Menu

Show posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Show posts Menu

Messages - 50mm1200s

#76
Quote from: a1ex on July 05, 2018, 09:54:11 AM
We will need an icon for this new menu.

What else to put in it? (possibly moved from other menus)
- Benchmarks
- File manager
- Shutter count
- Screenshot?
- Temperature? (used by astro folks and not only)
- Lens info? (I've never used it other than for debugging purposes)
- Overclocking?
- Warnings for bad settings? (that should be mostly a set-and-forget item, maybe even baked in the defaults somehow)
- ... ?

+1 for: File manager, Screenshot, Benchmarks and Overclocking.
#77
@dford would be nice a shot from a concrete wall,  with lots of detail :)

A question for the people working on this: since the algorithm seems to work by blending, wouldn't be better to denoise it first? I see some chroma noise on dford's shots that might affect the precision of the final blend...
#78
Quote from: ibrahim on July 03, 2018, 01:20:19 AM
Thanks for the shared video.

Could you explain what's not so good with canon's bokeh vs the older lenses. Do you mean that the old ones have a rounder shape than canon's?

I am planing of selling all my canon lenses and save money to buy the zeiss milvus for drama filmmaking in 2.35:1 on my 5d3. From what I have learned these primes are fantastic lenses for filmmaking but quite expensive.  :D

Well, I'm not an optics specialist, don't quote me on what I'm about to say, but here's my perspective. Each lens has its own "personality", that is, a blend of some variables that constitutes the formed image:
- MTF (resolution and contrast)
- Chromatic Aberration
- Distortion
- Vignette
- Flare

Canon lenses are very good, but generally they have this very diffuse light transmission. What I mean by that is that they try to replicate the human eye features and create a very smooth depth of field, high contrast and low focus breathing (the relationship of focus between objects). From my personal perspective, the image generated is very good for technical shots, such as architecture, but boring as an artistic choice. Also, they try to avoid lens distortion as much as possible, and that may be very good for movies.
That's why some of the most expensive lenses on the market (Cooke s5/i, Panavision G-series, etc) have what can be considered defects by academics. Cinematographers have to make artistic choices and these choices are not based on the most correct academic perspective, it's based on "look and feel" you want people to perceive in this exact story you want to tell.
Let me give you an example: many cinematographers, like Dan Laustsen, don't like to shoot on lenses that are too sharp, because they say the actors skin needs to be kinda soft while on screen or else it looks ugly. In the case of Laustsen, he even uses a Black Promist filter to soft the image even more. These cinematographers have the money to use the most sharp lenses out there ("diffraction-limited", like Zeiss Otus), but they choose not to.

I myself did many tests about this. I use old lenses of years now and when I get on my hands some newer lenses from Canon I always am dispointed by the result.
I made this test to show you what I mean: The first image is from a old Komura 135mm f/2.0, wide open. The second is the same lens, wide open, but with a oval filter, trying to simulate the anamorphic 'feel':



ps.: sorry for any grammatical issues... english is not my mother language.
#79
Quote from: masc on July 02, 2018, 03:06:13 PM
Hmm... it supports not really... it only supports adding the metadata, but that does not mean bt2020 is inside. You'll need a matrix conversion inside the ffmpeg command. But that won't "bring" more quality, because the input & encoding is still the same. There is this colormatrix=... command - this one works in 8bit - so you'll loose more than you'll get. You also can adjust colorspace via scale filter, this one works with higher bitdepth. Atm we convert and export to bt709 in this way. Otherwise you can export with AVFoundation, if you don't like bt709, or you must color correct in another software if thats not enough - you're right.

Ok, I get it now. Thanks.
#80
Quote from: Seruji on July 02, 2018, 02:45:01 PM
Filmmakers
¿Any tip for someone who has a 5D MK2 (mlv 2.0) and wants to shoot a shortfilm in raw and then post-edit in Adobe Premiere? In terms of LUT's order..if it's more important to apply a LUT in MLVAPP or export (proRes 4444) and apply the LUT in Premiere?

¿Which LUT for general color-grading do you guys apply? MLVAPP or in a video editing software (such as Resolve, Final Cut, Premiere..). I have dozens of LUT's, but don't know when to apply them, or if there are some very common on this community.

If you're going to apply a LUT, do it directly on MLVApp if possible. But, you might find MLVApp limited for color grading, since once it's encoded you can't change it (destructive workflow). So what I do is this:
- Import on MLVApp. Put the clips on Alexa Log-C. Fix exposure and WB. Export to ProRes 444
- Import on Premiere. Use Logarist to convert from LogC to Rec.709.
- Do color grading with Lumetri.

This way you have a somewhat solid color grading. I sometimes use ImpulZ LUT's (the "FPE" type, that is, the film color, but not density), but they have many issues on dark midtones (some color shift towards magenta).

You would get better results using a CinemaDNG on Resolve if you will be doing heavy color grading...
#81
@masc since we're already talking about encoding, about our discussion some time ago on color management... no improvements? I see advanced color management can't be done with ffmpeg, but at least it supports bt2020 out of the box...
#82
Quote from: bouncyball on July 02, 2018, 12:33:54 PM
It is not 12bits, it's 8 bit (indeed 4:2:0)

Well, exiftool is giving me wrong informations:


ExifTool Version Number         : 8.03
File Name                       : M21-1816.avi
Directory                       : .
File Size                       : 113 MB
File Modification Date/Time     : 2018:07:02 07:09:15-03:00
File Permissions                : rw-rw-rw-
File Type                       : AVI
MIME Type                       : video/avi
Frame Rate                      : 23.976
Max Data Rate                   : 4.857e+004 kB/s
Frame Count                     : 57
Stream Count                    : 1
Stream Type                     : Video
Video Codec                     : I420
Video Frame Rate                : 23.98
Video Frame Count               : 57
Quality                         : Default
Sample Size                     : Variable
Image Width                     : 1568
Image Height                    : 882
Planes                          : 1
Bit Depth                       : 12
Compression                     : I420
Image Length                    : 2074464
Pixels Per Meter X              : 0
Pixels Per Meter Y              : 0
Num Colors                      : Use BitDepth
Num Important Colors            : All
Software                        : Lavf57.82.101
Duration                        : 2.38 s
Image Size                      : 1568x882


ffprobe don't seem show the bit depth:


Metadata:
    encoder         : Lavf57.82.101
  Duration: 00:00:02.38, start: 0.000000, bitrate: 398007 kb/s
    Stream #0:0: Video: rawvideo (I420 / 0x30323449), yuv420p, 1568x882, 405004 kb/s, 23.98 fps, 23.98 tbr, 23.98 tbn, 23.98 tbc
#83
Quote from: Enrico on July 02, 2018, 01:41:11 AM
Is there any tool which convert AVI from MLVapp in the kind of AVI supported by premiere?

ffmpeg or Handbrake. You can easily install ffmpeg through Chocolatey (if you work on windows) or Homebrew (if on mac).

Quote
I record with a 700D in 10bit but are you saying that the AVI from MLVapp is only 8 bit? I don't understand, I'm so confused

No, uncompressed AVI from MLVApp seems to be 12bit, but the chroma subsampling is 4:2:0.
ProRes and DNxHR are nearly lossless, have subsampling of 4:4:4 and are 10-bit. Just use these... please. There's no need to complicate.
If you really want a lossless workflow, MLVApp is probably not what you're looking to use. Check Resolve, in this case.
#84
Quote from: Jeremy A. on July 01, 2018, 11:05:26 AM
I use mlv_dump to extract images from mlv files. Isn't that CinemaDNG already?

Not really. mlv_dump just 'extracts' individual dng's. The CinemaDNG format has multiple dng's, already in a sequence, with correct FPS and metadata. You can use MLVFS or MLVApp.

Quote
I use DNxHR HQX 10bits...

DNxHR is good. Keep it. Here's the guideline from youtube, for the new accepted format.
#85
Quote from: Enrico on July 01, 2018, 01:29:56 PM
But actually After Effects exports AVI uncompressed when it imports sequences of DNG or TIFF. Maybe the issues regard the codec, I'm not very knowledged in these things.

Well, just because AE exports, it doesn't mean it will open on PP. There's no codec on uncompressed AVI, it's just a container (.avi) with raw data. You should just use ProRes 444 or convert it as we said already.
#86
Quote from: Enrico on June 29, 2018, 09:21:23 PM
Firstly I'd like to thanks the developer of this great programme, anyway I have some issues when I export raw videos in uncompressed avi becouse in premiere it doesn't play them, instead everything works with the default windows 10 video player. Any possible solusion or even suggestion in the workflow?

AVI uncompressed doesn't seem to be supported by Premiere Pro. You'll have to convert it to something like Cineform or Lagarith, if you want lossless, or you could export directly to ProRes 444, inside MLVApp...
#87
Great job. I got this Ubi sunt feeling.

Positives:

- Screenwriting is good. Liked how it uses the rain as a symbol of sadness (pathetic fallacy)
- Soundtrack is pretty good.
- Actors are good. The kid did great.

Some critics (as you would expect from a technical forum):

- The bokeh is too "Cannon'ish". I think you would benefit very much from some old pieces of glass. I've been using old manual lenses for years now and I can say they do a great difference on the look. A Helios 44-2 with a low-end anamorphic lenses (like a Anamorphot-50), would make the images more interesting.

- Most scenes inside the house are underexposed. This is not wrong 'per se', but I think it would be better to overexpose the scene by +1.5 f-stops and then on postproduction make it darker, if it's an artistic choice. This way you preserve more information, less noise and have more flexibility on color grading.

Quote
the film was selected in a festival in Greece, but received no price

Keep your head up. I'm sure your next one will be even better. Ganbatte.
#88
Cool. The transitions are neat, although a bit overused. Some scenes have too much sharpening and there's this yellow cast on the highlights, but overhaull, it's nice. The monkey scene at 1:43 is the best (should make a wallpaper of it).

Quote
I know that pros go through...

Don't bother too much going to what "pros" say. You're doing nothing too wrong ;)

Quote
- Transformation of MLV into DNG.

You could use CinemaDNG instead. It's faster and better to organize.

Quote
- Color calibration of clips in Davinci Resolve (curve, exposure, contrast, white balance, etc..).

Why not do the color grading inside Resolve? I don't work too much with Resolve yet, but the less you transcode your footage, the better. See this.

Quote
- Export clips in 10 bits.

Is it in ProRes 422? Don't export it on H.264, or you'll loose more information. Also, see if you can work on Rec.2020 and 10-bit, this is the new upcomming standard (as I said in other posts, youtube is accepting this format now, as Vimeo Pro is too).
#89
Quote from: 70MM13 on June 28, 2018, 02:48:29 PM
I wish that this amplifier optimization would be taken seriously.

Indeed. I think this could be an option in the ISO menu. Maybe run a brutal force script to get the best DR and SNR regs for each camera model...
@dford compiled a version of 50D for me with the adtg branch and I can say it's stable enough to be on mainline.
#90
By the way, for anyone interested in more DR, check out Digicon Filter.
#91
Quote from: 70MM13 on June 28, 2018, 05:13:49 AM
http://drive.google.com/file/d/12-SEc4Xqd_n1ZzhtDXtyVzOHYt4YMABM/view?usp=sharing

File is way too big. Can you compress it? ProRes 422 or even 20MB/s h.264 would be nice.
Also, what are your reg values? Would be good to share on ADTG thread. With ISO changes and this could we finally beat Alexa? :p
#92
Quote from: Kanakori on June 28, 2018, 05:38:33 AM
Ok the dual iso problem seems solved but a new one came about...
How can I take that off?

Try the "Highlight Reconstruction" feature. If it doesn't work, maybe someone here could point a solution, but I would say you'll need to blow the highlights until the artifacts are gone (using curves).
#93
Nice! I will test and post some samples this weekend.
#94
Quote from: Danne on June 27, 2018, 08:55:39 AM
That idea is already working. A1ex put in a feature (HDR_extended) which normally is disabled. Been there for a while but forgotten. I had it reenabled in this branch:
https://bitbucket.org/Dannephoto/magic-lantern/branch/crop_rec_4k_mlv_lite_snd_sd_uhs_HDR

What camera are you using? If you want to test this concept I could give you a build and help you with steps in menu.

Oh, really? Wow. I'm using a 50D, but I also have a 600D too. If you can build it to me, that would be awesome!
#95
Quote from: Danne on June 27, 2018, 08:25:37 AM
I have a hard time understanding the deeper insights about that concept(if there are any)

The idea seems to be the same as with the old HDR feature in ML, but instead of using ISO change for switching between exposures, use shutter speed.

Quote
About ghosting in general. In my mind some heavy programming or maybe even some tweaking with hugin stuff might yield better results on moving subjects but thing like calculation movement inside footage is beyond my knowledge.

I think movement calculation wouldn't be necessary, maybe a averaging technique...

Quote
Putting this in mlv_lite module? I think a1ex or g3gg0 are the ones to answer this.

Yeah, but both are already working on other projects and no one is willing to make this idea work. Wish I could program it myself, but I have no real knowledge of programming. The idea is good, though.
#96
@Danne do you think this idea could be implemented using ffmpeg (without ghosts)? You seem to be more and more into ffmpeg than most people here. Maybe @dford could implement it on MLV module?
#97
Quote from: bouncyball on June 26, 2018, 09:26:48 AM
This second clip (first can not be downloaded) has very unusual WB out of the box (maybe some complicated mixed lighting?). I even straggled to white balance it manually.

Yes, mixed sodium (yellow) light with 5500K softbox. I first use AutoWB from ML on the gray card, and then do the precise WB in post. Tint is always a problem for me. Here's the other file:
https://we.tl/x59qxrDqHl

And here's a frame without the gray card (same ambient light):
https://we.tl/vpnSuESe93
#98
In case you guys want to test in a real gray card, here's two samples. These are precisely 17.8% reflectance material (according to the company):
https://we.tl/t45SiodapN
https://we.tl/7WNCIdIJnp

I think the idea of getting WB from skin tones is good if you don't have a gray card, but the 'correct' will always be to get a neutral balance so the colorist can work through it...
#99
Quote from: Danne on June 24, 2018, 12:46:33 PM
Temperature picker! Works great :)


Question. Could the picker stay in picking mode until picking the button again? Now one do one pick, has to go back to the picker button, push it test another white balance area, go back to the button etc. If it stays on we can pick until satisfied.

Nice! This will be really useful for me. Thanks.
#100
Quote from: reddeercity on June 11, 2015, 01:10:54 AM
(can & dose cause Banding etc.... in video)

From what I know, optics don't cause banding. A Polarizer can cause many optical issues (chromatic aberration, distortion, etc), but I don't think it causes banding. If you have any document proving me wrong, just link it here, please.