This is my original article written for "Digital Production" (Munich) in German:
https://drive.google.com/file/d/1cMYncu ... sp=sharing
For translation I recommend DeepL.
You may distribute it unaltered, but please no translations without my consent.
The benchmark is here. I used a few very demanding functions, which nevertheless may be needed regularly, like Depth Map and Fog (18 only), massive NR and Speed Warp.
https://www.dropbox.com/s/nig23ohng2gpu ... e.drp?dl=0
It's using demanding sources, 10 seconds each, like:
is.gd/arri_sample_footage (very high data rate, but easy to decode)
is.gd/bmd_sample_footage (we all know these, don't we?)
is.gd/red_sample_footage (wavelet RAW is hard to decode, but supported by an SDK for CPU and GPU)
is.gd/sony_sample_footage (Sony Venice at relatively high data rate)
is.gd/tony_r5_download (in case you don’t have a Canon R5 around, 8K RAW by Tony Mellinger)
And finally some 8K HEV out of Canon R5 with the highest data rate:
https://drive.google.com/file/d/1Z8y7Il ... sp=sharing
The short URLs are not seen as links by the forum, so copy/paste.
An Asus ProArt laptop with a Ryzen 9 CPU 5900HX and the Nvidia RTX 3070 mobile GPU with 8 Gbyte VRAM is rendering it into CineForm YUV 10 bit in 28 minutes 7 seconds (but crashing if trying to go straight to HEVC or DNxHR).
The Mac M1 Pro (details in my sig) does it in 26:25 for CineForm or with very similar numbers for the rest.
As a goody, a river scene for adding some fog with Depth Map:
https://drive.google.com/file/d/1dJzt_e ... sp=sharing
https://drive.google.com/file/d/1cMYncu ... sp=sharing
For translation I recommend DeepL.
You may distribute it unaltered, but please no translations without my consent.
The benchmark is here. I used a few very demanding functions, which nevertheless may be needed regularly, like Depth Map and Fog (18 only), massive NR and Speed Warp.
https://www.dropbox.com/s/nig23ohng2gpu ... e.drp?dl=0
It's using demanding sources, 10 seconds each, like:
is.gd/arri_sample_footage (very high data rate, but easy to decode)
is.gd/bmd_sample_footage (we all know these, don't we?)
is.gd/red_sample_footage (wavelet RAW is hard to decode, but supported by an SDK for CPU and GPU)
is.gd/sony_sample_footage (Sony Venice at relatively high data rate)
is.gd/tony_r5_download (in case you don’t have a Canon R5 around, 8K RAW by Tony Mellinger)
And finally some 8K HEV out of Canon R5 with the highest data rate:
https://drive.google.com/file/d/1Z8y7Il ... sp=sharing
The short URLs are not seen as links by the forum, so copy/paste.
An Asus ProArt laptop with a Ryzen 9 CPU 5900HX and the Nvidia RTX 3070 mobile GPU with 8 Gbyte VRAM is rendering it into CineForm YUV 10 bit in 28 minutes 7 seconds (but crashing if trying to go straight to HEVC or DNxHR).
The Mac M1 Pro (details in my sig) does it in 26:25 for CineForm or with very similar numbers for the rest.
As a goody, a river scene for adding some fog with Depth Map:
https://drive.google.com/file/d/1dJzt_e ... sp=sharing
Now that the cat #19 is out of the bag, test it as much as you can and use the subforum.
Studio 18.6.6, MacOS 13.6.6, 2017 iMac, 32 GB, Radeon Pro 580
MacBook M1 Pro, 16 GPU cores, 32 GB RAM and iPhone 15 Pro
Speed Editor, UltraStudio Monitor 3G
Studio 18.6.6, MacOS 13.6.6, 2017 iMac, 32 GB, Radeon Pro 580
MacBook M1 Pro, 16 GPU cores, 32 GB RAM and iPhone 15 Pro
Speed Editor, UltraStudio Monitor 3G