Results 1 to 10 of 10

Thread: Why are compression advances so hard to sell?

  1. #1
    Member SolidComp's Avatar
    Join Date
    Jun 2015
    Location
    USA
    Posts
    222
    Thanks
    89
    Thanked 46 Times in 30 Posts

    Why are compression advances so hard to sell?

    Hi all -- I just read Raphael Canut saying he was giving up on his codec after he contacted JPEG, MPEG, etc. and no one was interested. This isn't the first time I've seen developers on this forum struggle to sell or even give away their innovations to companies and standards bodies.

    Why are compression advances so hard to sell? Storage costs money. Bandwidth costs money. It's not economically rational to waste as much bandwidth and compute as companies are wasting right now with the suboptimal compression codecs in wide use (not to mention situations where they don't compress at all...).

    Some of you are prospering. Some of you are struggling. It's not clear to me what differentiates the successful from the unsuccessful in this field. I'd like to hear your thoughts. Also, I wonder if there's anything we can do to help Raphael and others market themselves and their codecs. Lack of perfect English fluency is a problem for some – any of you should feel free to message me if you need help with your English pitches. I'm happy to help you with English editing and the pitch. I'm a social psychologist who trained under Robert Cialdini, so I know a lot about the science of persuasion and can help with pitches. Still, it seems like such a hard sell to get people to listen to someone about better compression...

    Thoughts?

  2. #2
    Member
    Join Date
    Jun 2015
    Location
    Switzerland
    Posts
    667
    Thanks
    204
    Thanked 241 Times in 146 Posts
    Currently there are at least two standardization efforts going on about image compression, JPEG XL at ISO and AVIF at the Alliance for Open Media. Both have allowed open participation of new members, alliance of open media has 39 members, and ISO/JPEG XL had 8 algorithms participating initially.

  3. #3
    Member
    Join Date
    Nov 2013
    Location
    Kraków, Poland
    Posts
    645
    Thanks
    205
    Thanked 196 Times in 119 Posts
    SolidComp, building a good compressor is only one of factors in this game of big commercial players - whose main priorities are money and control, like the number of patents they can squeeze into a codec. Hence, e.g. MPEG compressor needs to use heavily patented CABAC, including genomic MPEG-G.
    If you want your compressor to be widely used, the most realistic way now is being hired by a corporation (like Yann) ... but it often means giving them all your intellectual property, you might be forbidden to use own work if changing employer (e.g. word2vec G patent whose author has later moved to Fb).
    Sure if your idea has turned out successful, you can count on their interest: trying to patent your work behind your back, don't even try thinking that you can get anything for expertise, helping them in such situation.
    The idea ownership system has made it hopeless, the only you can do is patenting yourself, or produce prior art to make it harder for them to freeze ideas for 20 years.

    ps. Finally a way little guy can get anything for working in data compression - you can find bugs in 7-zip ;D
    https://www.techspot.com/news/78051-...including.html
    Last edited by Jarek; 31st December 2018 at 09:35. Reason: bugs

  4. The Following User Says Thank You to Jarek For This Useful Post:

    compgt (31st December 2018)

  5. #4
    Member
    Join Date
    Sep 2018
    Location
    Philippines
    Posts
    29
    Thanks
    12
    Thanked 0 Times in 0 Posts
    Because JPEG and MPEG are the experts and it takes years to approve a standard algorithm. In my blog, i state that even the GUI archivers were already developed in the late 1970s to early 80s, like Windows. They are fine-tuned and enhanced through time then released to the public. I wonder if they have already holographic compression algorithms for holograms, in the waiting, for example, until commercially viable.

    Having stated in my blog that i probably helped in programming gzip, snappy, and arj during the Cold War era, i am now only limited to the C language, but if great compression ideas come, i would eagerly code the algorithms. Until then, my best general-purpose LZ77-based compressor (no better than gzip, by the way) is lzuf5.
    Last edited by compgt; 31st December 2018 at 12:55. Reason: Snappy

  6. #5
    Member
    Join Date
    Nov 2013
    Location
    Kraków, Poland
    Posts
    645
    Thanks
    205
    Thanked 196 Times in 119 Posts
    compgt, more specifically, because MPEG has covered this field with hundreds of patents - getting legal ownership of the entire field and paralyzing development.
    Now they are shifting e.g. to compression of genomic data, wanting to also get ownership of this (previously freely developed by academics) field by carpet bombing it with obfuscated patents mainly on known ideas: http://datageekdom.blogspot.com/2018...eg-g-ugly.html

  7. #6
    Member
    Join Date
    Aug 2018
    Location
    France
    Posts
    100
    Thanks
    7
    Thanked 5 Times in 4 Posts
    > it takes years to approve a standard algorithm

    Yes, it is actually more than 10 years that I have made the first version of my codec and that I try to "sell" it, but all the answers I had, and accumulated, over the years by the industry is sorry we can not test your work and are not open to external work... It required me a lot of courage and motivation to not abandon the NHW Project across all these years despite discouraging answers, depression, a lot of negative people,... because I always found it was with my eyes of better quality, more "neatness" of image, and also new and royalty-free and a lot faster to encode/decode...

    I will try now that the NHW Project is more complete to contact again some companies with SolidComp who will help me to notably better communicate/sell my work.Maybe we should focus on mobile devices manufacturers? Because I wonder if AOM AV1 and VVC that are very slow are desgined for shooting videos on a mobile phone for example?

    Many thanks!
    Cheers,
    Raphael

  8. #7
    Member
    Join Date
    Apr 2012
    Location
    Stuttgart
    Posts
    437
    Thanks
    1
    Thanked 96 Times in 57 Posts
    Quote Originally Posted by SolidComp View Post
    Hi all -- I just read Raphael Canut saying he was giving up on his codec after he contacted JPEG, MPEG, etc. and no one was interested.
    That's simply not true. The algorithm - at the deadline for the proposals - was simply not able to compress realistically sized images, and was not able to handle the image sizes in the test dataset. We are not going to hold an initiative because one interested participant is late. So please, stop the conspiracy theories, would you?

    Quote Originally Posted by SolidComp View Post
    Why are compression advances so hard to sell?
    Because these are only incremental advances. If something is good enough to solve your problem, it is probably good enough. Consider you would be able to save 30% more images at the same quality on an SD card. Would you be willing to give up compatibility with your tool chain for that, or would you just wait for 12 month and buy a bigger SD card for the same price?

  9. #8
    Member
    Join Date
    Apr 2012
    Location
    Stuttgart
    Posts
    437
    Thanks
    1
    Thanked 96 Times in 57 Posts
    Quote Originally Posted by compgt View Post
    I wonder if they have already holographic compression algorithms for holograms, in the waiting, for example, until commercially viable.
    No worries. Holographic compression is in the making, as part of the JPEG Holo project. Still in an early stage.

  10. #9
    Member
    Join Date
    Dec 2011
    Location
    Cambridge, UK
    Posts
    437
    Thanks
    137
    Thanked 152 Times in 100 Posts
    Quote Originally Posted by Jarek View Post
    compgt, more specifically, because MPEG has covered this field with hundreds of patents - getting legal ownership of the entire field and paralyzing development.
    Now they are shifting e.g. to compression of genomic data, wanting to also get ownership of this (previously freely developed by academics) field by carpet bombing it with obfuscated patents mainly on known ideas: http://datageekdom.blogspot.com/2018...eg-g-ugly.html
    AV1 proved there are still commercial incentives to not go down the patent-everything route. The whole "if it can't be monetised then we'll have no future developments" meme is just rubbish. If a major vendor can SAVE money, then this amounts to the same as MAKING money at the end of the day. I know some will want both, but AV1 demonstrates that innovation does not require patents.

    Genomics is an interesting situation because the field has been very patent averse and science lead until recently. Sadly I fear a greater and greater portion of my day job will now involve fighting pointless battles, rather than doing the things I love.

  11. The Following 3 Users Say Thank You to JamesB For This Useful Post:

    Cyan (2nd January 2019),Jarek (2nd January 2019),Mike (2nd January 2019)

  12. #10
    Member
    Join Date
    Jun 2015
    Location
    Switzerland
    Posts
    667
    Thanks
    204
    Thanked 241 Times in 146 Posts
    Quote Originally Posted by JamesB View Post
    AV1 proved there are still commercial incentives to not go down the patent-everything route.
    This is interesting. https://en.wikipedia.org/wiki/Alliance_for_Open_Media writes about patents, but doesn't directly support your claim. Where did you read about patent coverage in AV1?

Similar Threads

  1. Easy for you , very hard for me......aid !
    By plimo in forum Data Compression
    Replies: 23
    Last Post: 8th May 2014, 03:35
  2. Virtual Hard Disk Compress/Dedupe
    By JayM in forum Data Compression
    Replies: 6
    Last Post: 22nd July 2013, 00:00
  3. Replies: 4
    Last Post: 17th March 2008, 21:19

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •