Lightroom and ptgui - Should I use DNG, TIFFs or JPG?

[Jesper]

Hi all

I am figuring out the best workflow to get the most out of my photos from my drone.

I capture RAW

Import to Lightroom, and tweak as I like.

What is the best way from Lightroom to import into ptgui?

• Export the Importing as JPGs. These photos look the most saturated and looks like how I wanted them in Lightroom
• Export as TIFFs (16 bit). Looking a bit less saturated, but a lot bigger.
• Or export as DNG. Actually I tried this because I thought that it was the best, but I cannot seee the changes i made in Lightroom. They still look flat like RAW DNGs. Why dont the changes reflect in the exported DNGs`?

Time is no problem I want the best if quality can be better.

What would you do?

Best regards 

Jesper

[John Houghton]

Jesper, Output tiff files for PTGui (preferably 16-bit).  This will give you the highest quality.  Adjustments made in Lightroom will not be actioned when the raw files are opened directly with PTGui, which uses DCRaw to do the conversion.

John

[Jesper]

Hi John

Thankyou for your quick answer.

Would you describe the TIFF-solution as significantly better as the JPGsolution?

As in 5-10% better (just guessing) :o)

Is it in the colors or the resolution?

I thought that the quality - when ingested in PTGUI was what I got out and PTGUI would not alter that, but merely stitch.

Jesper

[Erik Krause]

Am 31.07.2016 um 22:46 schrieb Jesper:

> Is it in the colors or the resolution?

Both. JPG is highly compressed, which means that information is lost.
Furthermore you have only 8 bit per channel, that means only 256
different values.

> I thought that the quality - when ingested in PTGUI was what I got out and
> PTGUI would not alter that, but merely stitch.

"Merely stitch" means complex remapping with heavy interpolation. After
that it needs to be recompressed if output as JPG. Any post processing
reduces quality even further. To avoid that it is best to use 16 bit
TIFF for anything and reduce to 8 bit JPG only for the very final output.

--
Erik Krause
http://www.erik-krause.de

---
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus

[John Houghton]

On Sunday, July 31, 2016 at 9:46:14 PM UTC+1, Jesper wrote:

Would you describe the TIFF-solution as significantly better as the JPGsolution?

As in 5-10% better (just guessing) :o)

A lot depends on the nature of the subject in the panorama and what post-processing (if any) is performed on it.  Often, there will be no obvious difference between tiff and jpeg workflows.  Bland areas like blue skies are the most likely to suffer from visible artifacts such as banding and blocking. This 100% crop from a panorama shows typical differences that can be expected, even when using very little jpeg compression: http://www.johnhpanos.com/tiff-jpeg.jpg .  This is the output from PTGui with a couple of minor adjustments in Photoshop (Shadows/Highlights and Levels).  The artifacts will be much more severe when higher levels of compression are used.  It's therefore safest keep to 16bit tiff working for as long as possible before finally converting to a final jpeg image.

John

 

[Jeff Wischkaemper]

In addition to what John said, the difference will be most pronounced when doing HDR panoramas. Remember that even though a 16-bit file contains twice as much data as an 8-bit file (i.e. there are twice as many bits), the 16-bit file can represent up to 256 times as much information (I say "up to" here because you can always use more data to represent the same amount of information - obviously you can represent 8-bits of information with 16-bits worth of data... in this case the extra bits of data don't actually translate into additional information). 

To use a hypothetical example, if your raw files contain 12 effective bits of resolution (after accounting for noise), then converting to an 8-bit workflow will result in a loss of approximately 95% of the information in contained in the original file(s) (2^12 / 2^8 = 2^4 = 16, 1/16 = 0.0625). As John rightly says, this loss of information usually manifests itself most clearly in areas where you would typically like to see smooth gradients, like a blue sky. If you want to think about it this way, if an 8-bit file has 255 values of color for a certain channel (say, blue), the 16-bit file can represent 254 additional colors between each of those steps (so, for example, the 8-bit file could represent 46 and 47, but nothing in between, where the 16-bit file could represent something like 46.02083 and 46.02127). 

It is also important to say that this analysis applies to all 8-bit workflows, including 8-bit TIFFs. Compared to an 8-bit TIFF workflow, using JPEGs as a source file would also introduce compression errors, which can be more or less problematic, depending on the composition of the scene. 

One other matter to consider when using JPEGs as source files in PTGui concerns control points: namely, JPEG compression can have an adverse effect on control point detection algorithms, resulting in fewer (or poorer) matches. Again, this may or may not matter depending on the composition of the scene, but I occasionally find that PTGui easily matches all images on TIFFs when the same scene fails using JPEGs. As always, your mileage may vary. 

[Jesper]

Hi Guys

Thankyou for your efford.

I will stick to TIFFS in production and end with JPGs

Jesper