Messages

......
- DIY / 3D printing folks around the world are making face shields and donating them to hospitals. Consider helping them.
.......

Yeah but to do that you need a very good 3D printer and good printing skills. I own a prusa i3 rework built a few years ago (pieces printed from other 3d printers and CNC milled alluminium frame, should be the second version made) and even if it prints good is not that good as these guys' printers

.......
- There are several suggestions of COVID19-related open source projects, on the GSoC mentors mailing list (unfortunately not public). I can summarize if there is interest.
.......

Folding@Home is a simple way to help by mostly doing nothing at all. You install the software and let the network spread. Many of us are videomakers/photographers and have a quite good hardware, no doubt it could help.

..... U.S having the worst spread it seems.

I live in Italy  almost everything is closed here, except for a few industries (like the one i work in).

Can you share some of the tricks with us? 

Is there any information which can be useful especially to ML RAW users?

The first thing is about not using luts, there are tools (like stills) that are a lot more powerful. The biggest difference is that a LUT "cuts" the signal above and under, while still doesn't.
Then I learned the use of every node and also mixer and splitter node.

The biggest thing that helped me on ml raw is how to manage caching and the color space (input/output/viewer and so on) in the settings.

Then I knew the general workflow to be really faster, like using flags, smart bins, shared nodes, stills managing, caching and optimized media, etc..

It would be really too long to write everything in one post, maybe an entire forum would be needed for that.

Pheraps the fastest thing that made me be more precise is using the 10bit mode for the color picker (even if you have a 10bit monitor, like I have, your eyes can't see what davinci could) abd the thing that made me be faster is lightbox and stills (you can use lightbox to mark your clips with a flag, then correct just one clip, take a still of that clip and paste the still over the similar clips, then use the split window to match with another clip, take still of the new clip and paste over similar clips, then repeat again untill you corrected all the clips, then a couple of shared node and secondary is done).

Okay here is what I'm talking about:

Both images I used the dropper to find the white balance, and both I pushed up the shadow .67.

Resolve looks like absolute garbage....why?

That's because you're using log wheels e not primary bars (learn to use primary bars instead of primary wheels), did you maybe use a LUT? That would be an orribile mistake (at least on the first node)

There's a reason that makes Davinci Resolve a perfect software for color correction and grading and that's his "YRGB color space".

Not many YouTubers really know how to properly use DaVinci correctly, find a BMD trainer and ask for a course (I did it, if you work with videos it's worth the investment, I learned a lot of tricks that made me save hundred hours of worktime)

You should first take a look at google's support page where it describes the optimal rendering parameters before uploading.

Then you also have to check legal values for luma and saturation.

77d is a digic 7 camera.
Digic 6 porting is in progress slowly as there are many changes in the architecture of processor, instructions and many other things that are still to be figured out.

Probably after porting a digic 6 camera it will be easier to Port a digic 7 camera but there's no release date of course.

If you want to help you can, but at this stage you'll need a digic 7 camera, coding skills, arm processor knowledge and A LOT of time

Hi dfort, tell me where I can find your build and I'll test (I need a build cause I lost everything on my PC. Faulty disk )

Esiste qualche guida su come muovere i primi passi per un porting su 750d?

Beh inizia facendo un dump del firmware poi il primo passaggio è fare il boot in qemu.

Fatto ciò sei a buon punto, bisognerà risolvere i problemi di cache, per farlo però, una volte che sei riuscito, dovrai testare direttamente se funziona in camera perchè qemu, per quanto sia un ottimo emulatore, non riesce a riprodurli.



Benvenuto

PS se fai un dump passamelo che provo anche io

The first clip is a little noisy it's true but not too much, i see moirè and shaky footage probably due to the stabilizer on when on a tripod,

in the second clip the sharpness is bad due to high diffration at f23, if you try the same shoot at f11 you get better results.

after that i think a lot has been lost in youtube compression (it's a sad story but true, youtube kills videos). Try a diffrent compression ad upload to a better straming service (for quality) as vimeo.

i think you fighting against the aliasing filter in 5d3

Avoid putting too many MLVs in the same folder or many folders under mlvfs Scan base directory. When refreshing scans every file so probably means a overflow somewhere o too much cpu work

Hi guys,
I'm here as i written (if i can say that about a vb.net application ) a veeeery simple gui for mlv_dump.

It's very simple, just install it, open a .mlv file, choose output dir and the parameters you want to apply (there's also a little preset possibility, for example i apply 2x2 chroma smoothing to my 700D to get rid of focus pixels)

It's nothing will make you scream but.. It works and makes mlv_dump easyer for who doesn't know how to use command line

IT'S JUST FOR WINDOWS

you can donwload the installer here

If you want to help (next thing after i will have moved to my new house will we re-write this using a framework like ultimate++ to make it cross-platform and implement mlv_dump natively) just check the project on here

It's still alpha release. Help me to discover bugs.

@levas no, as Alex tested a few weeks ago the compression would be the same as a 10bit stream would give the same results of a 14 bit as it's used as a 10bit with leading zeros. So it's not worth to use a 10bit stream with this method.

Just I'd like to know how I can help to Port this on other cameras and if this would get rid of focus pixels on many cams like 700D. Would be nice to know if it's the same routine to compress and delete focus pixels

So how can we help to Port this on other cameras?

Probably just spam but no, it's not compatible

just increase the light outside. 600D doens't like high iso too much, already at 800 you can see noise... better shoot in better light. Probably using 1600 ISO gets better result overexposing a little

(like shooting ETTR)

why don't you try to work with proxies?

Simple rule: what you see in lv is what you record and what your PC will see and everybody will see

In live view you see what you get. The fact you see Just a Little of the whole screen is because what is visibile is what you get.

There's no "uncrop", when you record cropped Just a part of the sensor is recorded, while lv takes all the sensor. So se put Black bars to tell you "Hey what's behind the Black is not recorded"

So my interest is in short movies and i own a 700D

i tought to go with prime lenses for many reasons.

First of all i've got time, i'm not in a run&shoot where i have no time to change the lens (like in a wedding or any event)
Then i'll be forced to use that focal lenght so i would be learning better i think (but someone may not agree).

The last reason is that one day i could go FF (5d3 or 6D, or even 5d4) so investing in EF lenses instead of EF-S would be better but i also know that lenses for full frame are not so good on APS-C sized sensors...

I would have bought the 24mm pancake but i'm worried about the possibilty to use the follow focus on my rig as if it's like the 18-55mm STM lens the focus belt may not attach properly (focus ring too thin and the 24mm would be too close to the camera body for my follow focus)...

So do you think that a tamron 17-50 2.8 may be the best solution for me? The tamron allows me to keep all my vitals

The SD controller on the camera is limited. For example 700D couldn't write faster than 40MB/s, the limit for 60D is about 20MB/s

You're not italian so it doesn't matter, it was a joke about the "new" first citizen of Rome. Just don't care

Hi guys,
i'm here because i'd like to expand my set of lenses for my 700D

as i don't earn a lot of money every month  i have to correctly choose what are the best lenses i could get.

i'd like to go for prime lenses (wide aperture and many other things) and i'm here to ask you to share what you do use most in your videomakeing activities

of course there is no one simple solution to everything as there are people making short movies, others making events (like weddings and so on), others for documentaries etc..

so what are the lenses you use most and for what?


i'm asking that as i'm interested in buying a wide angle for my aps-c cam but still can't decide if it's better to spend on a samyang (rokinon/whatever it's called in your country ) 14mm or the 16mm for making short movies in both open area and indoor


So.. share you're lens kit, what are your favourite and why can be useful for many other guys here

do you mean double exposure? a few cameras like 5d3, 70d and 1dx can already do it (outputting a jpeg of course)

Nice vid, i like this type of editing even if i think that in a long video may seem confusing

Se sei di Roma non farlo vedere alla Raggi che lo ruba e non ti da nemmeno un euro (come agli artisti di capodanno )

Sorry guys if i'm taking so long but i'm really in a short of spare time.

I'm still trying to understand what's the big difference between digic 4 and digic 5 for raw_slurp

Io to now i think i just discovered hot water as i'm trying ti figure out how Memory is dinamically allocated in Canon firmware. As a1ex said a few posts ago probably the raw buffer address is a variabile inside a structure. My Guess is that in digic 5 this struct is statically allocated while in digic 4 this is dinamically allocated via a Memory lib (for garbage collection? With debugging symbols and assert management i think)
This should also explain why in digic 5 se get a pointer while in digic 4 we get a pointer to a pointer.

What i don't understand is why the offsets from the struct base address are so different from camera to camera.

Is it possibile to dump a certain part of RAM while running? From address X get N bytes?

Why not Just hardcore using compiler directives for every ml camera?