was the big breakthrough for reducing most of the shaking effect seen in earlier mounting attempts. The soft black foam is from a special sound reduction foam and it has a thin foil on one side. I keep the foil side toward the plastic. I believe the soft foam slides just as well without the foil but have not done enough detail testing to be sure.

6)Assembly of the Foam Camera housing, plastic sheet and soft black foam.

7)Whole system mounted on the DF

8)Construction material used. I used the electric foam cutter shown which really makes the job easy. A substitute for the back foam that I used is also shown. Preliminary tests showed that it worked about as well as the black foam.

Whew! Lots of typing there but I hope everything is clear enough. I especially hope that this serves as a starting point for even more discussion of better approaches to vibration free video. The total weight of the Camera and Mount was 7 oz. The mount was about 1.5 oz out of that total weight.

Old Man Mike

The new mount also makes a big difference in taking still shots. Three samples are show below and here is a link showing even more detail for one of them: [link=http://www.af9y.com/P1030664crop16x12.jpg]Higher Resolution Sample[/link]

A few notes about taking stills with this camera:

1) Setting ISO to 400 seems to give the best results since the camera is forced to use a faster shutter. These shots were at F/8 and 1/400 second shutter.

2) The jpeg process (even at the highest quality) within the camera now appears to be the limiting function in picture quality. Still, the pictures appear quite good for such a small camera and lens.

3) Antishake Mode1 was enabled but I'm not sure that it is helping much. I need to do more testing with both Mode 1 and Mode 2 antishake compared to no antishake.

4) The camera is shooting in continuous burst mode which produces around 400 6megpixel shots for a typical flight.


Although the DF flys a little heavy with this package, it certainly is manageable. I'm getting quite comfortable at flying it above 100 feet now.

NEWS FLASH for Sky High!!!:

I flew the camera with TI ON this evening! It was getting too dark to get good video but WOW what a difference! Both LEDs were lit during the arming but unlike previous attempts, this time I gunned the motors and tried to do the trimming in the air. I guess the whole problem is that I expected the trim to be about the same as with TI OFF. For some reason my DF needs a heavy trim to push the nose down. This is true with or without carrying the extra weight of the camera. Once set, it flys wonderful and appeared to have no problem with the extra camera weight. I could almost set the remote control down on the ground (not stupid enough to try) and let it just hang in space. Amazing stability and I am really excited getting some video while flying in this mode. I am puzzled why so much trim (almost 100) is required in TI mode. It also feels like I am flying it in training mode even though the active switch is up. I can now see how you can be comfortable flying at the 300 foot level when in TI mode.

Mike

Sky High

It just gets better and better! It's not necessarily your DF that needs that much nose trim. I think it's the location. I have had huge differences in calibration with two different locations 50 feet apart and miles apart with 5 minutes and 15 minutes of changing locations. In other words not necessarily different days. I would go to an open field like a pasture or football field and see if it still needs that much nose trim. I quit trying to figure out why along time ago.

BB_DF

Jeez. You just wouldn't believe us, would you?

Just to add a little to what Sky said. Don't worry about the trim - just do whatever you need. What happens is when you set the level (Throttle Down/Right) on the ground, it's almost impossible for the Ti to not be affected by something nearby. For instance, the afternoon sun could heat up a wall quite some distance away, and the sensors will pick it up easily, and skew the horizon. To test this theory, just set the thing down and set the level again in the same place on the ground, but facing in the opposite direction. You'll probably need just as much trim in the other direction after that.

I'd suggest you change the Expo in the Normal mode to 20%. I think you'll like it a lot better in Ti Mode. Some people fly it with no Expo in Ti. I fly 20% with Ti on and 30% with it off. It also comes with D/R set to 50%, 50%, 80% for training mode, and I set them all to 100%.

Good goin', Mike!

Sky High

I had an interesting time with extreme trim when shooting [link=http://www.rcuniverse.com/forum/fb.asp?m=4191489]this[/link] AP video. I got good Ti in the center of this huge concave field which as you will see was in the shade because the sun had just sunk below the treeline. So, when I calibrated Ti and thermal level in the center of the field which was in the shade, after climbing out into the sun the DF suddenly leaned way back like a stallion raring up. I quickly reacted by inputing full forward stick and eventually had to apply full forward trim because if I had let off the stick too much too fast it would have lunge backwards. I've said it before and I'll say it again. Ti is a great lowcost solution to help fly this machine but it's just too conditional and unpredictable for consistent results.

BB_DF

That's nice clean video, though. Did you run it through Adobe Premier? I think there might be something wrong with my hardware. Does Spectrolutions work on the video circuits, too?

- BB

BB_DF

I usually pop it up to about 10', and trim it out, then up to 20' and trim, then up to 50-75' and trim. I try to get it to hover pretty well at that altitude before flying off to do whatever.

Self-leveling is a tough problem to solve. Horizon-sensing is probably the simplest, and therefore maybe the best overall scheme. However GPS/barometer/magnetometer works better near the ground. [link=http://www.fmadirect.com/detail.htm?item=1489&section=20]HERE[/link] is a product that uses the same patent design as the DF Ti, that you can put on other models.

-BB

Sky High

That IS the Ti circuit. FMA licenses it to DFI & Spectrolutions.

Old Man Mike

Sky High: That was a nice smooth video. How high were you at the highest lpointl?

By the way, the little Panasonic camera I've been using also has the a real time video connector that could feed the SAVS transmitter while it is recording the video. Also, power could be supplied via another connector which would elimination then need for the Camera battery and reduce the system weight by almost 1 oz. The only problem is finding the connectors to mate. I've checked Digikey and several other houses with no luck yet. I may have to buy the external power supply unit just for the connector.

I still think there could be improvements which would provide additional suppression of vibrations to the camera. If the light level is low enough, the vibration effects are still visible. Maybe a foam ball inside a plastic ball with the camera foam enclosure hanging at the end. I'm going to invite an aerospace mechanical engineer over to see the system in the next few weeks and hope to get some advice from him on alternate approaches.

Let's hope there is little to no wind tomorrow. I'm ready to do some TI flying.

Mike

Sky High

I'd say at the most it was about 125-150 ft.


Now that reminds me of something but I can't quite remember what it is ...?

Old Man Mike

I was able to gain additional improvements by going to a suspension approach using 1/8" soft stretch elastic instead of rubberbands. This approach allowed me to return to the original carbon fiber base since there is enough room to access the top camera controls as it hangs below the base.

After a few experiments, I found the best length of elastic was around 2 inches. Too short results in not enough motor vibration reduction. Too long results in large camera swings as the DF moves around.

To help keep the swaying down, two pieces of soft foam were glued to the top of the harder foam camera mount. This foam is just tall enough to dampen any rocking motion but is not tall enough to have any compression in the level position (this keeps down the conduction of motor vibration). In addition, the two added foam pieces provide a resting surface for the base place when the DF is not airborne.

Two aluminum 1/8" dia tubes were passed thru the foam camera mount and serve as the attachment points for the elastic. Each of the four elastic lines are tied at one end to a button that has been drilled with a single hole. The button slides over the ends of the aluminum tube and acts as a bearing during the suspension.

I think the single picture shows everything clear enough given the description above.

There are several nice things about this final (well maybe for a while) design:

1) The camera mount weight is now down to 1 oz. The combined added weight of Camera and mount is 6.6 oz (I remember when I thought 8 oz was good)

2) Inside testing is by far the best I have seen since I started this quest. Here is a short clip in MOV format: [link=http://af9y.com/TFMod4in.mov]DFMod4[/link]

3) The camera mount can be easily rotated 180 degrees by sliding off the buttons, rotating it and then sliding the buttons back on. This is quite nice when you want to keep all the flying in one direction (because of wind or skill) but need to get hovering shots in that opposite direction.

4) The large green hanging camera mount is much easier to see than anything else on the DF. It actually provides a third dimension of visibility which helps to maintain the visual perspective.

5) The whole camera mount can be attached or removed in seconds by simply sliding off the four buttons.

6) I think this foam camera mount can be adapted for almost any of the ultracompact cameras available today.

I will post the outside flying results as soon as we have a low wind day.

Mike

Old Man Mike

We had only a little wind today so I was able to test the new mount outside. This was shot with TI enabled and as you can see, it has certainly reduced my rocking motion. It is so easy to fly in TI mode (after trimming) that I almost feel like I am cheating. I've been experimenting with different software for converting from MOV to AVI (or WMV) files. This conversion was done with Total Video Converter:

[link=http://www.af9y.com/DFMod4out794.avi]Converted Video (14 Mb)[/link]

Here is the original file if you have the patiences for the long download:

[link=http://www.af9y.com/DFMod4out794.mov]Original Video (69 Mb)[/link]

By the way, can anyone recommend a high-end video converter software program? I've tried about a dozen for the MOV to AVI or WMV conversion and none really impress me.

Mike

Sky High

Have you tried using Windows Movie Maker?

Old Man Mike

I've heard some good things about Movie Maker but I do not believe it does any conversion of MOV files.


Here is one more video. Unlike all my previous video, this one has some additional processing. I wanted to get as stable as possible before applying any post processing. I think this is about as good as it gets:

[link=http://www.af9y.com/Deshake798.avi]Smooth Video taken with Custom Mount[/link]

Mike

Sky High

Wow. Now that was very smooth and as always the picture quality was great! But why was it crooked most of the time? You and BB_DF are frustrating me to death with crooked shots.

Old Man Mike

Thanks! That happen to be the only video out of a half dozen that had any slant to it. I must have had one of the buttons not seated aginst the foam mount which resulted in a the angle hang.

(And what are you doing up? I thought I was the only one up at this hour of the day)

Mike

dart340

Hi Mike

Looking at your videos, your mount is pretty close to removing all vibrations. The only thing now is to fly the DF smoothly so the stabilization software doesn't get confused. I tried stabilizing one of the videos with Virtual Dub and it turned out pretty good. You can download it here:

http://members.shaw.ca/the.photograp...out794xvid.avi

You will need the XVID decoder to play it. You can downloade it from here:
http://www.koepi.org/XviD-1.1.2-01112006.exe

dart340

Thanks to Sky_high for pointing out that I can edit an existing post.

Sky High

Okay. I can't wait to see it when everything is secured. What I would really like to see is a flight lasting around 2 minutes just sitting in the sky in one location maintaining a solid level frame and then after that some longitudinal flights about 75-100 yards away. I'm wanting to know the endurance of the DF with the weight of that camera during a long flight with not much movement such as the hover and then with the forward distance flight.

My internal clock is upside down right now.

Old Man Mike

Thanks Dart,

Actually, I think you can do better than that with a different set of parameters setup for the Deshaker plug-in to Virtual Dub. Compare this processed video to what you posted:

[link=http://www.af9y.com/Deshake794.avi]Deshake794[/link]


Sky High: I will look through the videos taken so far to see if I have anything close to what you are asking for. I do know that I've had it hovering up and down at 100+ feet for about 8 mins and that was with the heavier early foam camera mount. I'm sure I don't have any longitudinal flights at that distance so will try that in the future.

Mike

dart340

Hi Mike,

Your version looks great, It looks like your settings are perfect for your videos. Would you mind telling me what your settings are on the deshaker plugin or post the *.vcf file that virtualdub saves when you /file/save processing settings. This would help me in understanding the program better and improving my videos.

Regards,

Tim

BB_DF

Wow, you guys are making great strides with the combination of cameras, mounts & processing. I'm impressed! Mike, your rig sure casts an unusual-looking shadow!