Samples

You will find in this section images and videos samples compressed with the OpenJPEG codec, as well as the command line used for the compression.

J2K Sample files

• Bretagne1
  
  

  Image characteristics : 640 x 480 (3 components, 8 bits per component) Original image:   [ppm] [bmp] (901 K) Compressed image: [j2k] (90 K) Command line used: image_to_j2k -i Bretagne1.ppm -o Bretagne1.j2k -r 200,50,10 -> Three quality layers with compression ratios of 200, 50 and 10

  
  
  
  
  
  
  
  
  

 

 

 

• Bretagne2
  
  

  Image characteristics : 2592 x 1944 (3 components, 8 bits per component) Original image:   [ppm] [bmp] (14.763 K) Compressed image: [j2k] (5.275 K) Command line used: image_to_j2k -i Bretagne2.ppm -o Bretagne2.j2k -c [128,128],[128,128],[128,128] -r 100,20,2 -t 640,480 -b 32,32 -> Precinct sizes of [128,128],[128,128],[128,128],[64,64],[32,32],[16,16] for the resolutions ([16,16] corresponds to the lower resolution) ->  Three quality layers with compression ratios of 100, 20 and 2 -> Tile sizes of 640x480 -> Code-block sizes of 32x32

 

 

 

 

 

 

 

 

 

 

 

• Cevennes1 
  
  

  Image characteristics : 2592 x 1944 (3 components, 8 bits per component) Original image:   [ppm] [bmp] (14.763 K) Compressed image: [j2k] (1476 K) Command line used: image_to_j2k -i Cevennes1.ppm -o Cevennes1.j2k -r 10 -> Compression ratio: 10

  
  
  
  
  

 

 

 

 

 

JP2 Sample files

• Cevennes2
  
  

  Image characteristics : 640 x 480 (3 components, 8 bits per component) Original image:   [ppm] [bmp] (901 K) Compressed image: [jp2] (18 K) Command line used: image_to_j2k -i Cevennes2.ppm -o Cevennes2.jp2 -r 50 -> Compression ratio: 50

  
  
  
  
  
  
  
  
  
  

 

 

 

• Rome 
  
  

  Image characteristics : 640 x 480 (3 components, 8 bits per component)Original image:   [ppm] [bmp] (901 K) Compressed image: [jp2] (379 K) Command line used: image_to_j2k -i Rome.ppm -o Rome.jp2 -q 30,35,50 -p LRCP -n 3 -> Three quality layers with distorsions of 30dB, 35dB and 50 dB -> Progression order LRCP (layer-resolution-component-precinct) -> Three resolutions (which means two decompositions)

  
  
  
  
  
  

 

 

 

 

 

MJ2 Sample files

• Speedway
  
  

  Sequence characteristics : 200 frames, 352 x 288 (1 components, 8 bpp) Original sequence:   [yuv] (29.700 K) Compressed image: [mj2] (2947 K) Command line used: frames_to_mj2 -i Speedway.yuv -o Speedway.mj2 -r 10 -> Compression ratio: 10

  
  
  
  
  

 

 

 

 

 

JPWL Sample files

Facade


Image characteristics : 1850 x 2160 (3 components, 8 bits per component) Original image:   [pnm] [bmp] (11992 K) Compressed image: [j2k] (2124 K) Compressed image with purposely added errors: [j2k] (2124 K) Command line output: [txt] Index file: [txt] Command line used: JPWL_image_to_j2k.exe -r 40,20,10 -t 925,540 -i facade.pnm -o facade.j2k -W h,h0=64,h3=16,h5=32,p0=78,p0:24=56,p1,p3:0=0,p3:20=32, p5=64,p6=0,s=0,s0=6,s3=-1,a=0,g=1,z=1 -x facade_idx.txt -> Three quality layers with compression ratios of 40, 20, and 10 -> Tile sizes of 925x540 -> Index file named facade_idx.txt -> predefined Error Protection Block (EPB) in main header (MH): h -> EPB with Reed-Solomon code RS(64,32) for tile part headers (TPHs) starting from tile 0: h0=64 -> EPB with Cyclic Redundancy Check CRC-16 for TPHs starting from tile 3: h3=16 -> EPB with CRC-32 for TPHs starting from tile 5: h5=32 -> EPB with RS(78,32) starting from packet 0 of tile 0: p0=78 -> EPB with RS(56,32) from packet 24 of tile 0: p0:24=56 -> EPB with default RS starting from packets of tile 1: p1 -> EPB with no protection at all from packet 0 of tile 3: p3:0=0 -> EPB with CRC-32 from packet 20 of tile 3: p3:20=32 -> EPB with RS(64,32) starting from packets of tile 5: p5=64 -> EPB with no protection at all from packet 0 of tile 6: p6=0 -> Error Sensitivity Data (ESD) with relative sensitivity for MH: s=0 -> ESD with TSE from TPH 0: s0=6 -> no ESD from TPH 3: s3=-1 -> data addressing with automatic size of addresses: a=0 -> byte range addressing mode: g=1 -> 1 byte for each sensitivity value: z=1 In summary, we have specified to have predefined EPB in MH, rs(64,32) from TPH 0 to TPH 2, CRC-16 in TPH 3 and TPH 4, CRC-32 in remaining TPHs, Unequal Erro Protection with RS(78,32) for packets 0 to 23 of tile 0, RS(56,32) for packets 24 to the last one of tile 0, default RS for packets of tiles 1 and 2, no protection for packets 0 to 19 of tile 3, CRC-32 for packets 20 to the last one of tile 3, RS(64,32) for packets in tile 5, no protection at all for packets of remaining tiles, relative sensitivity ESD for MH, TSE ESD from TPH 0 to TPH 2, byte range with automatic size of addresses, and 1 byte for each sensitivity value. Now, we can test if JPWL stuff works. Use the compressed image with errors, purposely added as if it were transmitted over a Binary Symmetric Channel (BSC) with crossover/bit error rate of 1x10-4. Decoding with JPWL correction disabled: JPWL_j2k_to_image.exe -i FacadeERR.j2k -o FacadeOPJ.bmp Decoding with JPWL correction enabled: JPWL_j2k_to_image.exe -i FacadeERR.j2k -o FacadeOPJW.bmp -W c=3 -> an image with three components is expected: c=3 (this option is necessary, since if the compressed image is highly damaged, JPWL code can try to get in synch with the codestream from the very start of it) Below you can find the comparison between the image decoded with JPWL correction disabled (left) and the image decoded with JPWL correction enabled (right). Since tile 0, 1, 2, and 5 are well protected, they are retrieved. The remaining tiles have no error correction for data packets: their contents are garbled. Looking at the command line output, you can see that when JPWL is enabled many errors are corrected and the existence of unrecoverable data is signalled, differently from the JPWL-disabled case.