[WINDOWS] MLV to DNG Batch Converter v1.9

Started by tonybeccar, January 27, 2014, 05:08:15 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

eightcore

Amazing tool by the way! I already integrated it in my workflow. A question: Could you implement the specific feature which allows to see the progress in the taskbar?

Example:

tonybeccar

Quote from: eightcore on February 09, 2014, 03:16:20 PM
Amazing tool by the way! I already integrated it in my workflow. A question: Could you implement the specific feature which allows to see the progress in the taskbar?

Example:

Hehe.. it IS implemented already.. but made it only available for Windows 7 because of a lot of errors I was getting for other Windows OS's... I can set it up also for windows 8 IF it works the same way ;) will add it to next release.. Currently I am trying to make proxies perfect..

chmee

@tonybeccar
the magic for Premiere is to cut the blacklevel-bias and recalculate the upper bound to 12bit. Blackmagic-DNGs are 12bit/logarithmic, the Bolex-D16-Files are 12bit as well, but it seems, linear (or without a linearisation-table). In my converter the partymode-files work quite good with 12bit/linear inside Premiere CC.

and clearly saying: 12bit is another world than "only" 8bit, so losing ~2bits (compared to original canon-data) is not a dealbreaker.

regards chmee
[size=2]phreekz * blog * twitter[/size]

tonybeccar

Quote from: chmee on February 09, 2014, 03:55:00 PM
@tonybeccar
the magic for Premiere is to cut the blacklevel-bias and recalculate the upper bound to 12bit. Blackmagic-DNGs are 12bit/logarithmic, the Bolex-D16-Files are 12bit as well, but it seems, linear (or without a linearisation-table). In my converter the partymode-files work quite good with 12bit inside Premiere CC.

and clearly saying: 12bit is another world than "only" 8bit, so losing ~2bits (compared to original canon-data) is not a dealbreaker.

regards chmee

If you had just a command line I could incorporate it here... and direct MLV support would save lots of time also...

chmee

:) i just didnt found the time to write a commandline-version. sorry.
(mlv is working in raw2cdng - but actually without audio. tested once again yesterday)
[size=2]phreekz * blog * twitter[/size]

tonybeccar

Quote from: chmee on February 09, 2014, 04:13:42 PM
:) i just didnt found the time to write a commandline-version. sorry.
(mlv is working in raw2cdng - but actually without audio. tested once again yesterday)

Hehe, please!! I am doing this because currently I am on a holiday... if not, then I don't know.. :) MLV is working but with the .raw step in between right?

naturalsound

Hi,

i recorded a lot of MLV footage. Now I am going to update my current PC to cope with those huge data rates and the preprocessing that hast to be done.

My question (especially for your converter): Will I benefit from a powerful graphics card like Geforce GTX 770 or is it just the CPU that matters?
I think at least ACR can use the GPU, can it?

I suppose a SSD for the footag is a must? Or will a 7200rpm be good enough? (max. 1 video + transitions in PP CS5 maybe PP CC later...)

Markus


Markus

Quote from: tonybeccar on February 09, 2014, 03:12:21 PM
Oh wow REALLY? 8bit are you SURE? Been using lossy, made comparisons with HEAVY color correction and they looked the same... :S I don't have my big monitor though... only my crappy laptop screen. Lossy is 25% of the original
It is 8bit but it is optimized so that the 8bits are distributed were the information is as I understand it.

tonybeccar

Quote from: naturalsound on February 09, 2014, 05:33:28 PM
Hi,

i recorded a lot of MLV footage. Now I am going to update my current PC to cope with those huge data rates and the preprocessing that hast to be done.

My question (especially for your converter): Will I benefit from a powerful graphics card like Geforce GTX 770 or is it just the CPU that matters?
I think at least ACR can use the GPU, can it?

I suppose a SSD for the footag is a must? Or will a 7200rpm be good enough? (max. 1 video + transitions in PP CS5 maybe PP CC later...)

I don't think my converter will take advantage of that.. (I am not 100% sure though... you should check with mlv_dump, mlv2dng, adobe dng converter, ffmpeg and dcraw).. BUT anyway, you are moving to raw, you MUST have a powerful graphics card to process the data efficiently! Resolve, or speedgrade, or any NLE will thank you for it!

tonybeccar

Quote from: Markus on February 09, 2014, 05:36:07 PM
It is 8bit but it is optimized so that the 8bits are distributed were the information is as I understand it.

I just converted some lossy dngs and opened one dng in Photoshop... it says 16-bit at the bottom part of ACR... you sure Markus?

Markus


"7. MadManChan2000, Jan 11, 2012 5:44 AM   in reply to Fuchsteufel Report
The color components of the stored & compressed image data is 8-bit depth.

The linear scene-referred data (usually 16 bits after demosaicing) goes through a 1D perceptual mapping, then is quantized to 8 bits, then JPEG compressed. The DNG spec will not require these steps (in fact the only new part is a new supported Compression tag value, and some additional tags to deal with the downsampled case), but it is what we do to ensure high-quality results.  An opcode (from the existing DNG 1.3 spec) is embedded into the file; this is used to invert the 1D mapping when reading in the image, thereby returning the stored image data to linear scene-referred encoding.

Yes, the demosaic step generally expands 1 color channel to 3.  But, the quantization to 8 bit (after perceptual mapping) and JPEG compression more than makes up for that.

For example, a 5D Mark II full raw file is ~27 MB with lossless DNG compression (same with .CR2).  Lossy compressed DNG (full-res) is ~13 MB depending on scene content, ISO, etc.

The two other advantages of this setup are:

1. It makes it simpler to define downsampled (i.e., "small raw") versions of the image.  For example, the same 5D Mark II raw file described above can be saved as a 10-megapixel lossy compressed DNG (7 MB) or just screen-sized, like 2048 pixels on the long side (1.5 MB).

2. It makes loading images much faster, primarily because it skips the early-stage raw processing (linearization, demosaic, etc.)."

http://forums.adobe.com/thread/947446

Malex

Quote from: tonybeccar on February 08, 2014, 05:00:31 PM
About Dual Iso, g33g0 told me that the file header doesnt contain any information about it.. so there will not be an automatic detection for now.. when I implement it, you'll have to select which files are dual and which now... will be tricky and will have to expand the UI a lil bit I guess...
Hi Tony ,in order to select manually which footage is dual iso, maybe the thumbnail could be bigger, so we could notice the lines and use a special dual iso checkmark for it?  What do you think?

pilgrim

Hi Tony,

Thank you very much for your apps. It's great.
A question arise : is the folder "Luts" for the LUT files ? Which file will be use ? .CUBE ?
The progress bar is really very useful. It's nice also the GUI, with the future function so we start to dream the road map  ;)

tonybeccar

v1.8!!

- DRAMATIC increase in speed while extracting DNGs. Using mlv2dng, and it is like 10 or 20 times faster... amazing stuff. BUT, the audio is still extracting with mlv_dump... and the basterd (with all due respect g33g0) takes forever to extract only the audio! :S But overall the process IS much faster. If anyone has any ideas on how to extract the audio another way, awesome, share.
- Ffmpeg is faster too! Changed the version. Last one was 4-6fps for me, this one is 15-20fps. Using threads:8 by default.
- Proxies should be rendering PERFECTLY. No drop frames, no duplicate frames, etc. SYnc sync sync.
- Changed to ufraw, and no, it's not faster at all. But managed to get the colors looking ok, neutral for now, next version gonna add the "Lut" selection, so the user can input their own ufraw preset. Right now is just a curve I made.
- Fixed BUGs related with proxy creation, for all the users that were skipping, this should fix it!

https://www.dropbox.com/s/vp61d1urllc3we7/MLV%20Converter%201.8.zip

I think it's all. Been researching for the proxies... and well it seems that they will NEVER be the same as the dngs processed by ACR in AE. NEVER. Above all if we create a flat looking proxy, then you wont be able to push the colors as much.

Changing the topic, can someone OBJECTIVELY explain me WHY is it better to have a FLAT (lifeless) looking video? Seriously, I understand all the theories about the retained dynamic range, shadows and highlights.. but everytime I do a test, it's logical.. in order to achieve a normal level of saturation, (or a little more), you have to reeeally push the colors.. and in the end you have artifacts and the image breaks. That doesnt happen AT ALL if you do that color correction with the inicial file.. so, am I missing something? (BTW, I tried with prores HQ)

Anyway, I think the proxy thing with ffmpeg will ONLY be useful for its word, proxy. You'll have to go back to the original dngs for a final output. Let's hope AME comes out with something good.

Cheers!

chmee

Quote from: tonybeccar on February 09, 2014, 11:53:48 PM
OBJECTIVELY explain me WHY is it better to have a FLAT (lifeless) looking video? .. and in the end you have artifacts and the image breaks. That doesnt happen AT ALL if you do that color correction with the inicial file..
Most common argument is, dont do that with 8bit-videodata - 10, 12 or 16bit is not only because of highlight-recovery or shadow-pushing, but mainly because of more headroom for these postpro-work.

regards chmee
[size=2]phreekz * blog * twitter[/size]

arrinkiiii

Version 1.8 start to extract audio and then it say finish but only extract 1 file in the folder, no dng's... and i think this MLV file i haven't recorded with  audio.

arturochu

Chu

tonybeccar

Quote from: arrinkiiii on February 10, 2014, 12:34:27 AM
Version 1.8 start to extract audio and then it say finish but only extract 1 file in the folder, no dng's... and i think this MLV file i haven't recorded with  audio.

my god! sorry!!forgot to unpack mlv2dng... uploading now

tonybeccar


tonybeccar

Quote from: chmee on February 10, 2014, 12:06:06 AM
Most common argument is, dont do that with 8bit-videodata - 10, 12 or 16bit is not only because of highlight-recovery or shadow-pushing, but mainly because of more headroom for these postpro-work.

regards chmee

Well yeah that's my point. It's ok for the highlight recovery of course.. but that headroom, really kills the color richness.. anyway, maybe it's not the thread to talk about it.. Just to try and push the colors to the same level, one is great and the other one is crap :S

arrinkiiii

It works good now, thanks Tony.

One thing that i notice is wend it start to convert, the first thing the app do is to extract audio... In my laptop for 300 frames it take kind of 37 seconds for extract the audio and then start the conversion from MLV to DNG... In the end i don't have any audio file in the folder of the DNG's (i think i didn't have recorded audio with this MLV files that im testing)

tonybeccar

Quote from: arrinkiiii on February 10, 2014, 02:01:07 AM
It works good now, thanks Tony.

One thing that i notice is wend it start to convert, the first thing the app do is to extract audio... In my laptop for 300 frames it take kind of 37 seconds for extract the audio and then start the conversion from MLV to DNG... In the end i don't have any audio file in the folder of the DNG's (i think i didn't have recorded audio with this MLV files that im testing)

Ok, so I have to disable for non-audio mlv files.. I wonder if I can :S. BUT, do you think it's faster than extracting everything with mlv_dump still? mlv2dng does it in the blink of an eye...

ted ramasola

trying out 1.8.1 Tony,

when i select one of the clips it gives "cannot find FORMAT.DLL" . I have reinstalled the plugins from irfanview but still the same error. no thumbnail is generated. am i missing something?

It is faster though than previous versions.
5DmkII  / 7D
www.ramasolaproductions.com
Texas

Panz

Forgot i had an acc here. First post, hi everyone!

I don't think using mlv2dng is good, because i think stuff like vertical band correction and so on is not implemented(maybe i am wrong, but i ve been watching things for some time and i believe
that development on mlv2dng stopped before that). Possibly other things are missing.

Best solution imo would be to use chmee's awesome program, because of the metadata you later get on cdngs(not able yet to have davinci read metadata from dng files from mlv_dump)
Still, even chmee's app haven't implemented vertical stripes yet, so better to use mlv_dump for now(quality-wise)

@Chmee, i thought i read a post(correct me if wrong) where you said that raw/mlv2cdng isn't important anymore coz of premiere and resolve importing dngs natively. I d like to disagree, your
app is very useful, especially if it can tranfer the metadata from mlv correctly(lens info, framerates and stuff).

THANK you both for your time and apps, they are really useful!