Where does M8's dual camera save the depth/distance data?

lachdanan

Well-known member
Aug 9, 2013
93
0
0
Visit site
Hi,

Is it possible to use the depth data saved by the dual camera outside the phone? For example in Photoshop, or in a compositing app like Nuke?

I find the one in the phone hard to use as it overwrites on the original image, etc. Also I could control the blur and apply different effects based on the depth data saved.

I checked the images outside the phone but they are jpg and couldn't find an extra channel or layer in them.

Any ideas?


Thanks.
 

Darth Mo

Well-known member
May 17, 2010
1,306
52
48
Visit site
The data is stored in the JPEG, but outside the normal address location. This allows you to do a straight transfer of the files that are still be readable by ordinary software.

Even if you managed to extract the data, it's probably in a proprietary format that would be useless in photo editing software.
 

lachdanan

Well-known member
Aug 9, 2013
93
0
0
Visit site
So do you mean the depth is saved as another image in another directory where the original jpeg just has a reference to? Do you know this location? I would still wanna check it out and could write some code to read it into Nuke for example. But it depends on the data.
 

Darth Mo

Well-known member
May 17, 2010
1,306
52
48
Visit site
No, I mean the data is stored somewhere within the original JPEG. The camera app on the phone knows where the data is an how to interpret it.

I have no idea where the data is stored in the JPEG, just that it is there. This is why the files are larger than normal.
 

lachdanan

Well-known member
Aug 9, 2013
93
0
0
Visit site
Thanks alot, I will look into where one can save data into a jpeg and how to retrieve it. I haven't done this but I imagine there would be a way to get all the extra data from a jpeg using an API and then see if I can make sense of the data.
 

kronosqq

Well-known member
Apr 19, 2012
666
21
0
Visit site
Thanks alot, I will look into where one can save data into a jpeg and how to retrieve it. I haven't done this but I imagine there would be a way to get all the extra data from a jpeg using an API and then see if I can make sense of the data.

Keep us updated, I'd be really interested to try that out if you find a way to do it.