Horiz/Vert sampling factors reversed?

[drewnoakes]

Suresh said on the mailing list:

I found an issue with the JPEG sampling factors extracted by your library. I found the Horizontal and Vertical sampling factors switched (1,2 instead of 2,1) when I compared it with the information provided by JPEGTRAN.

I have attached a zip file with a sample program, a sample image and the jpegtran command I used to compare check the sampling factor.

Is it just an issue of switching of the 2 fields or is there some other bug? After finding this issue I am using a jpegtran crop command just to get this info which will make my date stamping a little bit slower.

Here is the ZIP file: sampling_factor.zip

[kwhopper]

They appear to be reversed. All you have to do is swap the method bodies in JpegComponent.java for getHorizontalSamplingFactor() and getVerticalSamplingFactor(). This is confirmed with ExifTool and metadata-extractor-dotnet testing.

[drewnoakes]

Closed via #156.