an early xmas present for 5d3 owners from TNT!

Started by 70MM13, December 19, 2019, 05:30:56 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

70MM13

every magic lantern user (stills AND video) needs to watch this:

https://youtu.be/Pvi4lwgKT9s

with your help, every magic lantern camera can do this!

...and more!

we've only scratched the surface.




Audionut

What on earth is going on in the highlights?  Why does the sensor response curve have a brick wall in the highlights?

70MM13

that's the perfect name for it, since we are replicating a brick wall limiter...  of sorts ;)

Audionut

So you are intentionally causing pixels at highlights to have less SNR?

Why would you want some brick wall filter other then saturated pixel wells?

70MM13

why is the domain of...

have you seen my videos?

Audionut

Yes, I have.  In some other video you show how adjusting the gains affects the highlights.

In the video above you describe the effect of 0xFE as boosting DR by 0.5EV (in the shadows).  You then go on to talk about trust in the codebase, even though the developer describes a lack of trust, and that reducing analog gain only improves things by about 0.1 EV.

From what I can tell, the "tnt builds" simply (no disrespect to the dev) expose easier adjustment of the already found and tested regs.  This is great.  Spamming the forum with videos, and links to said videos all over the place, with claims that the tnt build can give 0.5EV better shadow noise.........

Look at the SNR curve you posted above.


Talk to me about what is going on in the highlights with the adjustments you have made.  Why are there a line of highlight pixels going below 2EV?  Why is there a big hump in the sensor response curve near the highlights?

70MM13

do you really think that a graph (of dubious accuracy according to a1ex himself) generated by a build you don't know the origin of, using a version of the raw diag you don't know the origin of, under conditions you don't know about is of more value than the ACTUAL videos i am creating and sharing with my methods-- not the howto that is made with my phone, but actual real video...  something that some of us actually do around here...

some of us are helping others to want to explore and innovate, and others are simply discouraging users and generating noise.

yourboylloyd

The dynamic range indicator could be a huge help itself. I'm not sure what it is you're doing or how you're doing it, but pleaseeeee bring this to 5d2!
Join the ML discord! https://discord.gg/H7h6rfq

histor

Quote from: yourboylloyd on December 23, 2019, 07:57:34 PM
but pleaseeeee bring this to 5d2!

Already done: adtg_gui.mo works on all Reddeercity's builds for 5d2 + raw_diag.mo (portable module from Alex)
Some fast settings to start with - from my previous posts. Not consider as spamming  ;)

And the whole CMOS/ADTG/Digic register investigation on ISO thread to study.

Audionut

Thread locked due to blatant misinformation.

The OP has also previously labeled results from the shadows taken at ISO 1600 as being taken at ISO 100 and refuses to apply scientific method or follow advice regarding accurate test procedures.

It is ok to be excited with developments, it is not ok to spread misinformation and not follow advice.




Quote from: yourboylloyd on December 23, 2019, 07:57:34 PM
The dynamic range indicator could be a huge help itself. I'm not sure what it is you're doing or how you're doing it, but pleaseeeee bring this to 5d2!

Available in all recent builds with raw based exposure feedback.
https://www.magiclantern.fm/forum/index.php?topic=12096.0