ERROR: unsupported JPEG in cmd_loadimage()

[christian]

Hi,

we are trying to display a jpeg image on a display connected to the BT815. But we got

ERROR: unsupported JPEG in cmd_loadimage()

followed by multiple

ERROR: display list overflow

The error message is clear but we can't find a reason why the jpg is not supported. The data sheet says that progressive jpeg is not supported but our jpeg is not progressive.

I attached the unsupported jpeg. What do we need to change to display the jpeg?

[Rudolph]

I just tried to feed my BT816 this image and it will not display for me either.
Then I opened it in IrfanView, checked the available information for it and saved it under a different name again
using the same 90% setting that is in your image.
And this one works just fine with the BT816.
I had both files in the µC memory and sent them over the SPI to be unpacked into gfx-mem.

There must be some difference.
But I can not find it.
Except for the new file beeing a little smaller which could be because of re-compression and therefore a binary compare is useless, there does not seem to be any difference.
I even used the services of fotoforensics.com and imageforensic.org and these both show the same data.
JPEG image data, JFIF standard 1.01
Baseline DCT, Huffman coding
No Meta-Data

So, no idea what the problem is, maybe something is added to the raw data that should not be there and is not picked up by pc software or the web-tools.
But re-save it with a tool like IrfanView and it works.
And you can even batch-process with IrfanView.

[christian]

Hi Rudolph,

Thanks for your response.

We did the open and save again trick with Gimp too and we also tried some jpeg analysis like you did. But we can't find the difference. Unfortunately this jpeg is only one of an ongoing jpeg stream so we can't just save the jpeg again with an external tool to display it. We need to know what we have to change at the stream source.

[christian]

By analysing the binary my colleagues found that jpegs with one big quantization table won't work. If you open the jpeg and save it again with Gimp (or another image tool) the quantization table will be split up into multiple tables of smaller size. Can anybody confirm this?
   

[Rudolph]

At least https://29a.ch/photo-forensics/ shows a difference then between the original file and the one I saved with IrfanView.

Original structure:
    SOI
    APP0
    DQT
    SOF0 (Baseline DCT)
    DHT
    DRI
    SOS
    EOI

New structure:
    SOI
    APP0
    DQT
    DQT
    SOF0 (Baseline DCT)
    DHT
    DHT
    DHT
    DHT
    SOS
    EOI

So the new file does not only have two DQT entries but also four DHT entries.
Oh yes, and no DRI entry, whatever a "Define Restart Interval" is good for.

[BRT Community]

Hello,

The IC designers have confirmed that this is a limitation of the BT815, and can only be resolved in the next iteration of the silicon.

Best Regards,
FTDI Community