More > JRiver Media Center 31 for Windows
NEW: JRVR Configuration Profiles
jmone:
Nice - I just did a quick test creating a "High Frame Rate" profile for material where "[FPS]=>30" and to use FSRCNNX 8 instead of 16. Works as it says on the tin! This will be great for mid range GPUs where FSRCNNX 16 works well on "film" type material but can't keep up with 50/60 Material.
Next is to test a different profile for Interlaced Material to see what I can find to reduce combing artifacts for these.
jmone:
Is there any existing MC Field that reliably indicates if a video is Interlaced or Progressive? I see in my SOT Media Info output that there is:
Scan type : Interlaced
Scan order : Top Field First
I'd rather use an existing MC Field(s) to make the determination but else I'll just add a New MC Field called "Scan Type" that SOT can populate.
Hendrik:
I don't think the database has that info, but I can probably add a custom data entry for it - at least in some circumstances. If it gets deinterlaced before JRVR sees it, it would have no clue.
jmone:
Yeah I could not see anything in the existing DB fields that may work. No probs - I'll just get the info from SOT--> Media Info to populate a MC Field that would then work in all circumstances.
jmone:
OK - so that all worked as expected. I've added new MC Database Field called "Scan type" that is then populated with "Interlaced" or "Progressive" from SOT. I've setup a JRVR Rule called "Interlaced" under Scaling that is activated by the rule "[Scan type]=[Interlaced]". The "Scaling Profile" then has the following set:
- Scale in Sigmoidal Light = Unchecked
- SuperRes Enhancement = Unchecked
- Advanced Scaling / Luma Image Doubling = Disabled
The resultant image is certainly very soft, but the artifacts (mostly from FSRCNNX or RAVU) is greatly reduced. No a great image but better than adding sharpened artifacts.
I know that FSRCNNX has options for processing Interlaced material, so I'm now wondering in JRVR's implementation if it is always set to "Progressive"?
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version