Profile

Cover photo
François Simond (supercurio)
Works at SpectraStudy
Lives in Chambéry, Savoy, France
4,754 followers|6,961,605 views
AboutPostsPhotos

Stream

Pinned
 
First article written for http://SpectraStudy.com is live!

Android Camera2 API promises to revolution digital imaging on smartphones or altogether with advanced capture and processing capabilities that have never been accessible to third party applications before.

Some of what becomes possible is manual controls, computational photography, RAW #DNG  capture, full control over video recording, custom image processing. Limits are few.

Here's an analysis on how much of this new API is supported by the #Lollipop  devices released during #MWC15 , including the highly anticipated Samsung Galaxy S6 and HTC One M9 flagships.
Since Google announced Android 5 Lollipop at Google I/O 2014, mobile photography and video enthusiasts are hoping for a sizable upgrade of their smartphone's camera capabilities. History The histor...
32
3
François Simond (supercurio)'s profile photoFriedrich Sinofzik's profile photoSpectraStudy's profile photoTK Bay's profile photo
6 comments
 
+Marton Varga awesome, you're a very attentive reader :)
I reformulated and changed the first link as its author deleted the post (which makes sense now!) and 21 indeed!
Thank you very much.
Add a comment...
 
Tired of waiting, I just built my own AOSP 5.1.1+ for the Nexus 9.

The original 5.1.1_r1 segfaults during compilation for aarch64, so I built the current "master" branch instead.
Not sure it'll boot after flashing it, let's see !

Edit: fixed the segfault: the machine needed more RAM+Swap. 
c112341 Merge "common.py: Add support for squashfs" by Mohamad Ayyash - 15 hours ago master · a990534 common.py: Add support for squashfs by Mohamad Ayyash - 15 hours ago; 0a3b080 Merge "Revert "Revert "Pack relocations for mips64""" by Dmitriy Ivanov - 2 days ago; 8387d99 Revert "Revert "Pack ...
22
Michael Panzer's profile photoKarl Ramstedt's profile photoFrançois Simond (supercurio)'s profile photoLuis Ojeda's profile photo
18 comments
 
The past 2 years Google has brought shame to the nexus family... No longer do they provide fast updates to all their devices. They need to bring more people into their android division. 
Add a comment...
 
Some facts after a quick analysis of both +HTC  #ONEm9   and +LG Mobile Global  #LGG4  produced DNG.

- Both lack flat-field correction
- Both provide incomplete matrix-only color profiling: no DCP
- Neither use compression
- HTC One M9 DNG is 10 bit stored in 16bit uncompressed data: 39MB per 20 Mpixel image.
- LG G4 DNG is 10 bit stored uncompressed, 20MB per 16 Mpixel image
- Both have non-optimal noise profiling settings: HTC One M9 set noise reduction too high and LG G4 lacks noise profiling entierly.

Notes on flat-field correction:
Mobile camera modules require such correction to correct both vignetting and color cast (like pink spot / greenish or blueish corners).
HTC One M9 requires less correction than the LG G4.
It is only possible to compensate for light fall-off, in RAW image editors, not color cast.
As a result, the color cast in corners is essentially non-fixable.

Attached: the #LGG4  DNG sample provided by +Colby Brown rendered in Lightroom with only modification an increased contrast and exposure slightly, to make both vignetting and color cast more obvious.

The least I can say is that there's room for improvement, both DNG implementation being non-optimized and incomplete.
63
5
Matthew Johnson's profile photoJulien SALAMI's profile photoTerry Choo's profile photoSimeon Marinov's profile photo
25 comments
 
+Davi Porciuncula​​ unfortunately no, Android DNGCreator does not support dcp color profiles at all nor compression.
Also I see that the made they flat-field correction gainmap optional: it was required in previous versions of the SDK if I remember correctly.
Documentation: https://developer.android.com/reference/android/hardware/camera2/CaptureResult.html#STATISTICS_LENS_SHADING_CORRECTION_MAP
Add a comment...
 
+Dart​​ team at Google prepares an alternative to both the Java language and current SDK for Android app development.

Think of it as an in-between Java + Android SDK, Phone Gap and HTML5 apps, implemented using Dart runtime, Google Chrome technologies and an original, direct implementation of Material Design as UI widgets.

It won't replace Java: Android OS itself is mostly written in Java but that could become the way to write Android apps in scripts loaded from a remote server instead of compiled and updated through the Play Store like today. 
22
2
Zak Taccardi's profile photoMichael Panzer's profile photoLyndon Bredenkamp's profile photogeorge oloo's profile photo
10 comments
Add a comment...
 
Android runtime was expected, but IOS apps support, Wow!

Well, technically it requires to compile iOS apps using their toolchain, you won't be able to run any iOS app as-is on Windows - if I understood correctly.

A few third party re-implementation of iOS libs were in the works for many years now but never were completed.
Microsoft making it happen eventually is what impresses me the most.
 
Here's the big news from Microsoft​ #Build2015: Android​ and iOS apps are coming to Windows​ 10.
Company says that it's "building bridges" to developers on other platforms.
6 comments on original post
15
2
François Simond (supercurio)'s profile photoMichael Panzer's profile photoRohan Blake's profile photoJoseph B's profile photo
9 comments
 
+François Simond cough got steam and a GTX970 for that 😁
But I kind of see your point a bit...
Add a comment...
 
Congrats to my friends and neighbors at +Kolor for joining +GoPro!

I'm very proud having worked with them years ago when the company was just the two founders :)
I always admired their vision and strength to make things happen and wish them the best in this new configuration.
A message from Alexandre Jenny, founder of Kolor.
4
Aden J Purcell's profile photoFrançois Simond (supercurio)'s profile photo
Have him in circles
4,754 people
ALI RAZA's profile photo
Casey Harkins's profile photo
lely c's profile photo
Christian Robecchi's profile photo
Fedor von Bock's profile photo
Doanh Doanh's profile photo
Hưng Thịnh Phát's profile photo
gaurav gupta's profile photo
Lars Wold's profile photo
 
After installing and trying a bunch of Hacker News readers tonight, here's my new favorite 👍
An ad-free, open-source Hacker News reader with material design look and fe...
1
David Poole's profile photobarnassey トーマス's profile photoFrançois Simond (supercurio)'s profile photo
3 comments
 
+David Poole sorry I am unable to understand your question.
+barnassey トーマス​ opps, better report that crash then 😉 it's open-source on github and stuff! 
Add a comment...
 
Servo Web browser layout engine is a pretty cool project to follow.

Nobody knows how much better.. or worse it will be in the end, once completed compared to existing alternatives like Webkit, Blink or the current one powering Mozilla products: Gecko.

It is promising however for several reasons that make it unique and interesting:
- it's written in Rust, a new compiled system language without GC that manages memory allocation securely at compilation instead of runtime.
- the design is brand new and is modeled after the observation that popular machines today are equipped with many small cores.

The state of mind of the community appears rather lovely too! 
Servo is a new prototype web browser layout engine written in Rust that was launched by Mozilla in 2012 with a new architecture to achieve high parallelism on components like layout and painting. It has been progressing at an amazing pace, with over 120 CSS properties currently supported, ...
8
1
Karl Ramstedt's profile photoPhil Sweeney's profile photoFrançois Simond (supercurio)'s profile photogeorge oloo's profile photo
3 comments
 
+Phil Sweeney while both Servo and Rust generates tons of enthusiasm, the layout engine itself is still very much a work in progress.
This video is a good illustration of that ^^  https://www.youtube.com/watch?v=pZGhnqtXVdc
Add a comment...
 
I said it after the iPhone unveiling and camera coverage and I'm gonna say the same for +LG Electronics G4.
An article at the core of the marketing strategy, written by someone paid to promote a product is not a review

From +Colby Brown website's bio: http://www.colbybrownphotography.com/bio/

Colby creates & helps run various social influencer marketing campaigns for some of the biggest companies and destinations in the world

The same +Colby Brown appearing in +LG Mobile Global marketing videos on +YouTube : https://www.youtube.com/watch?v=o_bBgVjQUSU

Then he goes as far as adding at the end of the article:

FULL DISCLOSURE
While I partnered with LG to create both images and videos to help promote the global launch of the LG G4, they never asked me or payed me to write this review. These thoughts and opinions are mine and mine alone. Even if you think that I am biased because of our business arrangement, I have provided enough details and high resolution images for you to come to your own conclusions about this device.

Yeah, right..

Note that it doesn't remove anything from the quality of the pictures published, and having image samples is a great resource.
However, such marketing strategy, claiming to publish non-biased reviews, written by someone selling his services of social influencer marketing campain specialist is unethical and dishonest.
"Rays of Light" Upper Antelope Canyon, AZ One of my favorite places to photograph in the US Southwest has to be around Page, AZ. Here you not only found… - Colby Brown – Google+
21
3
daniel cruz's profile photoHector Malave's profile photoFatman Lee's profile photoRachid Otsmane-Elhaou's profile photo
18 comments
 
Those photos are sizzling though. 
Add a comment...
 
Currently teaching myself OpenGL basics, I'm enjoying the lessons from +Jeffrey Chastine​ available on +YouTube​ 
https://www.youtube.com/user/jeffchastine/videos

You probably guessed: there's interesting things to do about color and display calibration with GPUs using OpenGL ES shaders ;)

Is any resource you particularly liked when learning about this 3D API?
2
François Simond (supercurio)'s profile photoClemens Schartmüller's profile photo
3 comments
 
Well, University project for a 4 sided cave ;) but thanks and gl getting into opengl
Add a comment...
 
Latest Samsung 840 EVO SSD firmware update
Version: EXT0DB6Q

Breaks TRIM function.. seriously. don't they somehow test their disks firmware before releasing them?
Now let's see if there's a way to downgrade...

# fstrim -v /
fstrim: /: FITRIM ioctl failed: Input/output error

kernel log:
[ 1165.961733] ata4.00: exception Emask 0x0 SAct 0x1400ff0 SErr 0x0 action 0x6 frozen
[ 1165.961740] ata4.00: failed command: WRITE FPDMA QUEUED
[ 1165.961748] ata4.00: cmd 61/40:20:e8:46:83/00:00:1c:00:00/40 tag 4 ncq 32768 out
[ 1165.961748]          res 40/00:01:00:00:00/00:00:00:00:00/40 Emask 0x4 (timeout)
[ 1165.961751] ata4.00: status: { DRDY }
[ 1165.961754] ata4.00: failed command: WRITE FPDMA QUEUED
[ 1165.961760] ata4.00: cmd 61/10:28:98:62:83/00:00:1c:00:00/40 tag 5 ncq 8192 out
[ 1165.961760]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1165.961763] ata4.00: status: { DRDY }
[ 1165.961766] ata4.00: failed command: WRITE FPDMA QUEUED
[ 1165.961772] ata4.00: cmd 61/08:30:00:88:de/00:00:1b:00:00/40 tag 6 ncq 4096 out
[ 1165.961772]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1165.961775] ata4.00: status: { DRDY }
[ 1165.961778] ata4.00: failed command: WRITE FPDMA QUEUED
[ 1165.961784] ata4.00: cmd 61/30:38:b0:30:27/00:00:1c:00:00/40 tag 7 ncq 24576 out
[ 1165.961784]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1165.961787] ata4.00: status: { DRDY }
[ 1165.961789] ata4.00: failed command: WRITE FPDMA QUEUED
[ 1165.961795] ata4.00: cmd 61/08:40:00:68:de/00:00:1b:00:00/40 tag 8 ncq 4096 out
[ 1165.961795]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1165.961798] ata4.00: status: { DRDY }
[ 1165.961801] ata4.00: failed command: SEND FPDMA QUEUED
[ 1165.961807] ata4.00: cmd 64/01:48:00:00:00/00:00:00:00:00/a0 tag 9 ncq 512 out
[ 1165.961807]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1165.961810] ata4.00: status: { DRDY }
[ 1165.961813] ata4.00: failed command: READ FPDMA QUEUED
[ 1165.961819] ata4.00: cmd 60/08:50:e0:30:27/00:00:1c:00:00/40 tag 10 ncq 4096 in
[ 1165.961819]          res 40/00:01:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1165.961822] ata4.00: status: { DRDY }
[ 1165.961824] ata4.00: failed command: READ FPDMA QUEUED
[ 1165.961830] ata4.00: cmd 60/10:58:48:6a:d1/00:00:16:00:00/40 tag 11 ncq 8192 in
[ 1165.961830]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1165.961833] ata4.00: status: { DRDY }
[ 1165.961836] ata4.00: failed command: WRITE FPDMA QUEUED
[ 1165.961842] ata4.00: cmd 61/18:b0:a0:17:4a/00:00:16:00:00/40 tag 22 ncq 12288 out
[ 1165.961842]          res 40/00:ff:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1165.961845] ata4.00: status: { DRDY }
[ 1165.961848] ata4.00: failed command: WRITE FPDMA QUEUED
[ 1165.961854] ata4.00: cmd 61/60:c0:d0:ae:44/00:00:1d:00:00/40 tag 24 ncq 49152 out
[ 1165.961854]          res 40/00:ff:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1165.961855] ata4.00: status: { DRDY }
[ 1165.961858] ata4: hard resetting link
[ 1166.453625] ata4: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 1166.453820] ata4.00: supports DRM functions and may not be fully accessible
[ 1166.454097] ata4.00: supports DRM functions and may not be fully accessible
[ 1166.454150] ata4.00: configured for UDMA/133
[ 1166.469581] ata4.00: device reported invalid CHS sector 0
[ 1166.469585] ata4.00: device reported invalid CHS sector 0
[ 1166.469588] ata4.00: device reported invalid CHS sector 0
[ 1166.469590] ata4.00: device reported invalid CHS sector 0
[ 1166.469593] ata4.00: device reported invalid CHS sector 0
[ 1166.469595] ata4.00: device reported invalid CHS sector 0
[ 1166.469597] ata4.00: device reported invalid CHS sector 0
[ 1166.469600] ata4.00: device reported invalid CHS sector 0
[ 1166.469602] ata4.00: device reported invalid CHS sector 0
[ 1166.469605] ata4.00: device reported invalid CHS sector 0
[ 1166.469619] sd 3:0:0:0: [sdb]  
[ 1166.469621] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
[ 1166.469624] sd 3:0:0:0: [sdb]  
[ 1166.469626] Sense Key : Aborted Command [current] [descriptor]
[ 1166.469630] Descriptor sense data with sense descriptors (in hex):
[ 1166.469632]         72 0b 00 00 00 00 00 0c 00 0a 80 00 00 00 00 00 
[ 1166.469641]         00 00 00 00 
[ 1166.469646] sd 3:0:0:0: [sdb]  
[ 1166.469648] Add. Sense: No additional sense information
[ 1166.469651] sd 3:0:0:0: [sdb] CDB: 
[ 1166.469653] Write same(16): 93 08 00 00 00 00 17 07 c3 d8 00 00 00 08 00 00
[ 1166.469664] end_request: I/O error, dev sdb, sector 386384856
[ 1166.469680] ata4: EH complete

After a  that the disk is in a wonky state and the system needs to be rebooted, or the filesystem crashes at some point later.
Frigging waste of time.
Samsung SSD Downloads - Find the latest firmware, data migration tools, and the Magician program.
5
4
Renaud “和彦” Lepage's profile photoDaniele Pantaleo (ozone89)'s profile photoCédric Andrieu's profile photogeorge oloo's profile photo
9 comments
 
Takes me back to the days of rfs file system 
Add a comment...
 
Excellent question from +Aden J Purcell:
 
Apparently the display on the G4 follows the DCI color space instead of sRGB. Do you know if this is a good thing? I thought movies were mastered in sRGB, not DCI? And even if LG has accurately calibrated the G4 to DCI, could they have messed up the gamma? Or is a 2.2 gamma part of DCI meaning LG can't mess with it otherwise they can't market their device as DCI compliant?

Like +Samsung Mobile has done before, marketing their Super AMOLED devices as color accurate according to the Adobe RGB standard to appeal to photographers, +LG Electronics described the LG G4 display as following another standard they call "DCI"

First of all, DCI acronym stands for "Digital Cinema Initiatives, LLC" and not the name of a color standard.
Assuming they are talking about DCI-P3 color gamut, it's the gamut part of one of the color encoding formats for professional cinema projection.

DCI-P3
- white point coordinates x: 0. 314, y: 0.351 vs sRGB x:0.3127 y: 0.3290
- gamma 2.6 vs sRGB around 2.2 average 
- pretty wide gamut, color primaries color hues not being the same as sRGB: green has less yellow, red has less orange, 

A comparison of sRGB and DCI-P3 color gamuts by +Jeff Yurek's blog:
https://dotcolordotcom.files.wordpress.com/2012/12/pantone_2013_dot_color.png
Additionally, the white point is not even the same.

Today, Android applications lack color management abilities which would allow to convert one color encoding to another and display content as intended on various displays seamlessly.

+LG Electronics throws in the "DCI" name to impress reporters and customers with the intent to convince them of the benefits in color accuracy and  true-to-life color reproduction of their new display.

However, when checking what they're mentioning stands for, it is clear that displaying today's content, which color are encoded using sRGB (Rec.709 gamut and around 2.2 gamma) on a "DCI-P3" display (wide gamut and gamma 2.6) would lead to particularly inaccurate color rendering.
Colors would not be rendered with the right hue, look too intense (over-saturated look), and annoyingly too dark due to the gamma 2.6, which increases saturation even more.

In conclusion, +LG Electronics is proud of their new wide-gamut display but please don't be fooled by the marketing mumbo-jumbo employed.
It doesn't correspond to any professional standard or any standard altogether, it will distort colors instead of rendering them faithfully, and this is mostly a response to Samsung, just as bogus Adobe RGB accuracy claims.

Source: DCI-P3:
http://www.hp.com/united-states/campaigns/workstations/pdfs/lp2480zx-dci--p3-emulation.pdf
38
12
Luca Zaninello's profile photoStuart Blakeley (SB13X)'s profile photoHRISHIKESH BISEN's profile photoDomenico Majorana's profile photo
11 comments
 
+Karl Ramstedt​ Yep.
Also note that the supposedly Adobe RGB mode has very little to do with Adobe RGB colorspace: wrong gamma, wrong RGB primaries as well. 
Add a comment...
People
Have him in circles
4,754 people
ALI RAZA's profile photo
Casey Harkins's profile photo
lely c's profile photo
Christian Robecchi's profile photo
Fedor von Bock's profile photo
Doanh Doanh's profile photo
Hưng Thịnh Phát's profile photo
gaurav gupta's profile photo
Lars Wold's profile photo
Work
Occupation
SpectraStudy.com founder
Skills
Analyst, developer, mobile / display / camera engineer
Employment
  • SpectraStudy
    Analyst, writer, software engineer, 2015 - present
  • Project Voodoo
    Android applications and drivers developper, 2010 - present
  • Cyanogen Inc.
    Software Engineer, 2014 - 2014
    Color, sensor, optics profiling of Cyanogen Camera RAW DNG capture on the OnePlus One smartphone
Basic Information
Gender
Human
Birthday
January 28, 1982
Other names
supercurio
Story
Tagline
Mobile display and camera calibration
Bragging rights
Autodidact
Places
Map of the places this user has livedMap of the places this user has livedMap of the places this user has lived
Currently
Chambéry, Savoy, France