Personal View site logo
Make sure to join PV on Telegram or Facebook! Perfect to keep up with community on your smartphone.
Please, support PV!
It allows to keep PV going, with more focus towards AI, but keeping be one of the few truly independent places.
Driftwood VY Canis Majoris: This thread will now be integrated into 'Driftwood Series 6 Settings'
  • 901 Replies sorted by
  • @beavis Ah, do you mean the black horizontal section contained in the bottom third? It was a little deceptive because it could be tough to tell if it was detail or noise. For the green channel, the other areas of the image were more representative and there was generally more detail contained in CM Night than in Sedna AQ1 A.

    But I am glad that the originals made it clearer. It's been tiring shooting and creating the test images. :)

  • @thepalalias thanks again for this huge comparsion, it´s very rare and much appreciated.

  • @beavis Flattery will get you... even more test pictures. :)

    These were made from the same images specified in my earlier post ( http://www.personal-view.com/talks/discussion/comment/59978#Comment_59978 ) except that I included 4 extra this time.

    They are in a different order (though the first image is still the JPEG shot in stills mode without eletronic shutter) and have all been processed with maximum contrast and maximum gamma as PNG files. Note the huge difference in the color cast of the shadow areas.

    Control JPEG.png
    1920 x 1080 - 2M
    Sedna_AQ1_A0000.png
    1920 x 1080 - 2M
    CM_Night0000.png
    1920 x 1080 - 2M
    Mysteron_Burst0000.png
    1920 x 1080 - 1M
    CM_SkinTone_Soft0000.png
    1920 x 1080 - 1M
    CM_Day_PM0000.png
    1920 x 1080 - 1M
    Sedna_AQ1_C0000.png
    1920 x 1080 - 1M
    Mysteron0000.png
    1920 x 1080 - 1M
    ClusterV10000.png
    1920 x 1080 - 1M
    Dark_MatterV10000.png
    1920 x 1080 - 1M
    Dark_MatterV20000.png
    1920 x 1080 - 1M
    Dark_MatterV30000.png
    1920 x 1080 - 2M
    EOSHD_Vanilla_440000.png
    1920 x 1080 - 1M
    FlowMotion1.110000.png
    1920 x 1080 - 1M
    GOLGOP3-13V4-130000.png
    1920 x 1080 - 1M
    Quantum_V9c0000.png
    1920 x 1080 - 1M
    Sanity_V50000.png
    1920 x 1080 - 1M
    Stock1.10000.png
    1920 x 1080 - 1M
  • @thepalalias Nice test.
    It certainly appears Canis Majoris Night edges out over Sedna A AQ1. Hmmm....

  • @proaudio4 Thanks. What I am struck by is not just how the blocks sometimes appear finer, but the overall coherence within given solid areas. It really helps that the red channel appears to be so much less noisy.

  • @thepalalias "It really helps that the red channel appears to be so much less noisy"

    This is exactly the news I like to hear!

    Have you checked how it holds under hard daylight conditions? I have not had a chance to check out Canis Majoris Night.. looking forward to it.

  • @proaudio4 Outdoor testing is exceedingly difficult to control for when there is 3 minutes between shots. At best you can compare a handful of settings before everything changes. :)

    I did shoot some limited tests with CM Night vs one or two other settings yesterday, but with the processing I used at the time, the differences were a lot more subtle than in today's testing. I may revisit the footage and see what I can do, now that you mention it.

  • So so far CM Night is appearing to be the best settings? Does this also work for day time conditions or only just night. I'm planning to shoot a 1-2min short in the subways and wondering which is the best settings to go with, Sedna A AQ1 or CM Night, or AM/PM.

  • @proaudio4 @thougts2uk Here is Canis Majoris Night vs Mysteron vs Sedna AQ1 A on a rather bright (but still cloud covered) day.

    No processing. Shot with 20mm Pancake, at Smooth -2, -2, -2, -2. Focused at lens' minimum focusing distance.

    Note: It was windy - don't judge sharpness on any objects that might potentially be in motion.

    Canis Majoris Night0000.png
    1920 x 1080 - 2M
    mysteron0000.png
    1920 x 1080 - 2M
    Sedna AQ1 A0000.png
    1920 x 1080 - 2M
  • @thepalalias that was a lot of work and darn informative...thanks!

  • @peternap Thanks! I am looking forward to the next sunny day when I can check to see if I have the same experience with the color of the sky as you did. As you can see, the sky was rather unhelpful when I shot yesterday. :)

  • @thepalalis Thanks for all the comparisons.

  • @thepalalias..Wow did you do all those tests with 1 gh2. Man that's a lot of work!! Good job.

  • Thank you, indeed, thepalalias.

  • @thepalalias Thanks mate! I can see small benefits in Canis but I think we reached the point of diminishing returns. Most people couldn't tell the difference between Canis and Sedna without amplifying the footage in photoshop. But it is good to have choice. I think a good day and a good night setting fill my needs.

  • @Kihlian @rumi @cosimo_bullo @sohus You're very welcome. :)

    @sohus Canis Majoris Night really comes into its own in dark lighting. I was frankly surprised by how big the improvement was. This is one of the first times I didn't need to do an enlargement to really catch the differences. While a less demanding shot (like the outdoor one I posted above) might not show much difference between the settings (even with a great deal of processing) the more demanding shots really do.

    Look at the image quality delta between Quantum v9c and Stock 1.1 in the comparison above. Then look at the image quality delta between Canis Majoris Night and Quantum v9c. The difference between the latter two may be as great as between the first two. The improvement over Sedna AQ1 A is still quite evident, but not to the same extent.

    EDIT: In other words, the low light performance MAY have come further in just over three and a half months since mid-January, than in the 6 months prior from the time VK first released the GH2 hack in July. Only additional testing can confirm or disprove this initial perception.

    The issue isn't that we have hit a point of diminishing returns, it's that frequency of updates has been so high that we sometimes lose sight of the progress. @Driftwood has posted at least 27 new .ini files since Quantum X V1 came out on 2012-02-18. Dark Matter, Orion, Sedna and Mysteron are much closer to each other in low light performance than any of them are to Canis Majoris Night. The silent progress in the month Mysteron appears to have allowed some rather noticeable improvements.

    Things have come a long way since you assembled the starter pack in October, don't you agree? :)

  • @thepalalias I agree. In no way I tend to downplay the results and efforts of @Driftwood! The difference will never be as big as the first GOP1 patch by him though ;)

    How do you think Canis night performs in daylight compared to Sedna? In other words, if the night results are better, and the day results are as good, can we use night as allrounder?

    I think it is time for a new starter pack soon. Times have changed indeed :)

  • @sohus I haven't shot a sufficiently demanding daylight scene to gauge it yet. Here is the video comparison used for the exterior shots above. Cake vs Mysteron vs Sedna AQ1 A vs Canis Majoris Night

    Password: gh2

    What was @Driftwood's first GOP1 setting? I joined in September and I've been looking through the archives to catalog such things, but there are big gaps in my understanding before mid-November.

  • @thepalalias

    Many thanks for your huge quantity and excellent tests! The matrix difference was found well. I am also looking forward to the test with GOLGOP3-13_SednaAQ1C. If it is only a test in darkness like this time, a better result may be brought by what combined AQ1A. However, since the codec of GOLGOP3-13_SednaAQ1A wavers slightly on a bright scene, I have recommended AQ1C. :-)

  • @thepalalias The first setting was the 220 or 240mbit GOP1 setting. I don't have it in my archive anymore.

  • Is CM Night targeted for any particular ISO range? Does it expect higher ISOs for its noise characteristic or would it do just as well at ISO 160? At night, we often have warmer tungsten lighting. Is it tuned for this or would it be expected to do just as well under cool moonlight?

  • @bmcent1 Regardless of how it was targeted, note that the tests I have shown in this thread so far are at ISO 3200 and ISO 200, both of them with preset (rather than custom) WB.

    The daylight test used daylight/sunny WB. The dark indoor test used incandescent WB. Originals have been provided of each.

    @bkmcwd Great, we'll see how it performs. :)

  • J'ai un gros problème sur Canis majoris et même beaucoup d'autre, Orion, Sedna ect...certain ont un problème de son, sur le téléviseur, le son Crack et saute :/ ma carte est une Sandisk 64gb 95mbs, je suis désespéré :(

  • I have a big problem on Canis Majoris and even many other, Orion, Sedna ect ... some have a sound problem on the TV, the sound jumps :/ Crack and my card is a Sandisk 64gb 95mbs, I 'm desperate: (

  • @ Driftwood I shoot today some footage in the same low light condition that I shoot few weeks ago with dark matter, but now with Canis Night. WOW! Much better low light performance, can also push much more the colour grading! No span problems (sandisk 95 Mbts) on 50p, and that all at slight lower bitrate. Donated again! Thanks!

This topic is closed.
← All Discussions