I always felt that the ball head I own while good is not that strong to properly hold my camera (1D) and a long lens (70-200). I was using a PhotoClam PC33.
I had itchy back side and end up buying a Markins Q10 with QuickShoe (QL60). Not cheap, not small but once you use it a bit, feels great. Managed to take some nice panning pictures night time somewhere in Clark Quay.
Looks bigger then in the picture, sure need to see and hold it to get the proper feeling, but the built quality and the details are great. Knobs turning smooth, big ball, ball get fixed in the position easy and there is virtually no camera shake, at least not from the ball head. The only thing I do not really like is the panning knob, while easy to lock, I do not find it that locks enough the ball head.
Tried to use it with L plates for 1D and 5D and works like a charm: very fast to lock and once locked is very sturdy. The QuickShoe clamp is great!, already considering to buy one for my PhotoClam ball head…
The spirit lever is a bit… funny, the position is pretty awkward as can not be seen once you attach any plate/camera, but I can live without it.
Overall, so far seems like a good investment, it does what is supposed to do and I do not need to break my piggy bank to buy a RRS ball head anymore. I just have less and less excuses not to go out and shoot.
For past two weeks my TimeCapsule started to slowly go to coma… Basically keep reseting itself, getting every time a new external IP address. I was mostly offline, could not use my internet connection, pretty much separated from the world… Just kidding, most of the time I’m using my iPad on 3G…
Anyway, after many thoughts of buying another TimCapsule, I decided that investing such a large amount of cash and in the end just getting basic network is not worth. I settle for the “new kid on the block”, the so much acclaimed ASUS RT-N56U, which sadly is still much cheaper than AirportExtreme option.
Just managed to set it up and is really working pretty well. Hopefully now I’ll have no excuses and start to post more too ^_^
Last Friday, I was clearing leave and cause of itchy backside, I started looking online for upgrading my bike. I was tempted to buy a new frame and build a new bike and sell mine, however, I realized that this is going to cost me 3-4 k$ at least… A bit out of my budget. So, after a previous visit to the bike shop next to my place, I decided to take a look on the Togoparts website for a second-hand option. I was tempted very much to go for a Cervelo P2 or even an aluminum Soloist from 2006, but after another visit to the bike shop, I end up exchanging my Giant TCR Team Alliance 2009 with a Scott CR1 2007 frame with DuraAce group set, used less than 1000 km, looking like new, if ignoring some scratches on the shifters and cranks cause by few falls most likely.
Overall, leaving aside a bit of rust on some bolts on the stem, looks pretty decent, full carbon frame and seatpost, coming with Shimano carbon clincher wheels (RS80), full DuraAce 7700 groupset, everything bellow 8 Kg. I changed back the wheels to my Neutron Campagnolo wheels, cause i like then so much and feel them like rolling better.
It feels quite stiff, a bit stiffer then my Giant, lighter, but still not used with the DuraAce shifters, which comparing with my previous 105 groupset, change the gears a bit smoother, but feel harder to press the lever for changing gears, however, it should not take long to get used to it.
Went out for 40 Km ride this morning, however, was quite windy on the east coast and also my legs still recovering from my previous weekend, not yet able to estimate a difference, anyway, this coming weekend will go with the guys for another “Reverse Round Island” ride, somewhere around 130-140 Km ride, should be able to tell a difference by the time will finish the ride.
Last weekend I went for the my first “Century Ride”, the Kuantan Century Ride. Actually a 164 Km on Malaysian roads, around 40-60 through palm tree plantations, another 60 Km on highways and federal roads, another 10 Km by the sea and the rest through cities and villages… Was a hot sunny day, with just a little bit of shades and more than 60 Km of rolling hills…
I managed to finish, but not very happy with the result. While i passed the first 3 checkpoints with more than 1.5 hours before the cut-off time, I managed to finish just before the cut-off time… The main reason been the hills and the lack of training, I admit. Anyway, is done, and while many gave up I managed to finish! Already looking forward for the Ipoh one!
Not very happy with the performance I get from Processing in displaying the images coming from my Arduino Module I decided to build a dedicated application to “see” the image instead. Is part of a bigger project, my “stereoscopic vision” project ^_^ Kind of a long shot, but sounds like a lot of fun.
Anyway, I start learning VB and for the past 2 days I tried to code a small application that connects to the serial port, retrieve the data streamed by the camera and display it at high speed (hmm… sort of anyway) in a window.
First, I am planning to use 8 bit per pixel, however, not happy with the 256 color per pixel option, looking more for a 4096 colors per pixel instead, so I decided to get pump 12 bits per pixel in only 8 bits… Cool, right? The trick is simple: for each pixel I stream 8 bits, each bit contain 4 bits for green (G) and 4 bits for blue (B) and next one 4 bits for green and 4 bits for red (R) so 2 adjacent pixels share the same B or R information at 12 bits per pixels, voila! got 50% compression of the data sent while getting 16 times more colors and images looks much much better.
Next step is to understand how the VB 2010 works and how to handle the serial port and then… how to draw pixels… Last was the worse part, basically could not find much about it and when I could I realized that was damn simple…
Anyway, while is not yet extremely stable and fast (320 x 240 = 76,800 bytes + some synchronization data, around 80+ Kbytes of data over serial port at 600,000 Bps take a bit less than 2 seconds per frame) is kind of working, however need to work a bit on the speed and to detect the errors and drop the lines that are not received properly.
I am actually going Botanic Garden quite often for past few years. I am usually to lazy to bring a camera along and when I do, I am too lazy to upload them… Now is one of these moments when I managed to finally do this.
By accident I ordered the LPC1114 instead of LPC1343 chips! Cannot believe it! I was lucky that I ordered at least the LQFP48 model and not the 33 pin ones…
Anyway, soldering them to the board is not easy at all! Only 0.5 mm between pins, just to align them before start to solder and feels like impossible. Anyway, managed to get them soldered and I did it without any short circuit, pretty clean I can say… Now I have to study on minimal configuration to get it programmed, guess should not be too hard.
Last night managed to spend some time to redo my panning code and got it moving much better. Still work to be done but is not too bad.
Then got a CF7670C-V1 module work with this setup, also tracking at high speed using the same piece of code, by doing a short circuit between pin5 and pin6 on the module. This way will behave like a V2 one, however here is no HREF signal going out, not used in my application.
Next will be an attempt to get it running with LPC1343 chip, cause this microcontroller is very cheap and can cut down the BOM quite a lot in case I will decide to make a commercial version for education purposes. Also the extra RAM can help in tracking multiple blobs of same or different color.
So far the tests to attach the OV7670 camera to a robotic platform seems to work. My “arduino camera” module (not sure how to call it since “AVRCam” and “ArduCam” are already used by other public projects) seems to work pretty fine attached to a small tracked robot, same base with the one I’ve use to connect my 4D uCAM module. However, the speed difference is quite impressive: from 3 FPS to 30 FPS and make it move smooth this time.
Anyway, got the “tilt” handled pretty fine with minimal oscillations, some tuning may improve even more the accuracy, and the “panning” almost working, cause is not that smooth, really need to work on my algorithm. Once this done, I will take a better video and upload it.
Looks really interesting… Using CF7670C-V2 module I can go much faster on my board… Is what I see is correct can basically track one color at 30 FPS… The movement looks so smooth on the Processing screen I use for displaying the data in real time. The good part: is a way to make a V1 module work as a V2 module, but this require a hardware change: shortcircuit between VSYNC and WRST, fast and dirty… ^_^
