burkphoto Loc: High Point, NC

IMHO, HP is a distant third in the PHOTO inkjet world. Epson makes the best high end pigment ink and pigment ink printers. Canon makes the best dye ink and dye ink printers. The Epson P-series is all pigment ink. Canon's Pro 100 is dye ink based. The Pro 100 is very reliable and affordable, and has a 13" throat.

I put several earlier Epson Ultrachrome pigment ink printers in a photo lab in 2003. We used them to make our largest prints, up to 44" by 96". So long as we used their ink, and their photo paper, the results were an order of magnitude better than the prints coming off our conventional RA-4 chromogenic process silver halide systems, using Kodak's best pro papers.

I have prints I made in 2004 on both systems... The chromogenic prints are fading and yellowing, while the Ultrachrome prints from the Epson look like they have not changed at all! In fact, new prints from the same files match the ones made 15 years ago on the Epson.

There is no correlation between your Nikon and any brand of printer... Once you have an image you like, you can print it on anything. It will look good if you adjusted it on a calibrated and profiled monitor.

Imagemine Loc: St. Louis USA

I have a Canon pixma pro-100 with 8 ink cartridges and have gotten good results, I'm a Nikon shooter also

burkphoto Loc: High Point, NC

Raw files from my 16MP camera are just under 20MB (19.6 to 19.8). I just exported the same file five different ways.

An sRGB JPEG exported at 100 quality from an uncropped file is about 9.5 to 10.5 MB, depending on subject content (detailed subjects take up more space, due to the way JPEG compression works).

An Adobe RGB JPEG exported at 100 quality from an uncropped file is about 9.0 to 10 MB, depending on subject content (detailed subjects take up more space, due to the way JPEG compression works).

A 16-bit uncompressed TIFF in sRGB exported from an uncropped file is 95.6 MB.

A 16-bit uncompressed TIFF in Adobe RGB exported from an uncropped file is 95.6 MB.

A 16-bit ZIP compressed TIFF in Adobe RGB exported from an uncropped file is around 81 MB. (File size will vary according to detail content, due to ZIP compression.

All that is perfectly normal! There is nothing strange going on here. Raw files contain much less data than do the bitmapped TIFF files created from them. For one thing, these are 12-bit files being interpolated to 16-bit files. They are raw sensel data being processed (demosaiced) to create file pixels with three of those 16-bit values per pixel.

In my next post here, I'll include enlarged segments of files you can examine and compare.

burkphoto Loc: High Point, NC

(See above)

Here is the same *cropped* 16MP image, first the original pixels as processed from raw in Lightroom to an sRGB JPEG. Original is raw converted to JPEG.

The next three are 1920x1080 pixel SECTIONS of a 40x22.5 inch, 240 PPI blow up of the image shown above. View in download at 100% and step back from your monitor (normal viewing distance for this size print is 50-75 inches! The idea here is to simulate what a viewer of this print would see at a normal viewing distance.

First is an Adobe RGB 16-bit TIFF (not sure it will display here).

Next is an sRGB 8-bit JPEG

Next is an Adobe RGB JPEG

All JPEGs are made at 100 quality in Lightroom or 12 quality in Photoshop (same actual effect, since Adobe uses different scales in each program. One point in PS = 8.5 points in Lightroom.

Note, your browser may not display the Adobe RGB files properly until you download them! This is a known issue with UHH. The color shift disappears on download, at least on the Mac running Safari.

(Download)


(Download)


(Download)


(Download)

natureslight Loc: Winona, Minnesota

Yes, the color shift did disappear upon download [interesting]. Was the 240 ppi a choice, or is that what was produced when cropped, because it seems 300ppi would be the choice for best resolution? I'm going to study some adobe videos on file sizing for export because there are a lot of variables available in lightroom, and I don't understand all the differences, which seems to be quite complex at this point. My goal is to understand what is needed for different size prints for good resolution and color prints that will match what is on the monitor. Some say sRGB is good enough, and others say to go with Adobe RGB. After viewing your file samples I honestly couldn't see any difference. You can see a huge difference when viewing a TV with 4K resolution versus one with 1080p, so is this not the same concept with image prints? You stated earlier not to mistake file size with image quality. That is confusing as it seems more data should equal better quality images provided it was captured correctly, but your samples don't support that either, so it's obvious there is a lot to learn here. Thanks again for all your help.

burkphoto Loc: High Point, NC

I was privy to a lot of information about print resolution when Kodak was guiding us from their research.

The dirty little industry “secret” is that an 8x10, viewed at 13 inches, only needs 240 original file pixels per inch to achieve extinction resolution — the point where your eye cannot see any pixels. Those pixels become dots. Different printers use different numbers and sizes of dots to make photo prints.

The myth of 300 dpi (PPI, really, because dots are used in scanners and printers, but pixels are just numbers in files) is very old. It traces back to MAGAZINE EDITORS’ requirements that an image survive 50% more enlargement if they had to make a last minute editorial decision.

You need MORE resolution for smaller prints. A 5x7 needs 300 PPI. A 16x20 is fine with 180 PPI.

But... you still need 240 PPI as minimum input to the printer driver. If you like 300, go for it. But at 50” from a 40x30, it won’t make any difference.

natureslight Loc: Winona, Minnesota

I really appreciate you taking the time to share all of this. I'm going to keep researching and learning about file sizing and image quality from lightroom. Thank you again for all your help. I wish I had your knowledge. Mark

Novicus Loc: north and east

yes indeed, I too have learned some, and thanks to natureslight for bringing this up,...hoping that Burkphoto will Not get tired of us anytime soon,the Help is Greatly appreciated :-)

Novicus Loc: north and east

I will look into this, it is appreciated, thank you

Wallen Loc: Middle Earth

Assuming we call dpi (dot per inch) & ppi (pixel per inch) as the same banana;
Think of resolution as the amount of sampling spot per specific area.
Example. 10dpi means every inch is divided by 10. The image is sampled 10 times vertically and 10 times horizontally. A square inch of a 10dpi image will have 100 colored spots(pixels).

A typical computer monitor has a 72ppi resolution. Hence, a 72dpi image will look fine when viewed at normal distances.

File DPI in truth is an illusion. An image actually have no fixed resolution. What it has, is a default viewing size and finite amount of pixels.
Example; You have a 3x4" 300dpi image file (7X10cm approx).
When opened and viewed, two things can happen.

Because the monitors screen is only 72dpi...
1. A full 900x1200 image is displayed or
2. the 300dpi file is displayed at a default size of 3x4 inches at 72dpi!

If the 3x4" imgae is zoomed in 200% (2x), then that same image will be displayed in the monitor as 6x8" at 72dpi (150dpi available).
If we zoom again (4x), that would be 12x16" at 72dpi (75dpi available)
The next time we zoom, individual pixels will show.

12x16 inches is about the same size of an A3 paper. Meaning we can print that image to A3 size and it will be the same quality/sharpness as the image we viewed 4x zoom in the monitor.


About the file color & size please read this post as backgrounder
https://www.uglyhedgehog.com/t-585722-1.html

My short answer would be, "Do not bother yourself with the file size. It is a complicated mess. Different file types and adjustments within the files themselves will dictate the size. As long as the image looks great, let the file size be".

Long answer is below and is purely subjective. This is just to visualize the variables involved:

A 3x4" 100dpi 8bit image means;
it is 300x400 or 120,000 pixels in size (notice DPI flies out the door).
Each pixel has 8bits per channel. That is 2x2x2x2x2x2x2x2 or 256 variations (computer geeks smile here) per channel.
Each pixel has 3 channels. RED, GREEN & BLUE. It is the combination of red, green & blue light and their respective intensity defines the color of each pixel. This is the basis of the RGB colors.

So our Theoretical* FILE SIZE would be;

FILE SIZE* = PIXELS x CHANNELS x COLOR BIT + OVERHEAD

= (300x400)(3)(8**)(OH) **8bit=1byte
= (300x400)(3)(1)(OH)
= 360,000bytes* +(OH*)
= 360kb* +(OH*)


*We say theoretical because aside from the pixels themselves, there are other overheads that need to be accounted for, some of which are the file type header, bit address etc. and would vary from file type to file type and also within the file themselves depending on user settings I.E. compression. Thus the actual file size can be way smaller or larger than the computation above.

Its a mess so just learn to live with it


Going back to colors, we have shown that an 8bit color channel is actually 256 shades for each R,G & B channels.
Meaning an 8bit channel file can show millions of colors. That is 256x256x256 or 16.77million colors!

This is why printers say 8bit JPG file is enough.

About TIFF, It is a file containing (many) other files. Think of it as Folder. Like a folder it contains the image/files and also tags or instructions on how to handle or interpret the files inside the TIFF. Similar to a word document with pictures telling the reader how to interpret the pictures and explaining the colors. So it is a good alternative because of the embedded meta data etc, the output can be accurately reproduced when read on another computer.
And since TIFF can be 16bit or more, it can be a more accurate reproduction.




Continuing from the above. The caveat here is that prints can never show exactly everything that the monitor can show. Quality in mind, Monitors produce light. Prints only reflect or in case of backlights, reduce from a light source.
If we wanted full accuracy, we should adjust the monitor to what the print can do and not the other way around. Meaning, just do your best in your part and try to find a printer that is acceptable to your taste.



Mark[/quote]

Below is an example. The bad image is actually a bigger file size than the good one below it. Quality starts at the capture. Generally speaking, even with all the software magic and user skills at post, it all goes down from there.

(Download)


(Download)

natureslight Loc: Winona, Minnesota

Thanks to you as well for all the time you've taken to help those of us who need it to understand all of this. I do think I understand most of what you've described here, but I still have a ways to go. Also, getting back to pixels, another variable is the camera sensor. Just because a sensor has more pixels, doesn't necessarily make it better as I know pixel size is a factor as well, and do you know if pixel size correlates to dynamic range for capture? Sorry, not trying to make this any more complex than it already is, just trying to understand as much as I'm capable of and you seem quite knowledgeable as well as some others in this thread. You do state "And since TIFF can be 16 bit or more, it can be a more accurate reproduction" which goes back to my original question about tiff versus jpeg. That is exactly what I thought, so when it comes to larger prints, it seems tiff would be better, yet a recent post from Burkephoto showed some samples comparing a tiff to jpegs and I could not see a difference when viewed at the correct viewing distance. However, that being said, I still here photographers saying if you want more accurate color representation, than print from tiff. So my original thoughts that the majority of these print labs only dealing with jpeg files may be correct. Either they can't or don't want to work with tiff files versus their claims of jpegs being good enough. Uhgg, not sure what to believe. Thanks again for all your help on this subject. Mark

burkphoto Loc: High Point, NC

The purists’ views are often misguided, because math seldom tells the whole story when it comes to visual perception. When you work in a lab that makes millions of prints from thousands of photographers’ work, you start to see the patterns and ‘practicals’ of photography.

High bit depth is most useful during image capture and manipulation. Once a 16-bit image is adjusted to look right, conversion to 8-bit loses little *perceivable* information.

Where printing from 16-bit files in wide gamut color space is useful, is in the high end, “giclee” inkjet world, and in certain advertising applications where matching a saturated logo color or product color is required by a finicky client.

The best scenario for doing that would be in a photo studio where the wide format Epson or Canon printer is connected to a computer running Lightroom and Photoshop.

The camera is also tethered to that computer, and the environment is set up for the standard print viewing conditions specified by PPA.

The image is captured, imported into Lightroom, and adjusted on a high end monitor until it looks like the product.

Any retouching or compositing is done in Photoshop.

The image is soft proofed using the profile for the printer/paper/ink combination in use. Then it is test printed directly from Lightroom through a 16-bit driver.

Finally, the print is held up against the product and compared. Tweaking ensues...

When the results are acceptable, resizing is done in On1 Resize. Then the image is printed.

That is a VERY expensive and labor intensive process.

rehess Loc: South Bend, Indiana, USA

At least some of the issue is because not everyone has exactly the same visual acuity, and some of us have an inferiority complex that our photos have some difference that certainly someone can see.

natureslight Loc: Winona, Minnesota

That all makes good sense. Sometimes having a little bit of knowledge on a topic, helps one to know who to listen to, as there are many voices on this subject. I am listening to you, for one. Thanks

Novicus Loc: north and east

Indeed , I too, wholeheartedly thank Everyone for their time and Patience ,and certainly for their Ability to explain things so that even i can understand.

Having knowledge is one thing, being Able to Convey said knowledge is a different matter, ..as I have tried, asked and still had trouble understanding the bits and pieces of Info gathered here and there , yet , now finally getting an Inkling of steps to take.

So, if I understand Correctly,...Set/ Use Recording Equipment ( Camera /Lenses etc. ) to Highest Standard.

Which means TIFF on Camera and then Downscale to a practical size with which the Computer/Printer is Comfortable with ( being able to cope within its capacity )

Said in darkroom terms : Start with a large negative..well exposed ..correctly developed etc., so as to be able , even after Cropping ,to produce an A3 in Utmost Quality.

Now, do i Need to resort to TIFF ( slowing camera severely ..nikon D3s / D3X ) or could I get away with NEF ( Nikon RAW format ) converted / scaled down accordingly using Nikon Capture NXD ,...or even Jpeg ERI ( Kodak Jpeg format ...Kodak SlrN ) using Photodesk.

I apologise to natureslight if I appear to be hi-jacking his posting as such is not my intent, it is just that my take is only slightly different as I intend to " do printing myself " as opposed to his wanting to "have it done", for I could Not afford otherwise.

Thanks a million :-)