|
|||||||||
|
Thread Tools | Search this Thread |
January 4th, 2010, 09:20 PM | #1 |
Inner Circle
Join Date: Jun 2003
Location: San Jose, CA
Posts: 2,222
|
SxS vs Nanoflash stills: Elephant Seals
Here are the 1080p30 stills of the same frame from a parallel recording using EX1 and Nanoflash (100 MBit Long GOP). The Nanoflash video looks more "real", and if I fixate on the area below the elephant seal's right flipper, the SxS video appears to have a screen door overlay where the water is quite blocky. I see very little if any such blockiness in the Nanoflash MXF video, and this is what I paid for. The Nanoflash is a good EX1 enhancement for complex scenes as the detail is better, though this scene probably needs more than 100 MBits/second to be indistinguishable from uncompressed video to this eye.
Index of /Videos/NanoTest_ESeal These are crops that have been enlarged 2x using nearest neighbor algorithm in Photoshop CS4, with EX1 on the left and Nanoflash on the right. http://www.gentlemensfightingclub.co..._R=Nano_2x.png http://www.gentlemensfightingclub.co..._R=Nano_2x.png Below are the frame grabs. I'm learning that the EX1 is really very good. Last edited by Gints Klimanis; January 4th, 2010 at 10:49 PM. |
January 4th, 2010, 09:33 PM | #2 |
Inner Circle
Join Date: Jun 2003
Location: San Jose, CA
Posts: 2,222
|
EX1 35 MBits/second on left, Nanoflash 100 MBits/second on right.
I don't know how control the size of the attachments. They seem to show up too small if you don't click on them. So check original files. Last edited by Gints Klimanis; January 4th, 2010 at 10:51 PM. |
January 5th, 2010, 03:52 AM | #3 |
Major Player
Join Date: Jun 2005
Location: Paris, France
Posts: 201
|
Thank you Gints, this is very instructive.
Macro blocking is clearly visible on the SxS cropped picture. The skin of the Elephant Seal is also more detailled on the Nano pictures (look at his nose). However I think this kind of test would be even more eloquent if the whole picture is moving during a pan or a travelling for instance.
__________________
http://www.songesdemoai.com/ |
January 5th, 2010, 02:37 PM | #4 |
Regular Crew
Join Date: May 2009
Location: Monument, CO
Posts: 109
|
I like seeing these side-by-side comparisons, thank you very much for that!
As an extension of this idea I would really like to see side-by-side comparisons after color grading to see the differences there. We know in theory that the Nanoflash footage should hold up better but it would be interesting to see how well the picture holds up in heavy post work compared to the native SxS footage. |
January 5th, 2010, 07:55 PM | #5 |
Inner Circle
Join Date: Jun 2003
Location: San Jose, CA
Posts: 2,222
|
Here's another set of 100% crops zoomed by 2x. Click on the images to show them at full size. To my eyes, it's easier to see the improvement in detail at 100%. The zooms I've provided just don't seem to convey the same increase in detail I see on the frame grabs when they are viewed in full resolution on a 1920x1200 24" LCD monitor. As before, EX1 on Left and Nano on Right.
Last edited by Gints Klimanis; January 5th, 2010 at 09:08 PM. |
January 5th, 2010, 09:07 PM | #6 |
Inner Circle
Join Date: Jun 2003
Location: San Jose, CA
Posts: 2,222
|
Check out the last two frames of chroma noise in the shadows, shot at 1/60 second at f/1.9 with either +6 or +12 dB Gain. I've added 100% crops from the master image, followed by a lightened image. Does the Nano reproduce the noise more faithfully or amplify the noise? As before, EX1 on Left and Nano on Right.
|
January 5th, 2010, 09:37 PM | #7 |
Inner Circle
Join Date: Jun 2003
Location: San Jose, CA
Posts: 2,222
|
Here's another where I don't see much of a detail difference other than some light compression echo artifacts and a some color difference on the edges of the whiskers. That's ok. Most of the frame is out-of-focus sand. which is reduced in contrast by the visible breath of the seal in the cold air.
Even so, after this frame grab experiment, I will always shoot with my Nanoflash. Life is too short to record visible macroblocks. |
| ||||||
|
|