Sunday, April 20, 2014

Lightroom 'plug-in' like functionality for CeroNoice

As I discussed in my recent post, I believe the CeroNoice workflow creates a serious alternative to other 32-bit TIFF routes, such as ‘Merge to 32-bit HDR’ or Photoshop. I believe the results are first class, for instance the image of Winchester Cathedral at the bottom of this post. 

Also in the last post I provided a small .bat script to partially automate the creation of the 32-bit images from within Lightroom. The only limitation is that this was for a Windows environment.

In this post I introduce ‘CeroNoice Bracketing.exe’, which can be used as an external editor in Lightroom.

To set up the new workflow, simply put the required files in a processing folder, as detailed before, ie CeroNoice.exe etc, together with my .exe, called CeroNoice Bracketing.exe (email me for a copy).

Set up an external editor preset in Lightroom, with the following attributes:

  • Export original (can be .cr2 or .dng); 
  • No file naming; 
  • Post processing = open in other application (ie point to CeroNoice Bracketing)

Finally, set up a ‘watched folder’ in LR (under File, Auto Import), where the 32-bit processed .tif will be returned. The returned file will have the image name of the lightest original, with _32-bit appended.

Once set up, all you do is select, from darkest to lightest, the input images, export via the CeroNoice Bracketing preset and carry on working! 

In a few seconds the output file will appear in the LR watch folder (which I call 32-bit Processing, for easy identification).

All working files, ie LR exports are deleted. You can move the 32-bit TIFF or simply leave in the 32-bit Processing folder.

I have found LR handles the CeroNoice created TIFFs well, eg able to fully use the +/- 10 Ev range in the floating point TIFF.

Bottom line: with the help of my little .exe, Windows’ users can access CeroNoice from within LR in a fully automatic manner, from either .cr2s or .dngs.

Thursday, April 17, 2014

Noiseless Post Processing: fact or fiction?

I have posted before about the ground breaking work being done by Alex and others over at Magic Lantern. For some, however, my posts must fall on deaf ears, ie Canon-specific. 

This post is different, it is about an ML-developed tool for all camera users, but especially those with Ligthroom.

It is an adaptation of a piece of software called Zero Noise: . The ML adaptation is called CeroNoice. Basically it creates a 'new' RAW negative from a bracket set, with no noise!

I think it is an exciting tool, but, as many things over at ML, to use it requires some handraulic work: that is until now.

Until this week, I had never written a Windows .bat script, but I sensed this is what I needed for my workflow with CeroNoice and Lightroom. Thus what follows is how I have done it and got CeroNoice running for me. To be clear, ‘all’ I want to do is use CeroNoice to create a noiseless 32-bit (tif) negative, starting with a set of bracketed .cr2 files. That is, I’m not interested in .dngs and not using LR. I want an LR workflow.

So this is my LR-based workflow: I have appended the .bat file I wrote to the end of this post (which I’m sure will have the experts cringing, ie I’m sure it could be more efficient: but it works for me).

First, set up a Folder where you will do your ‘out of LR’ processing, eg a folder called CeroNoice, or anything you like.

In that folder place the following executables, which you will find references to on the ML site CeroNoice.exe, dcraw.exe, dng_validate.exe and exiftool.exe. Note: I’m not even sure you need all these, but there is no harm placing them in the folder. Also note: read the stuff on the ML site, you must have the right/latest versions, ie each piece of software needs to be compatible with the others.

Open up a .txt file, eg in NotePad, and paste in the .bat text below and save this file as a .bat file, ie with a .bat extension. You can call it what you like, but make sure it is in your processing folder.

The Folder is now set up.

Now go to Lightroom (you don’t have to use LR to get your .cr2 files into the processing folder, but this is how I do it.) Set up an export preset, with the following attributes:

  • Identify the processing folder as the export location 
  • Don’t identify any sub folder 
  • Select file naming as Custom Name – Sequence 
  • Enter “in”, without quotes or spaces, as the custom text (key point) 
  • Select Original as the File setting format, ie the .bat file works with .cr2 files. You will need to change the .bat if you use .dngs. 
  • Save this as an export preset, eg called, say, CeroNoice

You are now ready to process .cr2 brackets.

Go to the brackets you are interested in processing, say, x of them. Select the darkest one (key step!) and holding the Ctl button down left click on the other brackets you wish to select, in a darker to lighter manner.

You will now have x images selected. Now right click on the brackets you have selected in order, and choose export…

Select the CeroNoice preset and press export. If you are doing multiple brackets, you only need select export with previous next time (unless you have used another preset in-between), BUT, only process one bracket set at a time!

Now go to the folder you set up, where you should see x .cr2 images and x .xmp files.

Now double click the .bat file.

The .bat file will do its stuff and create a 32-bit Tiff called out.tif. All input files will be deleted.

You can now play with the 32-bit Tiff, say, in PS-CC or ACR. But, as I said, this post is for the LR users.

Go back to LR and the library module, select import… , select the folder where the .tif is, there should only be one image there, ensure move is selected, select your ‘To’ folder, ie where you wish to move the 32-bit file to, and press Import.

The file will move from the processing folder, so this is ready for another bracket set, and now it will be in LR, ready for 32-bit processing!

The file will be large! Very large!

The file will, at first, be green and dark!

It will have a +/- 10 Ev range!

However, all the LR processing tools work at the 32-bit level, including the WB. So it is easy in ACR or LR to get the look you want.

For example, here is a processed image from three brackets:

Finally, here is the .bat text. As I said it is far from perfect, but is functional (although I do get warnings, it appears to not matter!) 

@echo off

title CeroNoice Processor

set brackets=0

for %%n in (*.cr2) do set /a brackets=brackets+1

echo %brackets%

set /a brackets=%brackets%+1

set counter=1

set cero=ceronoice.exe


set  cero=%cero% in-%counter%.cr2

set /a counter=%counter%+1

if %counter% EQU %brackets% (

goto continue

) else (

goto begin




dng_validate.exe -3 out out.dng

rem ensure all the input and working files are deleted

del *.xmp

del *.cr2

del *.dng



Bottom line: please don't take my word for the noiseless claim. I believe if you decide to put a little effort in setting the above up, you will also be amazed at the quality of your (new) digital negative!

Tuesday, April 8, 2014

Magic Exposures

As those that read my postings know, I tend to focus on the technical side of my (Canon-centric) photographic developments in this blog. Also, I try to give insights into some of my favorite technologies, eg: Magic Lantern, CamRanger and Promote Remote.
One technique I have advocated in previous posts is the so-called Expose To The Right (ETRR) technique, which, because of the way digital tonal data is captured and stored, means that you should be seeking to push your exposure as far the right as possible, without ‘going over the edge’.

In this case, going over the edge means that some pixel sites become ‘full up’, ie no more photons can be captured and thus the electron count at that pixel site has maxed out. So how do you stop the sensor maxing out at places, but ensure an ETTR strategy?

There is a lot of information out there on the relative benefits of a full frame (large pixel) camera over a cropped DSLR or a small Point & Shoot camera. For those interested in learning more about the ‘science’ behind the camera, I commend this page:

Yes, all the science is interesting; however, most of us have the camera we already own. So pixel size and sensor area are kind of irrelevant: we have what we have! Thus we can only maximize the quality of our captured data through using the ‘best’ in-camera workflow we can, eg using ETTR and, for very high dynamic range scenes, exposure bracketing.

In this post I offer the following two Magic Lantern enhanced strategies:

Strategy 1 covers a scene with ‘normal’ dynamic range, that is either fully contained within the sensor’s capture capability or has minimal clipping, ie at the shadow end. In this case I recommend adopting an ML Auto-ETTR-based workflow, ie: 

  • Enable Auto-ETTR in ML, ie enable the ETTR module and switch ETTR on;
  • I prefer to use the SET button to generate an ETTR exposure, but some prefer a double-press of the shutter button;
  • Set the ETTR parameters you need, ie clipping %, mid-range S/N override and or shadow S/N override; 
  • Compose and focus the scene, put the exposure on Manual and at 0Ev as a ‘first guess’;
  • If you think the scene has a broader dynamic range than can be captured in one image, consider enabling Dual-ISO, eg a 100/800 shot (but note this requires more post processing, although this is pretty much automated now in Lightroom);
  • Press SET (to generate the exposure time – note no image is taken); 
  • Press the shutter to capture the fully optimized ETTR image.

Strategy 2 covers a scene with a large dynamic range, ie even beyond that captured when using Dual-ISO ETTR processing, ie unacceptable clipping at both ‘ends’ of the histogram. In this case I recommend adopting an Auto-bracketing-based workflow, ie: 

  • Ensure Auto-ETTR and Dual-ISO are turned off in ML;
  • Enable Auto-bracketing in ML and set the shot-2-shot Ev delta, eg 1 or 2 Ev is usually about right - remember the lower the Ev delta the more shots you will capture, and the higher the Ev delta the greater the possibility for generating post-processing artifacts;
  • I prefer to initiate the auto-bracketing sequence from the shadow end. Thus what I would do is switch to LV and use the ML spot meter (set for Ev measurements) to ‘scan’ the scene in the areas where I wish to see shadow details and ensure this area is not clipped, eg place it in zone 3 or 4 (if you are exploiting the Zone System). If you use LV Exposure Simulation you can also get a visual clue and, of course, in ML LV you have Raw histograms, ie Canon’s are 8-bit JPEG-based ones. The LV scene will likely look blown out, but don’t panic!
  • [An alternative auto-bracketing approach would be to simply put the exposure at 0Ev and let ML auto-bracketing ‘do its stuff’]; 
  • Press the shutter to initiate the bracketing and the ML auto-bracketing algorithmics will do their magic; and because you started at the shadow end, ie no shadow clipping but with highlight clipping, you will end up with a bracket sequence that progressively brings the exposure down so that there is no highlight clipping; 
  • Post process in your favorite way, eg: PhotoMatix, HDR Efex-Pro or Photoshop etc.

Bottom line: for the Canon user Magic Lantern, with the right exposure workflow, guarantees you will never need worry about exposure. Thus you are free to think about the important things in an image, for example your artist vision.