Back to RasterLite2 doc index
Supported sample types
| Family | Bits | Mnemonic | Notes | ||
|---|---|---|---|---|---|
| Unsigned Integer | 1 | 1-BIT |
| ||
| 2 | 2-BIT |
| |||
| 4 | 4-BIT |
| |||
| 8 | UINT8 |
| |||
| 16 | UINT16 |
| |||
| 32 | UINT32 |
| |||
| Signed Integer | 8 | INT8 |
| ||
| 16 | INT16 |
| |||
| 32 | INT32 |
| |||
| Floating Point | 32 | FLOAT |
| ||
| 64 | DOUBLE |
|
Supported pixel types
| Mnemonic | Bands | Sample type | Notes |
|---|---|---|---|
| MONOCHROME | 1 | 1-BIT | the implicit photometric interpretation always is:
|
| PALETTE | 1 | 1-BIT 2-BIT 4-BIT UINT8 |
the implicit photometric interpretation is always the one defined by the Palette Color associated to the Raster Coverage. This actually corresponds to an RGB output image, the color of each pixel being indirectly determined by its palette index. |
| GRAYSCALE | 1 | 2-BIT 4-BIT UINT8 |
the implicit photometric interpretation is always of the Grayscale type. 0 always corresponds to full black, and the max value to full white. |
| RGB | 3 | UINT8 UINT16 |
the implicit photometric interpretation is always of the RGB type. |
| MULTIBAND | > 1 < 256 |
UINT8 UINT16 |
there is no implicit photometric interpretation. |
| DATAGRID | 1 | INT8 UINT8 INT16 UINT16 INT32 UINT32 FLOAT DOUBLE |
there is no implicit photometric interpretation. |
Supported compression algorithms
| Mnemonic | Compression type | Notes |
|---|---|---|
| NONE | lossless | no compression at all: simply an array of tightly packed values. |
| DEFLATE | lossless | a general purpose compression algorithm used by ZIP archives and implemented by zlib. You can find more detailed information about this algorithm here. |
| DEFLATE_NO | lossless | Exactly the same as DEFLATE except in that it will never apply a Delta Filter. |
| LZMA | lossless | an advanced general purpose compression algorithm used by 7-ZIP archives and implemented by liblzma. You can find more detailed information about this algorithm here. Please note: this algorithm is very slow while compressing and moderately slow while decompressing, but it usually ensures compression ratios by far better than DEFLATE (it's strongly optimized for high compression ratios, may be sacrificing speed). |
| LZMA_NO | lossless | Exactly the same as LZMA except in that it will never apply a Delta Filter. |
| LZ4 | lossless | an extremely fast general purpose compression algorithm implemented by liblz4. You can find more detailed information about this algorithm here. Please note: this algorithm is very fast both while compressing and decompressing, but it's usually unable to reach the same compression ratios than DEFLATE. (it's strongly optimized for speed, may be sacrificing compression ratios) |
| LZ4_NO | lossless | Exactly the same as LZ4 except in that it will never apply a Delta Filter. |
| ZSTD | lossless | a fast general purpose compression algorithm implemented by libzstd. You can find more detailed information about this algorithm here. Please note: this algorithm is intended to be a more modern replacement for DEFLATE. it's always faster than DEFLATE, and in it can often (but not always) reach better compression ratios than DEFLATE. |
| ZSTD_NO | lossless | Exactly the same as ZSTD except in that it will never apply a Delta Filter. |
| FAX4 | lossless | an highly specialized compression algorithm only supporting monochrome images initially developed for fax machines: implemented by libtiff. You can find more detailed information about this algorithm here. |
| PNG | lossless | a very popular general purpose image format supporting RGB, Grayscale and Palette-based colorspaces and fully supporting transparencies. Implemented by libpng. You can find more detailed information about this compressed format here. |
| JPEG | lossy | an universally spread photographic image format supporting RGB and Grayscale colorspaces but completely lacking of any transparency support. Implemented by libjpeg. You can find more detailed information about this compressed format here. |
| WEBP | lossy | an innovative photographic image format supporting RGB and Grayscale colorspaces and fully supporting transparencies. Implemented by libwebp. You can find more detailed information about this compressed format here. Please note: this algorithm is very slow while compressing, but is reasonably fast while decompressing; and usually ensures slightly better compression ratios than JPEG at the same quality level. |
| LL_WEBP | lossless | exactly the same of the previous one, but in lossless mode. This usually means accepting moderate compression ratios (about 1:2). |
| JP2 | lossy | the Jpeg2000 photographic image format supporting RGB and Grayscale colorspaces and fully supporting transparencies. Implemented by OpenJpeg. Please note: this algorithm is very slow either while compressing and compressing; it usually ensures slightly better compression ratios than JPEG at the same quality level. |
| LL_JP2 | lossless | exactly the same of the previous one, but in lossless mode. This usually means accepting moderate compression ratios (about 1:2). |
Pixel / Sample / Compression compatibility chart
| MONOCHROME | PALETTE | GRAYSCALE | RGB | MULTIBAND | DATAGRID | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| 1-BIT | 1-BIT 2-BIT 4-BIT UINT8 |
2-BIT 4-BIT UINT8 for Grayscale 16bit please see: DATAGRID UINT16 |
UINT8 | UINT16 | UINT8 (3 or 4 bands) |
UINT16 (3 or 4 bands) |
UINT8 UINT16 (2 or >= 5 bands) |
UINT8 UINT16 |
INT8 INT16 INT32 UINT32 FLOAT DOUBLE | |
| NONE | yes | yes | yes | yes | yes | yes | yes | yes | yes | yes |
| DEFLATE | yes | yes | yes | yes | yes | yes | yes | yes | yes | yes |
| DEFLATE_NO | yes | yes | yes | yes | yes | yes | yes | yes | yes | yes |
| LZMA | yes | yes | yes | yes | yes | yes | yes | yes | yes | yes |
| LZMA_NO | yes | yes | yes | yes | yes | yes | yes | yes | yes | yes |
LZ4 |
yes | yes | yes | yes | yes | yes | yes | yes | yes | yes |
| LZ4_NO | yes | yes | yes | yes | yes | yes | yes | yes | yes | yes |
| ZSTD | yes | yes | yes | yes | yes | yes | yes | yes | yes | yes |
| ZSTD_NO | yes | yes | yes | yes | yes | yes | yes | yes | yes | yes |
| FAX4 | yes | |||||||||
| PNG | yes | yes | yes | yes | yes | yes | yes | yes | ||
| JPEG | yes | yes | ||||||||
| WEBP | yes | yes | yes | |||||||
| LL_WEBP | yes | yes | yes | |||||||
| JP2 | yes | yes | yes | yes | yes | yes | ||||
| LL_JP2 | yes | yes | yes | yes | yes | yes | ||||
Multi-resolution Pyramid levels configuration
| Pixel type | Base level full resolution | Upper levels rescaled | Notes |
|---|---|---|---|
| MONOCHROME | 1-BIT | GRAYSCALE UINT8 compression PNG | applying an high-quality halftone interpolation rescaling/downsampling algorithm ensures a better visual quality but imposes a change in color model. |
| PALETTE | 1-BIT | RGB UINT8 compression PNG | same as above |
| 2-BIT | RGB UINT8 compression PNG | same as above | |
| 4-BIT | RGB UINT8 compression PNG | same as above | |
| UINT8 | RGB UINT8 compression JPEG | same as above | |
| GRAYSCALE | 2-BIT | GRAYSCALE UINT8 compression JPEG | same as above |
| 4-BIT | GRAYSCALE UINT8 compression JPEG | same as above | |
| UINT8 | same as in base-level | ||
| RGB | UINT8 | same as in base-level | |
| UINT16 | same as in base-level | ||
| MULTIBAND | UINT8 | same as in base-level | |
| UINT16 | same as in base-level | ||
| DATAGRID | INT8 | same as in base-level | |
| UINT8 | same as in base-level | ||
| INT16 | same as in base-level | ||
| UINT16 | same as in base-level | ||
| INT32 | same as in base-level | ||
| UINT32 | same as in base-level | ||
| FLOAT | same as in base-level | ||
| DOUBLE | same as in base-level |
- the same identical PixelType, SampleType and Compression are adopted for both base-level and upper rescaled-levels whenever is possible.
- except when a change in the color space is strictly required by high-quality rescaling/downsampling algorithm supporting halftone interpolation.
- all upper Pyramid levels (rescaled) are always implicitly compressed: these are always heavily reworked levels, so there is any real need to carefully preserve all input data exactly as they originally were.
- JPEG is used for either RGB and GRAYSCALE (UINT8).
- PNG is used for all limited extension samples: 1-BIT, 2-BIT and 4-bit.
All these samples could implicitly imply transparency, and only PNG will offer an adequate support. - DEFLATE is used for any other configuration.
Back to RasterLite2 doc index