CameraWrapper TDesC8* vs. CFbsBitmap*

[pdcb_pdcb]

Hi all,
this should be a fairly straightforward question but i couldn't find anything that clarifies this for me. I am trying to fully understand the CameraWrapper example. Taking a picture in the default way, i.e. preparing the camera as in:

TRAPD(err,iCameraWrapper->PrepareL(iCaptureSize));

let's me grab pictures really quickly but then i think that i get a compressed jpeg (the data is TDesC8*). Alternatively, I can do:

TRAPD(err,iCameraWrapper->PrepareL(iCaptureSize,CCamera::EFormatFbsBitmapColor16M));

In this way i get a CFbsBitmap but it's much slower (~1 second for 1024x768), probably because the image is much larger.


I need it to be fast so the second solution does not really work for me, but then i need to access the pixels of the image to do some image processing. I couldn't find a way of converting the compressed TDesC8* to uncompressed CFbsBitmap* or TUInt8*.

Is my only option that of saving the compressed image and then reload it as CFbsBitmap? Also, other formats such as EFormatYUV422, don't appear to be supported.

Does anybody have any idea?
Thanks so much!!!
pdcb

[chenziteng]

Hi,

You can use CImageDecoder:ataNewL() to create a CImageDecoder object and then use it to convert the jpeg data to a CFbsBitmap object, but I'm afraid the jpeg decoding is not faster than asking the CCamera for a bitmap object. (I guess the CCamera will do the decoding internally if you ask for bitmap object, and that is why it is so slow).

By the way you can use the following code to check the image formats supported by the camera.

...
iCamera->CameraInfo(info); // assume that iCamera is a valid CCamera object
info.iImageFormatsSupported; // a combination of several CCamera::TFormat values

Regards

Ziteng Chen

[pdcb_pdcb]

Thanks a lot Ziteng Chen,

So I did understand correctly: if you keep the compressed version there is no way i can actually access the pixels, right?

Thanks a lot for the other suggestion as well!
pdcb

[chenziteng]

Thanks a lot Ziteng Chen,

So I did understand correctly: if you keep the compressed version there is no way i can actually access the pixels, right?

Thanks a lot for the other suggestion as well!
pdcb

Hi,

I can not find any Symbian OS/S60 API to access the pixel values of jpeg data, but technically I think it is possible to do so. May be you should google for more information as this is a general image processing problem.

Regards

Ziteng Chen

[tamhanna]

Hi,

I can not find any Symbian OS/S60 API to access the pixel values of jpeg data, but technically I think it is possible to do so. May be you should google for more information as this is a general image processing problem.

Regards

Ziteng Chen

Decompress it to memory (to get a bitmap). Then access the pixel data of the bitmap.

JPG is like archive, needs to be decompressed to access!

[chenziteng]

Decompress it to memory (to get a bitmap). Then access the pixel data of the bitmap.

JPG is like archive, needs to be decompressed to access!

Hi,

I know jpeg data are compressed but I think technically it is still possible to get its pixel values.

"JPEG Pixel Interpolator"
http://www.zero-based.org/software/jpegpixi/

Regards

Ziteng Chen

[torsak1]

I am also interested in doing some image processing from a Tdesc8* structure. Did you find out a way to do that?

[wizard_hu_]

Then you will have to use CImageDecoder - you can find examples/documentation everywhere (Wiki, Docs section above, SDK Help).