Should jp2 be used in mosaic datasets?

1244
4
Jump to solution
08-20-2019 09:29 PM
MarkChilcott
Occasional Contributor III

Hi Peoples,

Anyone got any references or thoughts as to the advantages / disadvantages of using jp2 in mosaic datasets?

Cheers,

Mark

Tags (3)
0 Kudos
1 Solution

Accepted Solutions
PeterBecker
Esri Regular Contributor

ArcGIS can use JP2. JP2 provides lossless and lossy compression, typically with better compression than others. (Typically about 30-40% better than JPEG {lossy] or LERC {lossless} ). The format is computationally intensive and structure means that that is requires a lot of data access to read, hence is quite a bit slower then other formats. I would not convert imagery to this format unless storage size is the most critical aspects. There is a section in Imagery Workflows specifically on Imagery Formats . This also references OptimizeRasters that can be used to convert imagery to more optimal formats such as TIF (with JPEG/Deflate), MRF (JPEG/LERC) or COG (JPEG/Deflate)

View solution in original post

4 Replies
ModyBuchbinder
Esri Regular Contributor

As far as I know this is not recommended format,

The best is tiff with JPEG compression (but it depend on the number of bends and pixel depth).

I think you should put this in the Image form...

Have fun

0 Kudos
MarkChilcott
Occasional Contributor III

Hi Mody,

I am starting to get that feeling.

Having some issues with artifacts, which I believe are being caused due to the lossy compression which causes the edges of the image to overlap with NoData areas and you see white 'rips' around the edges.  Going back to the original tifs I see less issues.  Having to shrink footprints a fair bit to get around the issue.

Looking around for any blogs or references to confirm advantages / disadvantages.

Cheers,

Mark

0 Kudos
PeterBecker
Esri Regular Contributor

ArcGIS can use JP2. JP2 provides lossless and lossy compression, typically with better compression than others. (Typically about 30-40% better than JPEG {lossy] or LERC {lossless} ). The format is computationally intensive and structure means that that is requires a lot of data access to read, hence is quite a bit slower then other formats. I would not convert imagery to this format unless storage size is the most critical aspects. There is a section in Imagery Workflows specifically on Imagery Formats . This also references OptimizeRasters that can be used to convert imagery to more optimal formats such as TIF (with JPEG/Deflate), MRF (JPEG/LERC) or COG (JPEG/Deflate)

MarkChilcott
Occasional Contributor III

Hi Peter,

Great links.  Thanks for that - it is what I have been searching for.

Cheers,

Mark

0 Kudos