New solution for people wanting a DIY HD camera from Micron and Ambarella
 

Hi all,

Las night I was searching for new possible DIY HD cam methods (Farhad, don't take it personal... It's just that I can't stop) and found a very interesting piece of hardware made by Micron (now Aptina) and Ambarella.

It's called DEVcam and it's basically a Do-It-Yourself Sanyo Xacti system with a Micron sensor. A whole working system powered by just two AA batteries with FullHD@30fps (doesn't say "progresive", so I guess it's interlaced), 720p60 and 480p60 capture mode including sound recording. It all comes in a single board and according to Aptina, it's quite cheap (below $700). It also comes with a 2" lcd and it's ready to work right out of the box. It records video and audio to SD cards using H.264 codec.

Pros:

- It's a fully working system.
- It doesn't need a computer to work.
- It records to sd cards.
- As it comes without any external case, it can be used for our personal projects.
- We all know Micron sensors can give very good quality and superb dynamic range (for those of you who don't know, just take a look at my first DIY HD cam thread), so even if compression is similar to other hybrid HD cams, image quality should be better.

Cons:

- I think 1080 mode is interlaced. I mailed Aptina asking for more info.
- FullHD mode is 1440x1080 real pixels. I don't know if those are simply 4:3 or 16:9 with non-square pixels.
- H.264 means heavy compression. Similar to the HD2 cam from Sanyo.
- I just asked but I think 1080@30fps and 720@60fps are fixed framerates. No possible 24fps mode.
- I also asked if there's a full manual mode for exposure, white balance...

Besides all that I think this board can help us develop a very nice cam for non-professional work. I've seen Xactis attached to homemade 35mm adaptors (and even Redrock M2s!) and they offer a very interesting option for people who want 35mm DOF for home videos and basic amateur short films.

This developing kit would hopefully take a step further offering better image quality and dynamic range than a Xacti cam.

As for me, I'm going to take everything I bought for the other DIY HD cam project, buy one of these boards and put everything inside a box with an EOS mount. We'll see what happens.

I leave a PDF flyer attached. For those of you who want to know more, just google "DEVcam HD" and you'll find lots of pages.

BTW, if you are sending them an email, please let them know that it would be quite useful if they opened access to datasheets for the different sensors they offer.There used to be a time when they were open, and so Elphel was using the sensors.Now it seems that just for reading a datasheet you need to be from CIA or something.If someone is going to steal their secrets, I guess they won't do it from a specs sheet.

What is the bitrate of the video? h264 doesn't mean it will be bad, bitrate does matter.

It could have external shutter, an mechanical one ...great
what does it out SDI ?????????

The Ambarella A2 Fact Sheet says it can support bitrates up to 40mbps.

I believe it has an HDMI out. I'm still waiting for an answer from Aptina.

Dear Jose
sounds to good to be true in any case could be a poor indie cam
a decent DOF Adapter, mechanical shutter, look to the other tread
a magma express Box a card in there and an apple laptop voila

well we are not there but beside Si Sumix is not 2 K and who cares

Please, have in mind that I'm presenting this product for people who want an already made, pc independent HD solution. This is not a Sumix cam or something you can make serious projects with, but as it just needs two AA batteries and a SD card, if you include a homemade DOF adaptor, you can have good quality HD home videos with a cinematic feel. Well, maybe even more. Yesterday I saw a couple of music clips made with a Xacti HD1000 and a Redrockmicro M2 and I just couldn't believe they were made with a hybrid cheap cam.

Anyway I'd love to find a FullHD progressive H264 hybrid cam. So far I didn't find a single one. Not even this DEVcam kit. Here's a list with the newest ones. Many of them say "1080p" but then you visit their site and they're all 1080i. The first brand to release a truly progressive full 1920x1080 H264 cam with a decent bitrate is going to sell it like hotdogs.

http://hiperdef.com/tag/videocamara/ (sorry it's in spanish).

So
serious project means
High tech overkill
Quad CPU
half million Pickup
and test shots but no movie bah NON

problem is
this board board is
with a tiny Cmos form factor
for cheap lenses.
The problem starts there

great electronics, lenses ...scrap

such a board with a 2 K Cmos will be heaven
if it will sell big numbers
I don't believe
it is like loudspeaker DIY a niche
where you can talk and meet great people

Hugs
Régine

Jose, this will be a toy, 1/4.5" sensor with 39 dB S/N. Economical wise it is waste of operator value if you provide him a tool that costs much less than his time.
Also, we can never compete with Sony in low cost consumer HD Video cameras. There is so much that goes into optimization and production efficiencies. A high performance camera has more than 50% of its cost in it precision case. How would you like to put a cheap camera in an expensive case and offer it to an even more expensive operator to use it. Economy must be economical and this idea is not.
Regine is right on the money, let's make movies instead of working on producing cheaper cameras. I have many stories from past years that people tried to make new modalities instead of using tools and flexibilities they already had, although at higher costs, but ready and useful.
In Sumix we work on technologies that can prevail in test of time. SMX-12A2C is based on well thought architecture. It has flexibility that people can exploit and make their work look unique and even better than expensive cameras that come with built in processing. Price is already low comparing to human expenses of film making even considering another $10k in lenses , computers, fixtures and batteries.

After saying all these things, let me admit that we are considering the same sensor MT9M032 B&W for our 3x3 3D camera. But this camera will have 12 megapixel effective pixels, No Bayer (avoiding aliasing) as fast as hell >60 fps , depth of focus will be much better than 35mm camera. There will be no need for focusing while taking the clips. Focusing is done in post, so less skillful operator can operate the camera and cost of camera and operator will match better. Perhaps you will need a much more expensive and imaginative director as possibilities will multiply.

Correction, 3D camera with 3x3 sensors, total of 9.

Farhad

Hi Farhad,

I think there's been a misunderstanding here. I presented this kit just for people who want a cheap and easy to use option to shoot very basic HD videos. I don't think it can be better than any of the many cheap HD hybrid cams out there (Xacti, Aiptek...) and of course I wouldn't recommend it for serious projects.

Just wanted to make it clear.

Règine,

Yes, this cam cannot be used for professional projects and the 12A2C is not yet ready to give its best if you want to shoot for real, but it will be in about a month, so please be patient. Sumix is working to make it happen. You'll just need the cam, a decent laptop and your audio equipment to shoot whatever you want using the Sumix software so it won't be difficult or messy at all.

I'm sorry if anyone took this thread and its contents as a possible replacement for the Sumix cam. It's quite clear it's not.

Jose, I can not thank you enough for your enthusiasm and your quick ideas. You just gave us a chance to rehash where we are going. I am sure soon there will HD webcams under $100. The goal is not commodity cameras, but to have HD pro cinema cameras available to independent film makers. A $10k system may be rented like $200 a week or less. On the other side, according to marketing 101, your product must be unique. A film maker must have ability to do his own post processing from raw data and produce his own signature work. This is where Red and Si are going wrong. They are spending their resources on post processing that must eventually be under the control of the film makers. We try to produce tools that you will understand, there will be no magic and no surprises. You find third party tools and you make your own and share. Future more capable cameras will use the same tools.
Again, Jose, thanks for inspiring us. Your independence and that of others on DVinfo will keep our communications responsible and our goals easier to achieve. Without responsible and unbiased communications, independent film makers and Sumix together will be food in this jungle.

I don't understand... You say film makers must have total control over raw datas, which Red offer, but at the same time you say they are wrong ?

What is the compression ratio of Red Raw?