As mentioned above, JPEG is a lossy RGB format and has 8bit/channel, which limits its potential for further development and editing.
TIFF, in its 8bit variant, and with/or without compression is about as flexible as a JPEG, but image quality should be better because TIFF is lossless, and can be, even when compressed.
The example shows a CR3 file and several other formats as exported by PhotoLab 8
The 100% JPEG turns out to be larger than the RAW (with lossless compression)
The 8 bit TIFFs are quite big already and the 16 bit TIFF is huge
The DNGs can be large (linear, by PhotoLab) or small (by Adobe DNG converter)
While TIFF is a long standing, proven file format, Iâd probably tend to use DNG, if I wanted to avoid manufacturer specific RAW formats. Please note that DNG file sizes depend on which version of Adobeâs converter was used and how it was set.
so to summarize, and add information from further reading:
focusing on file information and processing
JPEG is not lossless even at the 100 quality setting.
TIFF is lossless but is still processed. see next point
Processing a RAW file has linear and nonlinear processes. And the original RAW has greater dynamic range and color range than processed RGB files.
DNG is a RAW file. DNG Linear may have linear, reversible corrections applied, but no irreversible changes to the RAW-type data. I assume it also have other information on how to further correct, like a DOP. No side car is needed with a DNG.
RAW file plus sidecar leaves the RAW data unchanged, and corrections are specified to the sidecar, specific to the processing software you are using.
I speculate that a DXO saved DNG file can be reopened by DXO and further processed. But it may or may not be exactly the same mechanics under the hood as reopening the RAW file with DXO and the sidecar, and further processing.
from DXO sites:
What are Linear DNG files and how should you use themâŚ
and
âDxO PhotoLab also supports DNGs which were created with DxO PhotoLab and DxO PureRAW. However, automatic optical corrections and denoising will not be available for those DNGs as they have already been applied. So youâll see the icon for âno optical module availableâ.â
Corrections or elaborations from others are welcome
Some years ago i tested and investigated this export posibilityâs of dxopl.
And we estabilisted that jpeg on 90% (i think i need to check) the max quality is wile keeping the file reasonable.
The last 10% is a filesize builder but has no advantage, growd, in quality.
Tiff 16bit Adobergb, thatâs my setting to NIK export.
I donât bother to store RW2"s in dng or tiff.
The rawconverters like that of adobe will probably cover 10 years or more of file types to convert. Even when photo editers are not able anymore. So why double your fileâs and data storage for the sake off maybe.
You could use that datastorage space for incremental backupâs.
I delete often the tiffâs after export back to dxopl unles i dit a lot of work on them. Mostly itâs the carrier vehikel from dxopl to NiK or my stacking app, and back to dxopl.
Yes, but itâs not an image. Itâs a file that needs to be converted to appear on a screen.
The DNG file is presented as a raw file, even though it is demosaiced and has already been corrected.
The idea is to use a first program to âenhanceâ the quality of the raw file transmitted to a second program.
One of the advantages of using a DNG to edit is you speeding up the editting lags due recalculating of your computer when you change a slider. So less powerfull pcâs bennefit from this.
It has some disadvantages doh.
One of it is colorshifting when you alter CA settings or too much color WB changes.
Why? Well in fact a DNG is a container which holts a TiFF like file.
Itâs holds a full colorspace equal to the workingspace you set to convert it to DNG.
Itâs RGB based and demosaiced and if you activated denoising denoised. And the optical module is used.
I believe there are two types of DNG in export.
RawDNG and customDNG.
If i remember correctly the rawDNG is a smaller filesize.
So recap.
If you like to use tiff to speedup progres of editting?
You can use dng wich holds a bigger colorspace.
And if you are bought the rawplus of dxo.(that converter application) you can use that to transport a as wide as possible file for further editing. (same as dxoplâs rawdng.)
Tiff 16bit when you also did masking and other editting or if you like use the custom DNG. (like creating a negative.)
This info is build on my memory because i havenât used dxopl alot this year. Due too old computer and lack of patience to wait after every change. And too much home improvement work to have time to sit.
Old computer will be replaced soon, time will be a questionâŚ
(Now i am back to my floor tyling job. 60cmx60cm tiles are a pain in the bu⌠To get right. Yes i know you have tile carriers which are like mini lifts but i donât buy 25 of them for a few tiles.)
I would love DxO to start embracing JPEG XL. While still compressed, it is leagues above the legacy JPEG and specifically designed for better compression and image quality (ideal for web and photographers).
I agree. I am completely unable to see any difference in quality with JPEGs above 90-92% when exporting from PL, only a huge file size increase.
Stenis
(Sten-à ke Sändh (Sony, Win 11, PL 6, CO 16, PM Plus 6, XnView))
11
Isn´t the main problem at 100% that it still is 8 bits?
Is the loss really in the compression when it is saved in 100%?
If then I think 100% is misleading.
Is a TIFF in 8 bits all that much better than a JPEG in 100%.
It is really strange that we accept such a lousy file format as TIFF 16 bits 2024. I read that some was complaining that his high res. RAW-pictures got 150 MB big when using TIFF as a mediate format with Topaz Photo AI 3 as a plug in to Lightroom.
The Americans sent Neil Armstrong to the moon already 1969. The first version of TIFF (3.0) was released 1986 and after that we are still stuck with TIFF as an intermediate format. I find that pretty strange.
8 bit is only one of the problems. By the nature of JPG compression it is always lossy. It cannot be lossless.
Other similar but âlosslessâ compression algorithms can use something like JPG as their primary compression but then they calculate the lossy âerrorâ (the original file minus the lossy file) and compress the error losslessly. Then they can apply that error after decompressing and restore the original file perfectly.
1 Like
Stenis
(Sten-à ke Sändh (Sony, Win 11, PL 6, CO 16, PM Plus 6, XnView))
13
A JPEG in 100% is nothing compared to a 16 bit TIFF.
Using the Adobe DNG converter and creating a DNG with the same sized embedded jpg as in the RAW files results in the same file size.
Difference are in the embedded jpg.
And you still need the RAW converter for it.