Sur d’autres sites (7102)

• Your 6-step guide to increasing acquisition

  2 juillet 2019, par Matomo Core Team — Analytics Tips
  Your 6-step guide to increasing acquisition

  Want to save time and money, as well as increase conversions and acquisition ? Matomo Analytics is here to help with that !

  Let’s start by helping you create a website visitors’ acquisition strategy, without it you might be going in blind and missing opportunities that might’ve been easily found in your metrics.

  To help you craft a strategy for your site, check out the steps below !

  Step one : Get familiar with the Acquisition feature

  The easiest way is to start with Matomo’s Acquisition feature itself. Discover and take action on the marketing channels with the biggest ROI for your business. You’ll learn :

  How to get traffic from external websites : Find out who’s helping you succeed from external websites and convince them to do more of it. Get more traffic by proactively asking for : paid sponsorships ; guest blog posts ; or spending more advertising on the particular website.

  

  About Social Networks : Which social media channels are connecting with the audience you want ? Take the guesswork out by using only the ones you need. By finding out which social channels your ideal audience prefers, you can generate shareable, convincing and engaging content to drive shares and traffic through to your site.

  Campaigns : This helps you understand which marketing campaign is working and which isn’t. You can then shift your efforts to effectively gain more visitors with less costs. Keep track of every ad and content piece you show across internal and external channels to see which has the biggest impact on your business objectives.

  Enhanced SEO : Every acquisition plan needs a focus on maximising your Search Engine Optimization (SEO) efforts. When it comes to getting conclusive search engine referrer metrics, you need to be sure you’re getting ALL the insights to drive your SEO strategy. See keyword position rankings, integrate Google, Bing and Yahoo search consoles, and no longer be restricted with “keyword not defined” showing up in your keywords reports.

  >> Watch Acquisition introduction video (playtime : 2.54 minutes)

  Step two : Set your goals and monitor conversion funnels

  Let the Goals feature guide you

  Goals are essential for building your marketing strategy and getting new customers. The more goals you track, the more you learn about behavioural changes and modify pathways to impact acquisitions over time. 

  Are you checking :

  • Which channels are converting the best for your business ?
  • Which cities/countries are most popular ?
  • What devices will attract the most visitors ?
  • How engaged your visitors are before converting ?

  This way you can see if your campaigns (SEO, PPC, signups, blogs etc.) or optimising efforts (A/B Testing, Funnels) have made an impact with the time and investment you’ve put in.

  >> Watch Goals introduction video (playtime : 2.04 minutes)

  The Funnels feature leads you to success

  Conversion funnels give you the big picture on whether your acquisition plans are paying off and where they may be falling short. If the ultimate goal of your site is to drive conversions, then each funnel can tell you how effectively you’re driving traffic through to your desired outcome.

  >> Watch Funnels introduction video (playtime : 2.29 minutes)

  

  Step three : Measure the success of every touchpoint in your customer’s journey

  Multi Attribution feature

  Accurately identify channels where visitors first engage with your business, as well as the final channel they came from, before purchasing your product/service. This helps you make smarter decisions when determining acquisition spend to accurately calculate the Customer Acquisition Cost (CAC). Here you no longer falsely over-estimate investment in failing marketing channels.

  >> Watch Multi Attribution introduction video (playtime : 2.28 minutes)

  Step four : For ecommerce sites, understand who your customers are to increase sales

  Ecommerce feature to significantly increase $ potential

  If your website’s overall purpose is to generate revenue, the Ecommerce feature gives you comprehensive insights into your customer’s purchasing behaviours.

  This heavily reduces your risks when marketing products to potential customers as you’ll understand who to target, what to target them with and where further opportunities exist.

  >> Watch Ecommerce introduction video (playtime : 2.04 minutes)

  

  Step five : Make sure the forms on your website are easy to complete

  Form Analytics feature

  Once you get visitors through the funnel, the forms on your website are the final step to conversion and need special attention. If not done right, you could be missing out on converting a large portion of your visitors.

  Thankfully, you can now identify and fix pain points on the forms that are most important to your business’ success.

  >> Watch Form Analytics introduction video (playtime : 2.39 minutes)

  

  Step six : Discover what a customer journey looks like on a user-by-user basis and bring in key acquisition elements to your strategy

  Visitor Profiles tell you each visitors’ history

  The Profile feature summarises every visit, action and purchase made.

  Better understand :

  • Why your visitors viewed your website.
  • Why your returning visitors continue to view your website.
  • What specifically your visitors are looking for and whether they found it on your website.

  The benefit is being able to see how a combination of acquisition channels play a part in a single buyer’s journey.

  >> Watch Visitors introduction video (playtime : 1.46 minutes)

  

  To summarise

  This guide will set you on a path to creating a well-planned acquisition strategy. It’s the key to attracting and capturing the attention of potential visitors/leads, and successfully driving them through a funnel/buyer’s journey on your website.

  Because of Matomo’s reputation as a trusted analytics platform, the features above can be used to assist you in making smarter data-driven decisions. You can pursue different acquisition avenues with confidence and create a strategy that’s agile and ready for success, all while respecting user privacy.

• FFMPEG overlay by time not working in my case

  18 août 2021, par Patel Milan

  Image Overlay on image and enable by time

  


  ffmpeg -y -loop 1 -i .\1080.png -i .\021.jpg -i .\022.jpg -i .\023.jpg -filter_complex "  [1:v]scale=534:810[a]; [2:v]scale=534:810[b]; [3:v]scale=534:810[c]; [0:v][a] overlay=10:8:enable='between(t,0,8)'[o1]; [o1][b] overlay=264:778:enable='between(t,1,8)'[o2]; [o2][c]  overlay=534:1524:enable='between(t,2,8)'[o3]" -map "[o3]" -t 8 outImageOverlay.mp4


  


  Log

  


  ffmpeg version 4.3.1-2021-01-01-essentials_build-www.gyan.dev Copyright (c) 2000-2021 the FFmpeg developers
  built with gcc 10.2.0 (Rev5, Built by MSYS2 project)
  configuration: --enable-gpl --enable-version3 --enable-static --disable-w32threads --disable-autodetect --enable-fontconfig --enable-iconv --enable-gnutls --enable-libxml2 --enable-gmp --enable-lzma --enable-zlib --enable-libsrt --enable-libssh --enable-libzmq --enable-avisynth --enable-sdl2 --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxvid --enable-libaom --enable-libopenjpeg --enable-libvpx --enable-libass --enable-libfreetype --enable-libfribidi --enable-libvidstab --enable-libvmaf --enable-libzimg --enable-amf --enable-cuda-llvm --enable-cuvid --enable-ffnvcodec --enable-nvdec --enable-nvenc --enable-d3d11va --enable-dxva2 --enable-libmfx --enable-libgme --enable-libopenmpt --enable-libopencore-amrwb --enable-libmp3lame --enable-libtheora --enable-libvo-amrwbenc --enable-libgsm --enable-libopencore-amrnb --enable-libopus --enable-libspeex --enable-libvorbis --enable-librubberband
  libavutil      56. 51.100 / 56. 51.100
  libavcodec     58. 91.100 / 58. 91.100
  libavformat    58. 45.100 / 58. 45.100
  libavdevice    58. 10.100 / 58. 10.100
  libavfilter     7. 85.100 /  7. 85.100
  libswscale      5.  7.100 /  5.  7.100
  libswresample   3.  7.100 /  3.  7.100
  libpostproc    55.  7.100 / 55.  7.100
Input #0, png_pipe, from '.\1080.png':
  Duration: N/A, bitrate: N/A
    Stream #0:0: Video: png, rgba(pc), 1080x2340, 25 fps, 25 tbr, 25 tbn, 25 tbc
Input #1, image2, from '.\021.jpg':
  Duration: 00:00:00.04, start: 0.000000, bitrate: 286665 kb/s
    Stream #1:0: Video: mjpeg (Baseline), yuvj444p(pc, bt470bg/unknown/unknown), 1080x2340, 25 tbr, 25 tbn, 25 tbc
Input #2, image2, from '.\022.jpg':
  Duration: 00:00:00.04, start: 0.000000, bitrate: 337493 kb/s
    Stream #2:0: Video: mjpeg (Baseline), yuvj444p(pc, bt470bg/unknown/unknown), 1080x2340, 25 tbr, 25 tbn, 25 tbc
Input #3, image2, from '.\023.jpg':
  Duration: 00:00:00.04, start: 0.000000, bitrate: 298403 kb/s
    Stream #3:0: Video: mjpeg (Baseline), yuvj444p(pc, bt470bg/unknown/unknown), 1080x2340, 25 tbr, 25 tbn, 25 tbc
Stream mapping:
  Stream #0:0 (png) -> overlay:main
  Stream #1:0 (mjpeg) -> scale
  Stream #2:0 (mjpeg) -> scale
  Stream #3:0 (mjpeg) -> scale
  overlay -> Stream #0:0 (libx264)
Press [q] to stop, [?] for help
[swscaler @ 000001cd0b24d000] deprecated pixel format used, make sure you did set range correctly
[swscaler @ 000001cd0b286080] deprecated pixel format used, make sure you did set range correctly
[swscaler @ 000001cd0b2c9c40] deprecated pixel format used, make sure you did set range correctly
[libx264 @ 000001cd0a848500] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX FMA3 BMI2 AVX2
[libx264 @ 000001cd0a848500] profile High, level 5.0, 4:2:0, 8-bit
[libx264 @ 000001cd0a848500] 264 - core 161 r3027 4121277 - H.264/MPEG-4 AVC codec - Copyleft 2003-2020 - http://www.videolan.org/x264.html - options: cabac=1 ref=3 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=7 psy=1 psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=6 lookahead_threads=1 sliced_threads=0 nr=0 decimate=1 interlaced=0 bluray_compat=0 constrained_intra=0 bframes=3 b_pyramid=2 b_adapt=1 b_bias=0 direct=1 weightb=1 open_gop=0 weightp=2 keyint=250 keyint_min=25 scenecut=40 intra_refresh=0 rc_lookahead=40 rc=crf mbtree=1 crf=23.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
Output #0, mp4, to 'outImageOverlay.mp4':
  Metadata:
    encoder         : Lavf58.45.100
    Stream #0:0: Video: h264 (libx264) (avc1 / 0x31637661), yuv420p, 1080x2340, q=-1--1, 25 fps, 12800 tbn, 25 tbc (default)
    Metadata:
      encoder         : Lavc58.91.100 libx264
    Side data:
      cpb: bitrate max/min/avg: 0/0/0 buffer size: 0 vbv_delay: N/A
frame=  200 fps= 43 q=-1.0 Lsize=     208kB time=00:00:07.88 bitrate= 216.6kbits/s speed= 1.7x
video:205kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 1.538915%
[libx264 @ 000001cd0a848500] frame I:1     Avg QP:10.61  size: 68492
[libx264 @ 000001cd0a848500] frame P:50    Avg QP:16.79  size:  2508
[libx264 @ 000001cd0a848500] frame B:149   Avg QP:29.02  size:   104
[libx264 @ 000001cd0a848500] consecutive B-frames:  0.5%  0.0%  1.5% 98.0%
[libx264 @ 000001cd0a848500] mb I  I16..4: 82.0% 10.1%  7.9%
[libx264 @ 000001cd0a848500] mb P  I16..4:  0.0%  0.5%  0.1%  P16..4:  1.0%  0.0%  0.1%  0.0%  0.0%    skip:98.2%
[libx264 @ 000001cd0a848500] mb B  I16..4:  0.0%  0.0%  0.0%  B16..8:  0.3%  0.0%  0.0%  direct: 0.0%  skip:99.7%  L0:39.1% L1:60.9% BI: 0.0%
[libx264 @ 000001cd0a848500] 8x8 transform intra:28.0% inter:98.0%
[libx264 @ 000001cd0a848500] coded y,uvDC,uvAC intra: 35.8% 31.1% 23.2% inter: 0.0% 0.1% 0.0%
[libx264 @ 000001cd0a848500] i16 v,h,dc,p: 98%  1%  0%  1%
[libx264 @ 000001cd0a848500] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 19% 17% 15%  7%  7%  9%  8%  8% 11%
[libx264 @ 000001cd0a848500] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 18% 21%  9%  7% 10%  9% 11%  6%  9%
[libx264 @ 000001cd0a848500] i8c dc,h,v,p: 76% 10% 10%  4%
[libx264 @ 000001cd0a848500] Weighted P-Frames: Y:0.0% UV:0.0%
[libx264 @ 000001cd0a848500] ref P L0: 90.9%  0.7%  7.0%  1.3%
[libx264 @ 000001cd0a848500] ref B L0: 57.1% 42.0%  0.9%
[libx264 @ 000001cd0a848500] ref B L1: 93.2%  6.8%
[libx264 @ 000001cd0a848500] kb/s:209.46


  


  Overlay Video on Image Command

  


  ffmpeg -y -i love.mp4 -i .\1080.png -i .\021.jpg -i .\022.jpg -i .\023.jpg -loop 1 -i .\020.jpg -filter_complex " [2:v]scale=534:810[a]; [3:v]scale=534:810[b]; [4:v]scale=534:810[c]; [5:v]scale=8000:4000,zoompan=z='min(zoom+0.0020,1.5)':d=417:s=1080x2340,setsar=1[d]; [0:v]scale=1080x2340,setdar=1080:2340,colorkey=0x1CD51A:0.3:0.2[ckout]; [1:v][a] overlay=10:8:enable='between(t,0,8)'[o1]; [o1][b] overlay=264:778:enable='between(t,1,8)'[o2]; [o2][c]overlay=534:1524:enable='between(t,2,8)'[o3]; [d][o3]overlay[o4]; [o4][ckout]overlay[o5]" -map "[o5]" -pix_fmt yuvj422p -t 8 outvideoOverlayInImage.mp4


  


  Log

  


  ffmpeg version 4.3.1-2021-01-01-essentials_build-www.gyan.dev Copyright (c) 2000-2021 the FFmpeg developers
  built with gcc 10.2.0 (Rev5, Built by MSYS2 project)
  configuration: --enable-gpl --enable-version3 --enable-static --disable-w32threads --disable-autodetect --enable-fontconfig --enable-iconv --enable-gnutls --enable-libxml2 --enable-gmp --enable-lzma --enable-zlib --enable-libsrt --enable-libssh --enable-libzmq --enable-avisynth --enable-sdl2 --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxvid --enable-libaom --enable-libopenjpeg --enable-libvpx --enable-libass --enable-libfreetype --enable-libfribidi --enable-libvidstab --enable-libvmaf --enable-libzimg --enable-amf --enable-cuda-llvm --enable-cuvid --enable-ffnvcodec --enable-nvdec --enable-nvenc --enable-d3d11va --enable-dxva2 --enable-libmfx --enable-libgme --enable-libopenmpt --enable-libopencore-amrwb --enable-libmp3lame --enable-libtheora --enable-libvo-amrwbenc --enable-libgsm --enable-libopencore-amrnb --enable-libopus --enable-libspeex --enable-libvorbis --enable-librubberband
  libavutil      56. 51.100 / 56. 51.100
  libavcodec     58. 91.100 / 58. 91.100
  libavformat    58. 45.100 / 58. 45.100
  libavdevice    58. 10.100 / 58. 10.100
  libavfilter     7. 85.100 /  7. 85.100
  libswscale      5.  7.100 /  5.  7.100
  libswresample   3.  7.100 /  3.  7.100
  libpostproc    55.  7.100 / 55.  7.100
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'love.mp4':
  Metadata:
    major_brand     : mp42
    minor_version   : 0
    compatible_brands: isommp42
    creation_time   : 2021-08-17T05:35:07.000000Z
    com.android.version: 11
  Duration: 00:00:06.93, start: 0.000000, bitrate: 538 kb/s
    Stream #0:0(eng): Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 196 kb/s (default)
    Metadata:
      creation_time   : 2021-08-17T05:35:07.000000Z
      handler_name    : SoundHandle
    Stream #0:1(eng): Video: h264 (Baseline) (avc1 / 0x31637661), yuv420p(tv, smpte170m/bt470bg/smpte170m), 1080x2340, 334 kb/s, SAR 1:1 DAR 6:13, 25 fps, 25 tbr, 90k tbn, 180k tbc (default)
    Metadata:
      creation_time   : 2021-08-17T05:35:07.000000Z
      handler_name    : VideoHandle
Input #1, png_pipe, from '.\1080.png':
  Duration: N/A, bitrate: N/A
    Stream #1:0: Video: png, rgba(pc), 1080x2340, 25 tbr, 25 tbn, 25 tbc
Input #2, image2, from '.\021.jpg':
  Duration: 00:00:00.04, start: 0.000000, bitrate: 286665 kb/s
    Stream #2:0: Video: mjpeg (Baseline), yuvj444p(pc, bt470bg/unknown/unknown), 1080x2340, 25 tbr, 25 tbn, 25 tbc
Input #3, image2, from '.\022.jpg':
  Duration: 00:00:00.04, start: 0.000000, bitrate: 337493 kb/s
    Stream #3:0: Video: mjpeg (Baseline), yuvj444p(pc, bt470bg/unknown/unknown), 1080x2340, 25 tbr, 25 tbn, 25 tbc
Input #4, image2, from '.\023.jpg':
  Duration: 00:00:00.04, start: 0.000000, bitrate: 298403 kb/s
    Stream #4:0: Video: mjpeg (Baseline), yuvj444p(pc, bt470bg/unknown/unknown), 1080x2340, 25 tbr, 25 tbn, 25 tbc
Input #5, image2, from '.\020.jpg':
  Duration: 00:00:00.04, start: 0.000000, bitrate: 184663 kb/s
    Stream #5:0: Video: mjpeg (Baseline), yuvj444p(pc, bt470bg/unknown/unknown), 1080x2340, 25 fps, 25 tbr, 25 tbn, 25 tbc
Stream mapping:
  Stream #0:1 (h264) -> scale
  Stream #1:0 (png) -> overlay:main
  Stream #2:0 (mjpeg) -> scale
  Stream #3:0 (mjpeg) -> scale
  Stream #4:0 (mjpeg) -> scale
  Stream #5:0 (mjpeg) -> scale
  overlay -> Stream #0:0 (libx264)
Press [q] to stop, [?] for help
[swscaler @ 00000230595cff40] deprecated pixel format used, make sure you did set range correctly
[swscaler @ 0000023059727e80] deprecated pixel format used, make sure you did set range correctly
[swscaler @ 00000230597768c0] deprecated pixel format used, make sure you did set range correctly
[swscaler @ 00000230597c3c80] deprecated pixel format used, make sure you did set range correctly
[swscaler @ 00000230597faec0] No accelerated colorspace conversion found from yuv420p to argb.
[swscaler @ 0000023059884cc0] deprecated pixel format used, make sure you did set range correctly
[libx264 @ 00000230536e2900] using SAR=1/1
[libx264 @ 00000230536e2900] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX FMA3 BMI2 AVX2
[libx264 @ 00000230536e2900] profile High 4:2:2, level 5.0, 4:2:2, 8-bit
[libx264 @ 00000230536e2900] 264 - core 161 r3027 4121277 - H.264/MPEG-4 AVC codec - Copyleft 2003-2020 - http://www.videolan.org/x264.html - options: cabac=1 ref=3 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=7 psy=1 psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=6 lookahead_threads=1 sliced_threads=0 nr=0 decimate=1 interlaced=0 bluray_compat=0 constrained_intra=0 bframes=3 b_pyramid=2 b_adapt=1 b_bias=0 direct=1 weightb=1 open_gop=0 weightp=2 keyint=250 keyint_min=25 scenecut=40 intra_refresh=0 rc_lookahead=40 rc=crf mbtree=1 crf=23.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
Output #0, mp4, to 'outvideoOverlayInImage.mp4':
  Metadata:
    major_brand     : mp42
    minor_version   : 0
    compatible_brands: isommp42
    com.android.version: 11
    encoder         : Lavf58.45.100
    Stream #0:0: Video: h264 (libx264) (avc1 / 0x31637661), yuvj422p(pc), 1080x2340 [SAR 1:1 DAR 6:13], q=-1--1, 25 fps, 12800 tbn, 25 tbc (default)
    Metadata:
      encoder         : Lavc58.91.100 libx264
    Side data:
      cpb: bitrate max/min/avg: 0/0/0 buffer size: 0 vbv_delay: N/A
frame=  200 fps= 11 q=-1.0 Lsize=    1411kB time=00:00:07.88 bitrate=1467.1kbits/s speed=0.435x
video:1408kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.226583%
[libx264 @ 00000230536e2900] frame I:1     Avg QP:20.68  size:121139
[libx264 @ 00000230536e2900] frame P:50    Avg QP:20.09  size: 15622
[libx264 @ 00000230536e2900] frame B:149   Avg QP:24.04  size:  3617
[libx264 @ 00000230536e2900] consecutive B-frames:  0.5%  0.0%  1.5% 98.0%
[libx264 @ 00000230536e2900] mb I  I16..4:  8.9% 81.4%  9.7%
[libx264 @ 00000230536e2900] mb P  I16..4:  1.5%  1.9%  0.3%  P16..4: 21.8%  8.1%  4.7%  0.0%  0.0%    skip:61.6%
[libx264 @ 00000230536e2900] mb B  I16..4:  0.1%  0.1%  0.0%  B16..8: 26.8%  0.7%  0.1%  direct: 0.3%  skip:72.1%  L0:46.3% L1:53.3% BI: 0.3%
[libx264 @ 00000230536e2900] 8x8 transform intra:59.8% inter:84.9%
[libx264 @ 00000230536e2900] coded y,uvDC,uvAC intra: 30.0% 30.8% 13.5% inter: 3.0% 2.2% 0.1%
[libx264 @ 00000230536e2900] i16 v,h,dc,p: 66% 30%  2%  2%
[libx264 @ 00000230536e2900] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 29% 12% 43%  3%  3%  3%  3%  2%  3%
[libx264 @ 00000230536e2900] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 39% 19% 11%  5%  6%  5%  6%  4%  5%
[libx264 @ 00000230536e2900] i8c dc,h,v,p: 57% 19% 21%  3%
[libx264 @ 00000230536e2900] Weighted P-Frames: Y:0.0% UV:0.0%
[libx264 @ 00000230536e2900] ref P L0: 73.3% 11.4%  8.9%  6.4%
[libx264 @ 00000230536e2900] ref B L0: 96.2%  3.1%  0.7%
[libx264 @ 00000230536e2900] ref B L1: 95.5%  4.5%
[libx264 @ 00000230536e2900] kb/s:1441.17


  


  Input images and output videos

  


  

  • 020.jpg

  


  • 021.jpg

  


  • 022.jpg

  


  • 023.jpg

  


  • 1080.png

  


  • outImageOverlay

  


  • outvideoOverlayInImage.mp4

  


  • love.mp4

  


  


  image on image overlay is working by enable given time but when video overlay into image is not working correctly

  


• IJG swings again, and misses

  1er février 2010, par Mans — Multimedia

  Earlier this month the IJG unleashed version 8 of its ubiquitous libjpeg library on the world. Eager to try out the “major breakthrough in image coding technology” promised in the README file accompanying v7, I downloaded the release. A glance at the README file suggests something major indeed is afoot :

  Version 8.0 is the first release of a new generation JPEG standard to overcome the limitations of the original JPEG specification.

  The text also hints at the existence of a document detailing these marvellous new features, and a Google search later a copy has found its way onto my monitor. As I read, however, my state of mind shifts from an initial excited curiosity, through bewilderment and disbelief, finally arriving at pure merriment.

  Already on the first page it becomes clear no new JPEG standard in fact exists. All we have is an unsolicited proposal sent to the ITU-T by members of the IJG. Realising that even the most brilliant of inventions must start off as mere proposals, I carry on reading. The summary informs me that I am about to witness the introduction of three extensions to the T.81 JPEG format :

  1. An alternative coefficient scan sequence for DCT coefficient serialization
  2. A SmartScale extension in the Start-Of-Scan (SOS) marker segment
  3. A Frame Offset definition in or in addition to the Start-Of-Frame (SOF) marker segment

  Together these three extensions will, it is promised, “bring DCT based JPEG back to the forefront of state-of-the-art image coding technologies.”

  Alternative scan

  The first of the proposed extensions introduces an alternative DCT coefficient scan sequence to be used in place of the zigzag scan employed in most block transform based codecs.

  

  Alternative scan sequence

  The advantage of this scan would be that combined with the existing progressive mode, it simplifies decoding of an initial low-resolution image which is enhanced through subsequent passes. The author of the document calls this scheme “image-pyramid/hierarchical multi-resolution coding.” It is not immediately obvious to me how this constitutes even a small advance in image coding technology.

  At this point I am beginning to suspect that our friend from the IJG has been trapped in a half-world between interlaced GIF images transmitted down noisy phone lines and today’s inferno of SVC, MVC, and other buzzwords.

  (Not so) SmartScale

  Disguised behind this camel-cased moniker we encounter a method which, we are told, will provide better image quality at high compression ratios. The author has combined two well-known (to us) properties in a (to him) clever way.

  The first property concerns the perceived impact of different types of distortion in an image. When encoding with JPEG, as the quantiser is increased, the decoded image becomes ever more blocky. At a certain point, a better subjective visual quality can be achieved by down-sampling the image before encoding it, thus allowing a lower quantiser to be used. If the decoded image is scaled back up to the original size, the unpleasant, blocky appearance is replaced with a smooth blur.

  The second property belongs to the DCT where, as we all know, the top-left (DC) coefficient is the average of the entire block, its neighbours represent the lowest frequency components etc. A top-left-aligned subset of the coefficient block thus represents a low-resolution version of the full block in the spatial domain.

  In his flash of genius, our hero came up with the idea of using the DCT for down-scaling the image. Unfortunately, he appears to possess precious little knowledge of sampling theory and human visual perception. Any block-based resampling will inevitably produce sharp artefacts along the block edges. The human visual system is particularly sensitive to sharp edges, so this is one of the most unwanted types of distortion in an encoded image.

  Despite the obvious flaws in this approach, I decided to give it a try. After all, the software is already written, allowing downscaling by factors of 8/8..16.

  Using a 1280×720 test image, I encoded it with each of the nine scaling options, from unity to half size, each time adjusting the quality parameter for a final encoded file size of no more than 200000 bytes. The following table presents the encoded file size, the libjpeg quality parameter used, and the SSIM metric for each of the images.

  Scale	Size	Quality	SSIM
  8/8	198462	59	0.940
  8/9	196337	70	0.936
  8/10	196133	79	0.934
  8/11	197179	84	0.927
  8/12	193872	89	0.915
  8/13	197153	92	0.914
  8/14	188334	94	0.899
  8/15	198911	96	0.886
  8/16	197190	97	0.869

  Although the smaller images allowed a higher quality setting to be used, the SSIM value drops significantly. Numbers may of course be misleading, but the images below speak for themselves. These are cut-outs from the full image, the original on the left, unscaled JPEG-compressed in the middle, and JPEG with 8/16 scaling to the right.

  

  

  Looking at these images, I do not need to hesitate before picking the JPEG variant I prefer.

  Frame offset

  The third and final extension proposed is quite simple and also quite pointless : a top-left cropping to be applied to the decoded image. The alleged utility of this feature would be to enable lossless cropping of a JPEG image. In a typical image workflow, however, JPEG is only used for the final published version, so the need for this feature appears quite far-fetched.

  The grand finale

  Throughout the text, the author makes references to “the fundamental DCT property for image representation.” In his own words :

  This property was found by the author during implementation of the new DCT scaling features and is after his belief one of the most important discoveries in digital image coding after releasing the JPEG standard in 1992.

  The secret is to be revealed in an annex to the main text. This annex quotes in full a post by the author to the comp.dsp Usenet group in a thread with the subject why DCT. Reading the entire thread proves quite amusing. A few excerpts follow.

  The actual reason is much simpler, and therefore apparently very difficult to recognize by complicated-thinking people.

  Here is the explanation :

  What are people doing when they have a bunch of images and want a quick preview ? They use thumbnails ! What are thumbnails ? Thumbnails are small downscaled versions of the original image ! If you want more details of the image, you can zoom in stepwise by enlarging (upscaling) the image.

  So with proper understanding of the fundamental DCT property, the MPEG folks could make their videos more scalable, but, as in the case of JPEG, they are unable to recognize this simple but basic property, unfortunately, and pursue rather inferior approaches in actual developments.

  These are just phrases, and they don’t explain anything. But this is typical for the current state in this field : The relevant people ignore and deny the true reasons, and thus they turn in a circle and no progress is being made.

  However, there are dark forces in action today which ignore and deny any fruitful advances in this field. That is the reason that we didn’t see any progress in JPEG for more than a decade, and as long as those forces dominate, we will see more confusion and less enlightenment. The truth is always simple, and the DCT *is* simple, but this fact is suppressed by established people who don’t want to lose their dubious position.

  I believe a trip to the Total Perspective Vortex may be in order. Perhaps his tin-foil hat will save him.