Note: as this a work in progress consider it to be a rough draft.
I use the RF5.2mm F2.8 L Dual Fisheye on an R5. Assume the RF-S3.9mm F3.5 STM DUAL FISHEYE works the same, other than the R7 lens is APS-C vs full frame and the R7 version has autofocus.
It has a very steep learning curve. Have confirmed that get good results with stills but am working on getting the same quality with video. This requires going through a ton of Canon camera settings, some of which are obscure. I.E. you don't even see a setting unless you have a specific setting for something else. The resulting video is not editable in DaVinci, Final Cut or Premiere Pro until you convert it to rectangular side by side in the Canon EOS VR utility. Here are just as few of the settings I am using right now:
Camera Settings (R5)
8K DCI Has an aspect ratio of 256:135 and 8192 x 4320 pixels. 8K DCI is a cinematography standard and is commonly used in movie projection.
Shoot 8KD All-I, IPB, or IPB Light to save storage space
AF-2 - MF Peaking Settings on, level high, color of choice (if grayed out turn off Zebra Camera)
Focus guide on
F stop 2.8 when setting focus if you can. Hit magnify button to get best focus.
Shoot 5.6 or above for best quality.
Hit info to check focus on right eye. Use hex wrench to adjust lenses if one is out of focus.
1.5m distance is stereo sweet spot
AF-2 - MF peaking settings - may turn off after achieve focus
Manual exposure & Clog gives best dynamic range, image quality and film like color
Camera-1 setting - 8KD & All-i "compress each frame one at a time"
EOS VR Utility:
Keep the metadata text files or CRM video files will not be seen.
raw quick development
set processor
set white balance
raw allows setting ISO
Cinema gamut
Gamma - camera log setting if used, otherwise DCI
denoise 60% however better to denoise in NLE such as FCP, DaVinci, Premiere or Topaz
turn on horizontal correction
turn on lens mask
remove chromatic aberration
export 100%, all metadata
VR utility parallax, lens correction
export 3d 180 maximum resolution
view in skybox
Denoising
Denoising in Final Cut Pro
Apply denoise effect to clip
Denoise in Topaz
Adjustments/Codecs set to H265 for 2.8 mm otherwise get the error:
the output resolution of the video is too large for the H264 encoder. Max size: 7,680x4,320
RF-S3.9mm F3.5 STM DUAL FISHEYE is 6960 x 4640 so h264 like will not work either
(can take a lot of memory - 20 GB or more)
1 3/4 hours for 1 minute of video
Table of Topaz processing times - Process/benchmark to get tables
Topaz Video AI v5.5.0
System Information
OS: Mac v15.0101
CPU: Apple M2 Ultra 192 GB
GPU: Apple M2 Ultra 144 GB
Processing Settings
device: 0 vram: 0.5 instances: 1
Input Resolution: 3840x2160
Benchmark Results
Artemis 1X: 04.15 fps 2X: 02.73 fps 4X: 00.71 fps
Iris 1X: 02.22 fps 2X: 01.09 fps 4X: 00.53 fps
Proteus 1X: 01.96 fps 2X: 02.32 fps 4X: 00.71 fps
Gaia 1X: 01.35 fps 2X: 01.02 fps 4X: 00.60 fps
Nyx 1X: 00.70 fps 2X: 00.95 fps
Nyx Fast 1X: 01.59 fps
Rhea 4X: 00.28 fps
4X Slowmo Apollo: 03.98 fps APFast: 18.94 fps Chronos: 01.26 fps CHFast: 01.99 fps
16X Slowmo Aion: 05.96 fps
may need to use a .mov container rather than .mp4
References
Hugh Hou has good instructional videos. Here are some of the videos I am still trying to process:
https://www.canon-europe.com/pro/infobank/video-formats/
https://support.usa.canon.com/kb/s/article/ART177388
https://mikewilley.com/blog-posts/mastering-8k-video-recording-on-your-canon-eos-r5
https://medium.com/@portemantho/can...best-practices-in-eos-vr-utility-ad9564ebda3a
https://support.apple.com/guide/fin...o-editing-workflow-ver3d865d8a4/11.0/mac/14.6
Besides the camera and workflow technical issues there a ton of creative issues you have to consider when actually taking the videos. If shooting on a tripod then you need to use an extender so that the tripod feet don't show, use an omni-directional microphone, etc.
I hope that I will eventually be able to get higher quality with the 4K Canon videos than the 1080p spatial videos that the iPhone produces. iPhone spatial videos are obviously much, much simpler to make. One has to decide whether it is worth the time investment to get the higher 8K (4k x 4k) quality.
I use the RF5.2mm F2.8 L Dual Fisheye on an R5. Assume the RF-S3.9mm F3.5 STM DUAL FISHEYE works the same, other than the R7 lens is APS-C vs full frame and the R7 version has autofocus.
It has a very steep learning curve. Have confirmed that get good results with stills but am working on getting the same quality with video. This requires going through a ton of Canon camera settings, some of which are obscure. I.E. you don't even see a setting unless you have a specific setting for something else. The resulting video is not editable in DaVinci, Final Cut or Premiere Pro until you convert it to rectangular side by side in the Canon EOS VR utility. Here are just as few of the settings I am using right now:
Camera Settings (R5)
8K DCI Has an aspect ratio of 256:135 and 8192 x 4320 pixels. 8K DCI is a cinematography standard and is commonly used in movie projection.
Shoot 8KD All-I, IPB, or IPB Light to save storage space
AF-2 - MF Peaking Settings on, level high, color of choice (if grayed out turn off Zebra Camera)
Focus guide on
F stop 2.8 when setting focus if you can. Hit magnify button to get best focus.
Shoot 5.6 or above for best quality.
Hit info to check focus on right eye. Use hex wrench to adjust lenses if one is out of focus.
1.5m distance is stereo sweet spot
AF-2 - MF peaking settings - may turn off after achieve focus
Manual exposure & Clog gives best dynamic range, image quality and film like color
Camera-1 setting - 8KD & All-i "compress each frame one at a time"
EOS VR Utility:
Keep the metadata text files or CRM video files will not be seen.
raw quick development
set processor
set white balance
raw allows setting ISO
Cinema gamut
Gamma - camera log setting if used, otherwise DCI
denoise 60% however better to denoise in NLE such as FCP, DaVinci, Premiere or Topaz
turn on horizontal correction
turn on lens mask
remove chromatic aberration
export 100%, all metadata
VR utility parallax, lens correction
export 3d 180 maximum resolution
view in skybox
Denoising
Denoising in Final Cut Pro
Apply denoise effect to clip
Denoise in Topaz
Adjustments/Codecs set to H265 for 2.8 mm otherwise get the error:
the output resolution of the video is too large for the H264 encoder. Max size: 7,680x4,320
RF-S3.9mm F3.5 STM DUAL FISHEYE is 6960 x 4640 so h264 like will not work either
(can take a lot of memory - 20 GB or more)
1 3/4 hours for 1 minute of video
Table of Topaz processing times - Process/benchmark to get tables
Topaz Video AI v5.5.0
System Information
OS: Mac v15.0101
CPU: Apple M2 Ultra 192 GB
GPU: Apple M2 Ultra 144 GB
Processing Settings
device: 0 vram: 0.5 instances: 1
Input Resolution: 3840x2160
Benchmark Results
Artemis 1X: 04.15 fps 2X: 02.73 fps 4X: 00.71 fps
Iris 1X: 02.22 fps 2X: 01.09 fps 4X: 00.53 fps
Proteus 1X: 01.96 fps 2X: 02.32 fps 4X: 00.71 fps
Gaia 1X: 01.35 fps 2X: 01.02 fps 4X: 00.60 fps
Nyx 1X: 00.70 fps 2X: 00.95 fps
Nyx Fast 1X: 01.59 fps
Rhea 4X: 00.28 fps
4X Slowmo Apollo: 03.98 fps APFast: 18.94 fps Chronos: 01.26 fps CHFast: 01.99 fps
16X Slowmo Aion: 05.96 fps
may need to use a .mov container rather than .mp4
References
Hugh Hou has good instructional videos. Here are some of the videos I am still trying to process:
https://www.canon-europe.com/pro/infobank/video-formats/
https://support.usa.canon.com/kb/s/article/ART177388
https://mikewilley.com/blog-posts/mastering-8k-video-recording-on-your-canon-eos-r5
https://medium.com/@portemantho/can...best-practices-in-eos-vr-utility-ad9564ebda3a
https://support.apple.com/guide/fin...o-editing-workflow-ver3d865d8a4/11.0/mac/14.6
Besides the camera and workflow technical issues there a ton of creative issues you have to consider when actually taking the videos. If shooting on a tripod then you need to use an extender so that the tripod feet don't show, use an omni-directional microphone, etc.
I hope that I will eventually be able to get higher quality with the 4K Canon videos than the 1080p spatial videos that the iPhone produces. iPhone spatial videos are obviously much, much simpler to make. One has to decide whether it is worth the time investment to get the higher 8K (4k x 4k) quality.