What's new

Any quality loss dng to tiff?

juicegoose

TPF Noob!
Joined
Sep 4, 2013
Messages
102
Reaction score
4
Is there any quality loss when converting a dng from lightroom to an 8 bit tiff in elements?
 
Well, you are at least going to lose bit depth, so, generally speaking yes.
 
Raw files start out as 12-bit or 14-bit depth Raw image data files.

14-bits can code 16,384 discrete colors in each of the 3 color channels that make a color digital photograph.
12-bits can code 4,096 discrete colors in each of the 3 color channels that make a color digital photograph.
8-bits can code 256 colors in each of the 3 color channels that make a color digital photograph.

Image Files
 
Is there any quality loss when converting a dng from lightroom to an 8 bit tiff in elements?

Yep, and that in a nutshell is what's wrong with Elements.

Joe
 
Lightroom is a professional grade database manager that has a module that is a parametric raw converter application (Adobe Camera Raw, or ACR).

Elements is a consumer grade raster graphics editing application that has about 1/2 of the same parametric raw converter application tools and functions Lightroom has.
 
Well shoot. I usually use elements to do some layer work and more advanced cloning. Guess ill have to look at cs6
 
CS 6 (Creative Suite 6) is still being sold as a Legacy Version ($654) Adobe Photoshop CS6 [LEGACY VERSION], but Adobe went to a subscription and Photoshop Creative Cloud (CC) for a variety of reasons.

Until December 31, 2013 Adobe is offering a Photoshop Photographers Package, 1 year subscription @ $9.99 a month that includes CC and LR 5. http://www.adobe.com/products/photoshop.html?promoid=JOLIW
CC has many features and functions CS 6 doesn't have. Many of those additional features and functions are from the no longer available $999 CS 6 Extended.
 
Wow thats seems like a great deal. I don't own a version of Photoshop though so I don't qualify. Oh well.
 
Wow thats seems like a great deal. I don't own a version of Photoshop though so I don't qualify. Oh well.

You don't have to own Photoshop. Its part of the package. You had better jump on it
 
Yep. The promotion is all about that you DON"T NEED TO ALREADY HAVE PHOTOSHOP - if you'd read the info Adobe provides.

You can lead a horse to water, but you can't make it drink.

I'll click the links for you - https://creative.adobe.com/plans/offer/photoshop+lightroom

Adobe Creative Cloud - FAQ

Photoshop photography program offer

Who is eligible for this offer?
This offer is available to anyone until December 31, 2013. There are no previous product ownership requirements.
 
Last edited:
Raw files start out as 12-bit or 14-bit depth Raw image data files.


14-bits can code 16,384 discrete colors in each of the 3 color channels that make a color digital photograph.
12-bits can code 4,096 discrete colors in each of the 3 color channels that make a color digital photograph.
8-bits can code 256 colors in each of the 3 color channels that make a color digital photograph.

RAW files are raw sensor data, and are not really an image file. Whether they are 12 bit or 14 bit does not affect the bit depth of an RGB (3 color) image file produced from the raw data. And the colors are not encoded in the 12/14 bit data, as they are in an RGB image file, so the number of colors is not really as indicated by the bit depth.

The "dynamic range", or exposure values, are recorded in the RAW file data. Hence a 14 bit RAW fine has 16,384 tonal brightness values, not colors. The color of each pixel produced when the data is interpoled is much greater! Consider that the minimum number of sensor locations used to determine the color of a single pixel will be a 3x3 matrix with 9 each of these 14 bit data bytes! Rather than 16,384 in each of 3 color channels it's 16,384 values in each of 9 different bytes, which is 2 ^ (14 * 9) and equates some huge number that is 85 with 36 zeros after it!

Of course when that is saved as a 16 bit TIFF file (which is a 3 color RGB file) is has to be narrowed down to 2^16 bits per channel, or some 65536 values for each of Red, Green, and Blue. That is commonly referred to a 48-bit color, with a total of 280 million different colors. There will be just as many colors from a 12 bit raw file that is converted to a 16 bit TIFF! The RAW converter will use 2 ^ (12 * 9) bits of data, which equates to 32 with 31 zeros, to generate the same 48-bit color. So the number of colors does not change, only the range of brightness values.

A 16 bit TIFF has 16 bits per each of three color channels, so it is a 48 bit color system, with 280 million different shades of color.

An 8 bit TIFF or JPEG has 8 bits per channel so it is a 24 bit color system with 16.7 millions different shades of color.
 
Yes, a "Raw image data file" is not yet a photography, but the raw sensor data never leaves the camera.

The image sensor in the camera is an analog device that cannot record color. Plus, the image sensor is a linear device while human eyes are non-linear.
So, the analog voltages the sensor pixels produce have to be amplified according to the ISO setting of the camera and would displayed as a grayscale image at that point.

The next step in the process, and the point the raw sensor data is lost, is for the A/D (analog to digital) converter in the camera to convert the analog pixel voltages to digital numbers and that's where the digital data file bit depth is set prior to the Raw image data file being written to the memory card.

The Raw image data file won't become a photography until the Raw image data file has been run through a Raw converter.

The Raw converter uses a variety of algorithms to interpolate the photo colors (demosaicing), set the gamma curve, tone map, and perform other necessary functions to make the Raw image data file an image that approximates what human eyes see.
 
Yes, a "Raw image data file" is not yet a photography, but the raw sensor data never leaves the camera.

The image sensor in the camera is an analog device that cannot record color. Plus, the image sensor is a linear device while human eyes are non-linear.
So, the analog voltages the sensor pixels produce have to be amplified according to the ISO setting of the camera and would displayed as a grayscale image at that point.

The raw sensor data does leave the camera. It is indeed initially analog and is converted to digital, stored on a memory card in a "RAW file" and that most definitily leaves the camera. Note that it isn't "per pixel" data, it's per sensor site data.

The image sensor does record color. It is not encoded into the bit pattern on a per byte basis though, and instead is encoded with the Color Filter Array pattern to the bytes.

The linear vs logrithmic nature of the sensor isn't really important to this discussion.

Even after the analog voltage is digitized, as well as before, it really is not image data, grayscale or otherwise. Out of four sensor locations, two are for green light only, one is for blue and one is for red. Even to generate a grayscale image they have to be interpolated to get the appropriate intensities relative to each other. That is done by interpolation, using a maxtrix of sensor locations for each pixel. Hence no pixel value is ever formed from just the sensor data that physically corresponds to the location of the pixel generated. That pixel gets its value from 9 different sensor sites at a minimum.

The next step in the process, and the point the raw sensor data is lost, is for the A/D (analog to digital) converter in the camera to convert the analog pixel voltages to digital numbers and that's where the digital data file bit depth is set prior to the Raw image data file being written to the memory card.

That is were the bit depth of the raw data file, which is digital, is set. It is not pixel data yet. And the bit depth is only brightness value, not color, as was noted in my previous article. This file is typically 12 or 14 bits per byte for a sensor location; while an image file is 24 or 48 bits per pixel (commonly called 8 bit or 16 bit, per color instead of per pixel).

The Raw image data file won't become a photography until the Raw image data file has been run through a Raw converter.

Correct! That may be done in the camera or external to the camera. The RAW data file, which is not pixel data, is transformed to pixel data. The color is encoded into the bit data, with three color channels. Regardless of the bit depth of the RAW data, the resulting image data is either 8 bit or 16 bit, typically. It is now per pixel data. It takes three bytes, of whatever size, to define a single pixel and those bytes define only a single pixel. That is very different from sensor data, where there are no pixels defined and any given byte is part of the data set used for many pixels.

The RAW file, even when is has been stored external to the camera, is still not an image file. And it's bit depth defines the number of fstops of potential brightness an image can have, but not the number of colors in the image.

The Raw converter uses a variety of algorithms to interpolate the photo colors (demosaicing), set the gamma curve, tone map, and perform other necessary functions to make the Raw image data file an image that approximates what human eyes see.

The primary point though is that RAW sensor data is not per pixel data, it is per sensor site location. Data from each sensor site use used for more than a single pixel. Each pixel is defined by a matrrix of sensor sites. Demosacing is what produces a file of pixel data. Note that "gamma curve, tone map" the "other" functions aren't really necessary to make it an image file, just demosaicing. The rest of that makes just it an image with particular characteristics (and that may or may not approximate the human eye).
 
KmH
Although your consistently smarta$$ attitude isn't necessary I appreciate your help. I had read reviews from scott kelby and dpreview that stated you needed cs3 or later. I chatted with adobe prior to your comments and had already verified they changed this policy due to negative response from customers. I did in fact download the suite thanks.
 

Most reactions

Back
Top Bottom