Emma is JooVuu’s product sourcer, helper, and all round amazing person. She lives in Shenzhen, China, and has a network of manufacturers and contacts that most people can only dream of. She has worked at JooVuu for a year and a half and in that time has become instrumental to the running of JooVuu. JooVuu would not be where it is without her, she has been a real driving force behind the JooVuu X, liaising with manufacturers, designers, engineers, support, and factories, sometimes all at the same time, whilst also pushing the JooVuu vision forward and reigning in my frustration and impatience. She has been a guiding force in the creation and implementation of JooVuu X. As she has been so instrumental with the JooVuu X I thought it is about time she showed what the process has been like from someone who has been ‘in the thick of it’. I have NOT changed any meaning or sentiment in this piece but I have corrected some spelling and grammar.
I’ll separately describe the three major different times, and the different accidents that have occurred and are occurring.
It has been a very long story for the whole developing process and was hard from the first actual task of producing a camera – finding a manufacturer.
1) Source a good supplier.
We have very limited finance and firstly want to trial produce a small quantity. I asked many factories who had experience of developing Ambarella products such as Dome, Iro, Apical, Koonlung, Matego, Chelong, Skylight , TCL, DWCO, SJCAM, plus a lot more. All of them felt the quantity we wanted was too small. Most manufacturers will not earn any money from the developing fee but instead from the unit price per camera, and as such want to see the camera sell very well.
Most of them felt that the developing duration period is extremely long, and the request we put forward to are hard to meet with.
We have spent long time to confirm a prototype fee, but get to know,have to give developing fee, then offer 10 to 20pcs prototypes. We have also performed a lot of investigation about using Ambarella A9, Ambarella A12, we want a camera with both action camera and dash camera. We were originally looking at developing an A9 camera but we finally found that A9 is exclusive for Gopro – nearly,we could pay an entry fee of 200,000RMB to get basic Ambarella support but no factory really knows A9 so it is a huge risk.
Another important reason is Xiao Mi action camera was released at around 50USD that it has crashed the whole market that very few factories are willing to create another action cam for fear they cannot compete. They felt that our quantity for our project is very small but Xiao Mi action camera their first trial production order qty can reach 100,000 – 1,000,000.
The divided cost for Joovuu to develop a camera can not compete with Xiao Mi. As such most factories do not have the confidence to accept the project of JooVuu X.
Then how about adopting Ambarella A12 chipset? We also think about it, but unfortunately, Ambarella A12 would be released until Oct 2015 or so. And only several big factory such as Dome, have the foundation of developing Ambarella chipset dare to try A12.
For IRO,they think they ever before had invested a lot of time,energy to develop A7 product , since they are the first batch of A7 chip that they solved those bugs, it just helped others to get direct experience, IRO refused the ODM project.
So they would not like to spend the time on developing a camera looks like no hope and not competitive product.
Well,we have some other similar supplier directly refused our ODM request,because they think quantity has to be 5000pcs above,or else,not worthy for them to spend the time when engineers workers are very limited.
Finally,after compare, analyse,we chose our factory who we will call Factory X, they can give us the small quantity JooVuu can afford to start the ODM project and sounds like with great confidence in the developing ability.
We finally settled on a supplier. Now we have to fine tune the specification.
2) What kind of spec do we finally can adopt?
A. Hardware part:
After we settle down the selection of supplier,we set about making a list of what we wanted/needed. Do we need dual microSD slots or not, external microphone port or not, touch screen or not, WiFi or Bluetooth 4.0, lens can be screwed in and out, CPL or not, splash-proof or not, 2k or 4k video quality or not,IR enable it or not. The list was long. We had done a lot of research on DashCamTalk and other forums so we knew some features, but others it was much harder to settle on even though we thought we knew what we wanted.
We have the idea of one camera has two microSD cards,and each microSD card can stream and save the recorded videos. Another idea is Smart Battery/Super Capacitor is built in – and the battery stops charging and sending a current to the batter when full even if still being charged from the wall, this extends its life.
However, Ambarella had not released the SDK for copying video files, and our other function of wanting to stream video files from the card to a computer was not possible due to WiFi limitations. The battery problem we have implemented a similar design. However, we had to really now choose what our camera was, a car or action camera, as an action camera cannot run with a super capacitor and a car camera is less reliable with a battery. We decided after much time to build a car camera but make it so a battery or capacitor can be installed. This took a lot of time, effort, and communication to solve and took close to one month.
B. Firmware part:
We have spent very painful period in decision of Ambarella A7LS or A7LA75 ,as different chipset means different production direction.
A7LS focus on sports camera mainly,can support 720P 120FPS, but can not support GPS
A7LA75 can not support 720P 120FPS but can support GPS.
Finally,we have decided that we focus on the dash camera first,then chose A7LA75.
For a simple function such as loop recording, because our camera do not have screen that most of the time,we could only get the default video resolution video clips, for our DQA,it is also hard to test more thoroughly. We have solved this problem but are still trying to overcome the save setting function, we have found the firmware error so we now know how to fix, unfortunately it is now Autumn Festival so cannot be fixed until early October.
For example: In order to get our user have the best video quality,we have tried many proposals in developing this camera such as Night Time Detection. The camera detects it is dark and increases exposure and decreases the frame rate to let more light in to get better video quality. But go straight to the answer and we had to make it so the current night mode is changing according to the time not the light, only at time set by the user can increase EV value and frame rate.
C. APP part:
This is actually the most headache part, as we start from April,now nearly Oct, spend 6months already on the APP development and only just nearing completion.
We waited nearly a month for the initial build, another month to decide the menu option and design, then create each single feature, firmware update day by day. This is very slow, but we have learned a lot from this experience.
It could be a very overwhelming work load for simply implementing a feature such as ” loop recording” since our APP company actually did not have Ambarella camera developing expeirence. Factory X initially spent a lot of time having to collect the API and teach APP company how to create a command,how to send a command,how to save a command in different environments. It may cost a week time to simply add one option,so you could imagine how slowly could it be for us to get a valid APP feature.
But anyway,with the help of our developer and our app company’s strong sense of self learning, after several months’s experience of writing the build of APP, they have been become more and more familar with add a feature and debug a feature as they can know how to do another option by learning from other examples
Now we are overcoming the bugs of APP such as WIFI disconnection, camera unresponsive after certain commands, sometimes stream can see full black or sometimes have green words,etc.
We are still struggling to solve the bugs,but we are basically close to finish with only a few bugs to iron out. No bugs are hardware related which is very good news. We will make the camera better and better as we have absorbed many failures and now on the road to success.