Page 1 of 2 12 LastLast
Results 1 to 30 of 57

Thread: Oodle LZNA API Documentation Missing

  1. #1
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts

    Exclamation Oodle LZNA API Documentation Missing

    Does anyone have API documentation for the Oodle LZNA?

    They have not answered my request for an evaluation version of their SDK. I was considering making a plug-in out of it for DiskZIP.

    The DLL is straightforward enough to find online.

    I'd love to benchmark their algorithms on some real world test cases before getting excited about a potential investment.

  2. #2
    Member
    Join Date
    May 2012
    Location
    United States
    Posts
    343
    Thanks
    200
    Thanked 58 Times in 42 Posts
    PM sent...

  3. Thanks:

    diskzip (3rd June 2017)

  4. #3
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    You can use the Turbobench compression Benchmark . Oodle plugin access is available in the windows binaries turbobench_win64 only.
    Just put your "oo2core_4_win64.dll" in the current directory and type:

    Code:
    turbobenchs -eoodle,ml file
    ex.
    turbobenchs -eoodle,89 enwik8
    m=method: 0:LZH 1:LZHLW 2:LZNIB 3:None 4:LZB16 5:LZBLW 6:LZA 7:LZNA 8:Kraken 9:Mermaid 10:BitKnit 11:Selkie 12:Akkorokami
    l = level: 1-9

    see also the Benchmark "LzTurbo vs. Oodle Kraken/zstd"

  5. Thanks:

    diskzip (3rd June 2017)

  6. #4
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    Quote Originally Posted by dnd View Post
    You can use the Turbobench compression Benchmark . Oodle plugin access is available in the windows binaries turbobench_win64 only.
    Just put your "oo2core_4_win64.dll" in the current directory and type:

    Code:
    turbobenchs -eoodle,ml file
    ex.
    turbobenchs -eoodle,89 enwik8
    m=method: 0:LZH 1:LZHLW 2:LZNIB 3:None 4:LZB16 5:LZBLW 6:LZA 7:LZNA 8:Kraken 9:Mermaid 10:BitKnit 11:Selkie 12:Akkorokami
    l = level: 1-9

    see also the Benchmark "LzTurbo vs. Oodle Kraken/zstd"
    Thank you for sharing the command line.

    The stream has to be a single file, right? This can be a disadvantage to the compressor when compared to multiple folder/multiple file capable compressors; especially if they sort files by compression order, and this would help increase compression ratios.

    How long do you think turbobench would take to compress 22 GB using this approach? That's the total uncompressed size of my testing data - I would pack it into a .TAR file and then feed that to the benchmark, as otherwise, I have tens of thousands of files sitting in probably as many folders.

  7. #5
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    There are several options in turbobench


    1 - Benchmark multiple files recursively with option "-R"
    Code:
    turbobench -elzturbo,10 -R -i0 -j1 -J1 DIR

    2 - You can tar your multiple files into one large file and:


    - Benchmark the whole file
    Code:
    turbobench -elzturbo,10 file.tar -I1 -J1

    - or benchmark the first 100M block (decoding speed test)
    Code:
    turbobench -elzturbo,10 -I1 -J15 -B100M  file.tar

    3 - You can also use multiple blocks mode with the option "-M" (build) and "-m" (benchmark)

    The benchmark duration is depending on the codecs/levels and iterations used.
    You can set the iterations/runs for compression/decompression with the options "-i", "-I"/"-j","-J".
    The max. time per run is also limited (default: 2min.) regardless of the number of iterations.

    For speed tests it is enough to use a file <100Mb.

  8. Thanks:

    diskzip (3rd June 2017)

  9. #6
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    Quote Originally Posted by dnd View Post
    There are several options in turbobench


    1 - Benchmark multiple files recursively with option "-R"
    Code:
    turbobench -elzturbo,10 -R -i0 -j1 -J1 DIR

    2 - You can tar your multiple files into one large file and:


    - Benchmark the whole file
    Code:
    turbobench -elzturbo,10 file.tar -I1 -J1

    - or benchmark the first 100M block (decoding speed test)
    Code:
    turbobench -elzturbo,10 -I1 -J15 -B100M  file.tar

    3 - You can also use multiple blocks mode with the option "-M" (build) and "-m" (benchmark)

    The benchmark duration is depending on the codecs/levels and iterations used.
    You can set the iterations/runs for compression/decompression with the options "-i", "-I"/"-j","-J".
    The max. time per run is also limited (default: 2min.) regardless of the number of iterations.

    For speed tests it is enough to use a file <100Mb.
    Thank you.

    I want to compare file size, and not speed. So, not use it as a benchmark, but just to see how big a file is created by Oodle LZNA vs. 7Zip LZMA2. Is this possible?

  10. #7
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    Yes, for multiple files just type:

    Code:
    turbobench -eoodle,79/lzma,9d29:fb273 -R -i0 -j1 -J1 DIR
    a result summary is displayed at the end.
    Better try it first with some small files.

  11. Thanks:

    diskzip (3rd June 2017)

  12. #8
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    Quote Originally Posted by dnd View Post
    Yes, for multiple files just type:

    Code:
    turbobench -eoodle,79/lzma,9d29:fb273 -R -i0 -j1 -J1 DIR
    a result summary is displayed at the end.
    Better try it first with some small files.
    Thank you.

    Where would that create the final output archive at the end?

    Does it also support dictionary size 30/31 or 1536 MB like the latest 7-Zip does with LZMA2?

    If I do not specify /lzma,9d29:fb273, would that harm the ratio?

    I have started the test with your parameters already, it will take quite a while to process.

  13. #9
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    The process took a very long time, and at the end was the single error message "access denied <DIR NAME>".

    I am not sure if it tried to create a new archive with the exact name of the directory it was compressing.

    Is there a particular command line parameter to specify the output archive name, is this even possible?

    Or was it trying to compress the "folder file" itself, and errored out as a result?

    It would have been great to see the final compression savings!

  14. #10
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    Turbobench is an In-memory benchmark. There is no archive file written.
    The results are displayed and a file "file.tbb" is written at the end of processing.
    You can also display the results with the option "-p1 -o" (Type turbobench -h for help)
    Code:
    turbobench -p1 -o file.tbb
    You can ignore the "permission denied" message.

    lzna is optimized for binary structured files only and is performing worse as general purpose compressor.
    Compression is very slow, decompression is 2-4 times faster than lzma.

    Maximum compression benchmark
    Code:
                      size  lzturbo 49     lzma,9    oodle,79 (lzna)
    A10.jpg         842468    834018++     845895      842380+                   
    AcroRd32.exe   3870784   1404693++    1406475+    1422382          
    english.dic    4067439    752857++     851380+     878059         
    FlashMX.pdf    4526946   3707306+     3706565++   3724842         
    FP.LOG        20617071    667804++     838971+    1763279         
    MSO97.DLL      3782416   1805152++    1816397+    1858869         
    ohs.doc        4168192    773397++     790240+     800190         
    rafale.bmp     4149414    944281++     976121+     981758         
    vcfiu.hlp      4121418    604926++     613169+     630763         
    world95.txt    2988578    559315++     573160+     587502
    lzna has worse compression particulary for text files (see fp.log)

    Better make the tests with your own data.

  15. Thanks:

    diskzip (4th June 2017)

  16. #11
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    Quote Originally Posted by dnd View Post
    Turbobench is an In-memory benchmark. There is no archive file written.
    The results are displayed and a file "file.tbb" is written at the end of processing.
    You can also display the results with the option "-p1 -o" (Type turbobench -h for help)
    Code:
    turbobench -p1 -o file.tbb
    You can ignore the "permission denied" message.

    lzna is optimized for binary structured files only and is performing worse as general purpose compressor.
    Compression is very slow, decompression is 2-4 times faster than lzma.

    Maximum compression benchmark
    Code:
                      size  lzturbo 49     lzma,9    oodle,79 (lzna)
    A10.jpg         842468    834018++     845895      842380+                   
    AcroRd32.exe   3870784   1404693++    1406475+    1422382          
    english.dic    4067439    752857++     851380+     878059         
    FlashMX.pdf    4526946   3707306+     3706565++   3724842         
    FP.LOG        20617071    667804++     838971+    1763279         
    MSO97.DLL      3782416   1805152++    1816397+    1858869         
    ohs.doc        4168192    773397++     790240+     800190         
    rafale.bmp     4149414    944281++     976121+     981758         
    vcfiu.hlp      4121418    604926++     613169+     630763         
    world95.txt    2988578    559315++     573160+     587502
    lzna has worse compression particulary for text files (see fp.log)

    Better make the tests with your own data.
    Thank you.

    My data set is very binary heavy, so it is a decent use case, at least for the faster decompression.

    But the extremely slow compression is an issue, no doubt!

    So is there no way to obtain total "archive size" at the end of the operation with TurboBench?

    I downloaded the sources, but I could not find where it calls the oodle DLL - any hints?

  17. #12
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    Quote Originally Posted by diskzip View Post
    So is there no way to obtain total "archive size" at the end of the operation with TurboBench?[/CODE]
    The total "archive size" is indeed displayed at the end of the benchmark and also stored in the DIR.tbb file.
    Please test first with a few number of test files.


    Quote Originally Posted by diskzip View Post
    I downloaded the sources, but I could not find where it calls the oodle DLL - any hints?
    The oodle wrapper is not contained in the turbobench source code.

  18. Thanks:

    diskzip (4th June 2017)

  19. #13
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    Quote Originally Posted by dnd View Post
    The results are displayed and a file "file.tbb" is written at the end of processing.
    You can also display the results with the option "-p1 -o" (Type turbobench -h for help)
    I missed this part in your earlier post, thank you.

    I searched my entire hard disk for *.tbb and it turned out nothing. Where is its default location?

    I must have also missed the total size displayed at the end of processing in the console window.

    Or do you think the access denied error came from the inability to create the tbb file somehow?

  20. #14
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    The "tbb" file is not created when the number of iterations is "1" (test mode)

    I've just used
    Code:
    turbobench -elzma,9 DIR -R -i1 -j2 -I1 -J1
    and the "DIR.tbb" with total size is created a the end regardless of the "permission denied" message.

  21. Thanks:

    diskzip (5th June 2017)

  22. #15
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    Quote Originally Posted by dnd View Post
    The "tbb" file is not created when the number of iterations is "1" (test mode)

    I've just used
    Code:
    turbobench -elzma,9 DIR -R -i1 -j2 -I1 -J1
    and the "DIR.tbb" with total size is created a the end regardless of the "permission denied" message.
    Thank you.

    So that would run two tests?

    That would be prohibitively expensive time wise.

  23. #16
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    Quote Originally Posted by diskzip View Post
    That would be prohibitively expensive time wise.
    No, -j and -J options are for decompression, but the benchmark time is largely dominated by compression.

  24. #17
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    These guys really don't want to license their technology.

    I got the following response today in answer to my email sent a couple weeks ago:

    The cost for Oodle is $1million. Should I send you the license agreement and invoice?

    Mitch Soule

    RAD Game Tools

    425-893-4300 ext 101

    www.radgametools.com



    Someone needs to replace this guy at Oodle.

    He's a liability for the company.

    Or, perhaps, more likely its a one-man shop about to fold any time now.

    Amazing they lasted so long with customer support like this!

  25. Thanks:

    encode (8th June 2017)

  26. #18
    Programmer Bulat Ziganshin's Avatar
    Join Date
    Mar 2007
    Location
    Uzbekistan
    Posts
    4,572
    Thanks
    783
    Thanked 687 Times in 372 Posts
    their engine is for in-game data compression. if your intention was to put engine into your own program, the price is meaningful. once it's here, people will probabl buy your software to perfrom the compression rather than their own.

  27. Thanks (2):

    diskzip (8th June 2017),RamiroCruzo (8th June 2017)

  28. #19
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    Quote Originally Posted by Bulat Ziganshin View Post
    their engine is for in-game data compression. if your intention was to put engine into your own program, the price is meaningful. once it's here, people will probabl buy your software to perfrom the compression rather than their own.
    IMHO for a heavily open source LZMA derived algorithm, asking for a million dollars - without even providing a trial SDK - is saying **** you to the customer.

    They don't even have an end-user play, its not like DiskZIP is a tool which competes with their B2B business of selling compression algorithms in the first place!

    If people actually cared about compressing better, WinZip would have been long dead - sadly, they don't, and will keep using whatever is dog-fooded to them.

  29. Thanks (2):

    hunman (8th June 2017),RamiroCruzo (8th June 2017)

  30. #20
    Member
    Join Date
    May 2017
    Location
    UK
    Posts
    13
    Thanks
    1
    Thanked 3 Times in 3 Posts
    Quote Originally Posted by diskzip View Post
    IMHO for a heavily open source LZMA derived algorithm, asking for a million dollars - without even providing a trial SDK - is saying **** you to the customer.

    They don't even have an end-user play, its not like DiskZIP is a tool which competes with their B2B business of selling compression algorithms in the first place!

    If people actually cared about compressing better, WinZip would have been long dead - sadly, they don't, and will keep using whatever is dog-fooded to them.
    I agree it's annoying that their pricing prevents enthusiast use, but I should point out that LZNA/Kraken completely demolish the opposition in their designed use-case: fast in-memory decompression of game binary data (floating point, ints, etc.).

    The competition in high ratio fast decompressors that can be compiled into games is lizard(39/49), lzham, lzsse8(17) and zstd(22).
    Here is a test utilizing binary floating point data:
    Code:
    TurboBench:  - Thu Jun 08 14:32:51 2017
    
          C Size  ratio%     C MB/s     D MB/s   Name            File
        38130687    30.3       2.62      62.03   lzma 9                           test1
        38882540    30.9       1.44    3302.37   oodle 86 (Kraken Optimal)        test1    
        38893010    30.9       1.48     299.65   lzham 3                          test1
        38928526    31.0      15.18    1222.79   zstd 22                          test1
        38933920    31.0       1.48     292.93   lzham 4                          test1
       114776987    91.3       2.81      21.14   lzma 5                           test1
       125760771   100.0      20.94    7071.80   lizard 39                        test1
       125756930   100.0      37.73    6402.41   lzsse8 17                        test1
       125760771   100.0       4.24    6956.89   lizard 49                        test1

    Needless to say, lzma is completely infeasible for this purpose at 62 MB/s decomp and lizard & lzsse8 can't deal with the data at all. Zstd is the closest, but for a game developer the speed might be the difference between waiting 10 seconds for a level load vs 30 seconds, the latter probably deemed unacceptable. Alternatively, it might be the decision between having higher resolution art vs lower resolution if realtime memcpy-like decompression is part of the game engine.

    Not defending their pricing, but I must admit Oodle is a unique product that their customers make a sensible decision to buy given their objective function.

    That said, Hamid could destroy their business model tomorrow if he released a DLL of LZTurbo that people could incorporate into their games/programs:
    Code:
    TurboBench:  - Thu Jun 08 14:32:51 2017
    
          C Size  ratio%     C MB/s     D MB/s   Name            File
        38822993    30.9       7.17    2285.84   lzturbo 39                       test1
        38882540    30.9       1.44    3302.37   oodle 86                         test1
        41589073    33.1      98.26    5355.49   lzturbo 32                       test1

  31. Thanks:

    diskzip (8th June 2017)

  32. #21
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    Quote Originally Posted by djubik View Post
    I agree it's annoying that their pricing prevents enthusiast use, but I should point out that LZNA/Kraken completely demolish the opposition in their designed use-case: fast in-memory decompression of game binary data (floating point, ints, etc.).

    The competition in high ratio fast decompressors that can be compiled into games is lizard(39/49), lzham, lzsse8(17) and zstd(22).
    Here is a test utilizing binary floating point data:
    Code:
    TurboBench:  - Thu Jun 08 14:32:51 2017
    
          C Size  ratio%     C MB/s     D MB/s   Name            File
        38130687    30.3       2.62      62.03   lzma 9                           test1
        38882540    30.9       1.44    3302.37   oodle 86 (Kraken Optimal)        test1    
        38893010    30.9       1.48     299.65   lzham 3                          test1
        38928526    31.0      15.18    1222.79   zstd 22                          test1
        38933920    31.0       1.48     292.93   lzham 4                          test1
       114776987    91.3       2.81      21.14   lzma 5                           test1
       125760771   100.0      20.94    7071.80   lizard 39                        test1
       125756930   100.0      37.73    6402.41   lzsse8 17                        test1
       125760771   100.0       4.24    6956.89   lizard 49                        test1

    Needless to say, lzma is completely infeasible for this purpose at 62 MB/s decomp and lizard & lzsse8 can't deal with the data at all. Zstd is the closest, but for a game developer the speed might be the difference between waiting 10 seconds for a level load vs 30 seconds, the latter probably deemed unacceptable. Alternatively, it might be the decision between having higher resolution art vs lower resolution if realtime memcpy-like decompression is part of the game engine.

    Not defending their pricing, but I must admit Oodle is a unique product that their customers make a sensible decision to buy given their objective function.

    That said, Hamid could destroy their business model tomorrow if he released a DLL of LZTurbo that people could incorporate into their games/programs:
    Code:
    TurboBench:  - Thu Jun 08 14:32:51 2017
    
          C Size  ratio%     C MB/s     D MB/s   Name            File
        38822993    30.9       7.17    2285.84   lzturbo 39                       test1
        38882540    30.9       1.44    3302.37   oodle 86                         test1
        41589073    33.1      98.26    5355.49   lzturbo 32                       test1
    It has been my experience that companies proving this level of customer disrespect and disservice do a favor to no one - not their customers, certainly, and not themselves either, as they promptly leave all of their employees (or the single founder/employee, in this case) without a job.

    I would encourage Hamid to capitalize on this business opportunity without delay! It would only help him make money, help his customers get better service and support, and even help Oodle improve their service&support.

    That's how the free market economy works. I'm not a fan of savage capitalism, but there are things we can learn from it.

  33. #22
    Programmer Bulat Ziganshin's Avatar
    Join Date
    Mar 2007
    Location
    Uzbekistan
    Posts
    4,572
    Thanks
    783
    Thanked 687 Times in 372 Posts
    i can't agree more. if you are ready to spend 1m for a compression library, why you are gone to customer-irrespectful guys selling a mere lzma clone? instead, go to Hamid and buy original lzturbo algo with famous german customer support

  34. Thanks:

    diskzip (8th June 2017)

  35. #23
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    Quote Originally Posted by Bulat Ziganshin View Post
    i can't agree more. if you are ready to spend 1m for a compression library, why you are gone to customer-irrespectful guys selling a mere lzma clone? instead, go to Hamid and buy original lzturbo algo with famous german customer support
    The original quote of a million dollars was a sham anyways.

    Makes me wonder if these guys are scammers of some sort, if they are unable to provide an SDK for evaluation online.

    It could be they have some kind of open source liability issue. There's many systems nowadays which check for source code license violations, even based on just binary file analysis.

    It could also be that the technology just doesn't work, or has to be "tweaked" for each data set individually, to avoid things like extraction errors, etc.

    One could argue that they're concerned about piracy, but that would just be a testament to their extreme immaturity in the industry. Piracy is unavoidable, and to a certain extent, it may even help get the word out about your product.

    If Hamid needs any help marketing his solution, I'd be more than glad to help out.

    By the way, I asked the Oodle guy to provide me with one credible reference of a top-ten/a-list game that uses his technology. I told him I'd be glad to stand corrected.

    He replied as follows:

    Hah!
    Good luck.

    Is there any non-Indie production using this technology?

  36. #24
    Member
    Join Date
    Feb 2015
    Location
    United Kingdom
    Posts
    176
    Thanks
    29
    Thanked 74 Times in 44 Posts
    If I'm not mistaken one of RAD's tools (Bink) is used within Rockstar's Grand Theft Auto 5; a very popular title. Their website also contains companies which have deployed their product http://www.radgametools.com/binkgames.htm which is backed up by a quick Google search. I don't think they are being bad company by offering a quote outright, to me that just makes getting the information faster which some businesses might prefer if it was a serious request to use RAD's tools. I would like to benchmark RAD's tools as much as the next guy on here but it's an enterprise grade tool and they probably don't want programmers tampering with it in a public manner.

  37. Thanks:

    diskzip (9th June 2017)

  38. #25
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    4,013
    Thanks
    406
    Thanked 403 Times in 153 Posts
    Anyway, I can't beleive that "official spokesman" of RAD Game Tools could talk in such disrespectful manner.

  39. Thanks:

    diskzip (9th June 2017)

  40. #26
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    Quote Originally Posted by Lucas View Post
    If I'm not mistaken one of RAD's tools (Bink) is used within Rockstar's Grand Theft Auto 5; a very popular title. Their website also contains companies which have deployed their product http://www.radgametools.com/binkgames.htm which is backed up by a quick Google search. I don't think they are being bad company by offering a quote outright, to me that just makes getting the information faster which some businesses might prefer if it was a serious request to use RAD's tools. I would like to benchmark RAD's tools as much as the next guy on here but it's an enterprise grade tool and they probably don't want programmers tampering with it in a public manner.
    If it was a legitimate quote, that may have been something.

    There are kinder ways to decline business from customers without insulting their intelligence by quoting an exorbitant fee and declining an evaluation.

    This is something which clearly no customer in their right mind is going to accept.

    Besides, no customer would pay 1m for their technology, which fails to rise above open source commodity level.

    For most people, and companies, compression is perceived as being free, regardless of the incredible difficulties of getting it to actually work.

    Bink appears to be a video player alternative, unrelated to compression altogether.

    Adding 1+1 and connecting all the dots, I am inclined to believe their compression SDK does not work, period, or does not work as advertised.

    All this million dollar quoting is a poorly disguised effort to save face by what appears to be an incredibly unprofessional, one-person shop.

  41. Thanks (2):

    encode (9th June 2017),hunman (9th June 2017)

  42. #27
    Administrator Shelwien's Avatar
    Join Date
    May 2008
    Location
    Kharkov, Ukraine
    Posts
    3,982
    Thanks
    298
    Thanked 1,309 Times in 745 Posts
    1. Bink is an actual video format - see https://github.com/libav/libav/blob/...avcodec/bink.c
    Although its unrelated to oodle developers (afaik).

    2. It could be a legitimate quote. For example, see http://www.edg.com/faq/price .
    In which case it might include library sources and RAD having to support it forever, including fitting it for a specific game etc.
    Also keep in mind that they likely have to support platforms like PS3/PS4.
    So it could be actually interesting to see that license agreement.

    3. " incredibly unprofessional, one-person shop".
    According to this, they have 35 people: http://www.zoominfo.com/c/RAD-Game-Tools-Inc/97338754
    Also, the guy who answered your mail is the owner: http://www.zoominfo.com/p/Mitch-Soule/141476812

    Although really, yes, it most likely means that they're working with specific game studios,
    and don't need any new clients (unless you're willing to pay $1m at least).

  43. Thanks:

    diskzip (9th June 2017)

  44. #28
    Member
    Join Date
    Aug 2016
    Location
    Seattle
    Posts
    4
    Thanks
    0
    Thanked 10 Times in 3 Posts
    Lordy, ok, was ignoring this, since I don't think it should be on encode, buf since Simon (diskzip) is posting emails... Here you go: the entire email exchange before Simon exploded. They were four short mails:


    From: Simon King
    Sent: Wednesday, May 31, 2017 6:38 AM
    To: sales3@radgametools.com
    Subject: OODLE SDK


    Hello,


    I build and ship a commercial data compression application, competing
    with the likes of WinZip and WinRAR.


    Would you let me license your SDK to bring your compression formats to
    market as well?


    Alternately, please let me know if you prefer to write your own
    plug-ins for my platform.


    My platform is archive agnostic; everything is supported by plug-ins
    which are third party extensible.


    Thank you!
    - Simon King
    ============


    Mitch responds:


    Simon,


    Thanks for the interest in Oodle. I'm sorry but we don;t have
    resellers of our product.


    Mitch Soule
    RAD Game Tools
    ============


    [Pretty standard for us - at this point in our product cycle, we're not interested in 3rd parties selling our tech directly - we're just doing games and a few other verticals like disk storage]


    ============
    Simon responds:


    On Wed, Jun 7, 2017 at 1:30 PM, Simon King wrote:


    Hi Mitch,


    I am not looking for a reseller, I would be glad (and prefer) to buy
    your SDK from you directly.


    Simon King
    DiskZIP Computing
    ==========


    [At this point, Mitch comes to me and asks "what do you know about this company." I look online and there isn't much - even the Diskzip website has an empty about page. However, I also see THIS VERY THREAD! The first message, he says he found our DLL online and was asking if anyone had documentation. I mean that's just nuts - he's selling retail software, but he just copied ours off the web and then asked for docs - that's fairly crazy thing to do, if 5 days later you are going to ask us to get it...? I mean our programmers have been posting on this site since 2010 - we're going to see this admission at some point. In any case, despite being irritated, I see that Diskzip also does disk storage compression, which is an area we do sell into now (NAS's and such), so I said just quote him what you quote the NAS people.]


    =============


    Mitch responds:


    The cost for Oodle is $1million. Should I send you the license agreement
    and invoice?


    Mitch Soule
    RAD Game Tools


    ===========


    And that's it - that's the sum total of emails we traded with Simon until explosion time. I don't think any of those are rude - just a salesman doing his job, maybe he could have explained that cost, but at this point, I probably made Mitch think that it wasn't a real sale anyway. But I really don't think he was rude or anything, maybe brisk at the most.


    ==========


    Simon explodes:


    From: Simon King
    Date: Wed, Jun 7, 2017 at 1:46 PM
    Subject: Re: Fwd: FW: OODLE SDK
    To: Mitch Soule


    If you think I am going to invest a fortune without access to an
    evaluation SDK, you are very mistaken, Mitch.


    Do you have a supervisor or someone I can talk to at your organization
    to file a complaint regarding your intentionally offensive email?


    With customer service like this, I cannot imagine you guys staying in
    business for long.


    Unless my complaint is addressed, I will be posting this thread publicly
    and making this a problem for you until you get it right.


    Simon King
    DiskZIP Computing
    ==========


    So, Mitch was surprised by this, and shows it to me, and I'm like, "wow, he got hot" - but I don't see how anyone could see Mitch's mail as "intentionally offensive". So, I offer to email him:


    ==========
    Jeff emails:


    Mitch sent me your email chain - just to have a second set of eyes read it. I don't see anything rude - we just don't resell our codecs to be used in other compression products (for obvious reasons, it competes with ourselves). We have sold it to companies using it in a way that might compete with us, but those have been $1m licences, which when we tell people about, usually just ends the conversation.


    So, I'll say good luck and wish you well, but it doesn't sound like we're a fit.


    ->Jeff
    RAD Game Tools


    ===========


    I just figured, wish him luck, explain that I don't think we're a fit (especially after such an angry mail), and we move on. Well...


    ===========


    Simon replies:


    Hi Jeff,


    We are a VC backed end-user company, in a completely different vertical
    (rather, horizontal), as Mitch could have found out for himself.


    We don't compete with your product as we don't build or license
    compression algorithm SDKs for software developers.


    That said, there's kinder ways to tell off customers you don't want to
    do business with, for starters, I would think.


    If you expect me to justify an investment of a million dollars without
    having access to so much as an evaluation SDK, you're very deluded.


    Expecting me to believe that you've sold any license at that price point
    without an eval insults your intelligence more than mine.


    And that's the kindest way I could put it for you.


    Have a great rest of your day, Jeff!


    ============


    So, I figure ok, kind of a shitty mail, but we're probably good, and I'll clarify on the eval thing:


    ============


    Jeff replies:


    We would have done an eval once you were comfortable with the price - that's how it always works. But the cost was just to give you an idea of what the licensing is outside of games, because that usually a dead stop for most people (understandably).


    Regardless, good luck.


    ============


    Figured that would be the end. I mean this is a fairly low-key misunderstanding as far as internet misunderstanding go. Well...


    ============


    Simon replies:


    How can I know what your product is worth before I've even seen it?


    To me it seems like you are a one man operation (nice try on the
    aliases) with a dubious product you cannot even put online (probably
    because none of your advertising claims are justified).


    Show me one top ten publisher using your "SDK" (which probably doesn't
    even exist), and I'll be glad to stand corrected.


    ==========


    At this point, it's so beyond silly. I mean, even the most cursory of web searches tells you about RAD - I mean we've been around 29 years <grin>, and I don't need to justify anything to a stranger about our business. So, I figure he's just trolling, and I should just get out of this:


    ==========


    Jeff replies:


    Hah!
    Good luck.


    ========


    And that's it. Everything else he has posted here happened after that, since I guess he wanted to follow through on his original threat of "making this a problem for you until you get it right", which I still don't even understand.


    So, there you go - I don't think Mitch was rude, maybe too all business, and don't think I was rude. But hey, read this and make your own decision, I'm not going to post here about this again.


    Anyway, I'll say it a fourth time: good luck, Simon.

  45. Thanks (5):

    encode (9th June 2017),JamesB (9th June 2017),jibz (9th June 2017),Mike (9th June 2017),schnaader (9th June 2017)

  46. #29
    Member
    Join Date
    Dec 2011
    Location
    Cambridge, UK
    Posts
    506
    Thanks
    187
    Thanked 177 Times in 120 Posts
    Thanks for clarifying things.

    Perhaps the email that started this rant was a bit abrupt, but it's still totally within your rights to conduct your business as you see fit and one of those is obviously protecting yourself from direct competition.

  47. Thanks:

    diskzip (9th June 2017)

  48. #30
    Member
    Join Date
    May 2017
    Location
    Australia
    Posts
    126
    Thanks
    92
    Thanked 32 Times in 21 Posts
    Quote Originally Posted by Shelwien View Post
    1. Bink is an actual video format - see https://github.com/libav/libav/blob/...avcodec/bink.c
    Although its unrelated to oodle developers (afaik).

    2. It could be a legitimate quote. For example, see http://www.edg.com/faq/price .
    In which case it might include library sources and RAD having to support it forever, including fitting it for a specific game etc.
    Also keep in mind that they likely have to support platforms like PS3/PS4.
    So it could be actually interesting to see that license agreement.

    3. " incredibly unprofessional, one-person shop".
    According to this, they have 35 people: http://www.zoominfo.com/c/RAD-Game-Tools-Inc/97338754
    Also, the guy who answered your mail is the owner: http://www.zoominfo.com/p/Mitch-Soule/141476812

    Although really, yes, it most likely means that they're working with specific game studios,
    and don't need any new clients (unless you're willing to pay $1m at least).
    I don't lend much credence to reports of companies having X number of employees, Y amount of revenue, etc. unless the information is verifiable.

    Needless to say, you can post anything online. Sites like zoominfo (or many others) are often editable by the companies themselves, where they volunteer the information in question directly.

    You cannot really take any of that information at face-value unless you want to be misled. The only way to know for sure is if the company is public. Unless a company is public and filing reports required by the government, you can never be sure.

    And even then, we know how many public companies lie in their reports

Page 1 of 2 12 LastLast

Similar Threads

  1. new compressor LZNA = "LZ-nibbled-ANS" - "Oodle 1.45"
    By joerg in forum Data Compression
    Replies: 22
    Last Post: 19th February 2018, 04:50
  2. Replies: 92
    Last Post: 14th September 2016, 16:42
  3. Intel Releases A Boatload Of Haswell Documentation
    By Bulat Ziganshin in forum The Off-Topic Lounge
    Replies: 1
    Last Post: 1st January 2014, 02:37
  4. Eugene D. Shelwien is missing.
    By kampaster in forum The Off-Topic Lounge
    Replies: 36
    Last Post: 23rd July 2013, 18:53
  5. Missing options
    By in forum Forum Archive
    Replies: 3
    Last Post: 20th May 2006, 23:19

Posting Permissions

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