File size and Print Quality

Howdy folks,

Something has me concerned, and I'm hoping it is just my ignorance. I'm just wanting to make sure I have the hightest quality files for printing, especially when presenting to a client. I'm using Lightroom 4.

When I import into LR, I do my inital edits, then export as a TIFF, moving the slider to the highest quality for export.

I shoot my photo with my 12mp Canon 5Dc. My original RAW files are somewhere around 11.2 or so mb.

Exporting from LR - my files as TIFF are 34mb!

After I do my final edits in PSE (10), save as a Jpg, and end up with a file that is anywhere between 4-12mb. Why such a divergence? I can expect a smaller file after converting to B&W, or cropping like crazy, but some files just end up smaller for seemingly no reason.

Can someone explain the relationship here - I just want to make sure if a client wants a 16x20 print, and their photo size is only 4 mb, there is enough resolution to print at that size. Mainly, peace of mind. Does file size have anything to do with MP?

thanks for reading my not-so-blissfully ignorant rant!

Michael

JPEG is lossy (er?) than TIFF. You won't see a difference in print. The size in megabytes has NOTHING to do with the inherent quality or resolution of the image itself...it's only an indication that there is a lot or a little detail (ie data). Shoot a landscape and then shoot with your lens cap on and you'll see two very different file sizes in MB even though they're both 12 megapixels.

The best way to determine if a photo is going to print well is to look at it. See if there is detail loss or banding because of over-compression. If it looks good, it'll print good.

Here is an extreme example. The first is Level 10. The second is Level 2. Make sure you don't have excessive amounts of what you see in Level 2.

Level 10

Thanks for the response! My eye must be severly untrained- I see very little difference in the two Level examples. A good thing?

Thank you for the explanation though- I'm clear on it now and resting a bit easier.

I appreciate your input!

Michael

CrackedLens wrote in post #15256965
Thanks for the response! My eye must be severly untrained- I see very little difference in the two Level examples. A good thing?

Look closer. Very clear difference in the skin detail and eyebrows for instance.

Here's a good read on LR jpg quality: http://regex.info …room-goodies/jpeg-quality

Look mostly in her right eye (left side of your screen) as that's where critical focus was and where it's most evident.

I probably should have used a different photo for greater effect. A car, for example, with its straight lines would have been more obvious.

This does show, however, that some images can be compressed quite a lot before they get really ugly so it's important to look at the photo itself rather than its EXIF data to determine whether it's printable. This also highlights why professional labs recommend 10 instead of 12...there's a significantly bigger difference in file size than file quality.

1. In RAW, each pixel (actually a 'sensel' at the RAW stage!) has 12- or 14- bit maximum for ONE color (R or G or B).
   
2. During the RAW conversion process, adjacent sensels are considered to derive and store 8-bits or 16-bits per RGB value, or 24-bits or 48-bits for a single Pixel -- stored as 16 bits for R, and 16 bits for G, and 16 bits for B = 48 bits!
   
3. So one single picture element changes from 12- or 14-bit sensel into 24-bit or 48-bit pixel...before any data compression occurs.
   
4. High amounts of JPG compression can smash subtle differences in RGB color down to a single value, losing subtle gradations and which can result in 'banding' seen in the sky, for example.
   

Gonna chew on that one awhile, Wilt. Thanks all for your insight in helping me understand this. I don't see any degradation in my photos (at least the images with good exposure, low iso, and acceptable focus) so I'm feeling better about print quaility.

Thanks again!

Michael

I'm sure I answered this same question phrased slightly different just a day or so ago...

The MB sizing etc. may be a fun exercise for some but the lab you want to do the printing has the last laugh.

If they are like most labs, both consumer and commercial, they are using equipment driven by software that has a optimal input file type. Generally, its sRGB at 300 pixels per inch. The RIP or other software used to drive the printer is probably happy with any thing between 150 and 600 pixels per inch.

The TIFF images for printing are generally reserved for CMYK printers used in publishing. Most often they need a specific ICC profile assigned that is fine tuned to the printing machine.

For book publishing, there are sophisticated pre-press systems that can make a proof print from the TIFF file with ICC profile assigned on an inkjet printer that the printer is willing to guarantee is an accurate representation of what the book will look like after printing.

Bottomline, 16x20 x 300 pixels is ideal. If your native camera file isn't large enough, have Lightroom upsize it on export from the print module as an sRGB JPG. Don't compress the file.

And to Wilt's point; JPGs are 8 bits per channel while TIFF can be 8, 16, 24 or even higher. 32 bit floating point TIFF files can be generated by combining 3 or more raw files in Photoshop. Once saved they can be processed in Lightroom 4.1 and higher. And then output as an 8 bit JPG for printing.

David, it sounds like you might be mixing up the question about file size (in MegaBytes) with the image size/dimensions (in pixels). If you understand the difference you may want to re-word the above reply to reflect that difference.

An image can have a given resolution in pixels but the resulting jpeg file size can vary widely related to the jpeg Quality setting (governing compression). Also, two images of the same resolution in pixels and saved/converted at identical jpeg Quality settings can have significantly different jpeg file sizes because the amount of detail in the image affects the amount of compression that is possible at a given Quality setting. Take a low-ISO well-exposed shot of a clear blue sky, then one of a scene with a lot of detail, Export/convert both at, say, a quality level of 8, and look at the difference in file size, pretty substantial.

These things have nothing to do with the specs of a particular printer because the resolution will be identical.

Now if you convert/save both the images as uncompressed tiffs, the file sizes should be pretty equal.

My understanding is that resolution should be selected to suit the output method and final print size such that generally the following settings should be considered:
for screen 72 dpi for an average 15" monitor this would be an image of 1080 pixels wide
for print 300 dpi is normally considered maximum needed to achieve good results and anything higher would be considered wasting space.
therefore for 16x20 print you would need am images sized 4800x6000

Jpeg compression should be considered in relation to available storage space and desired quality. Generally for Web only I use compression around 50% and for printing or probable further editing I keep it maximum quality. However I have seen tutorials and online comments suggesting that Jpeg doesn't need to be any higher than 10 for most applications.

I think the differences in different JPeg settings probably aren't as noticeable on screen as they would be in final print and therefore to try and show the actual quality difference using on screen pictures is not really going to be beneficial.

tonylong wrote in post #15262394
Now if you convert/save both the images as uncompressed tiffs, the file sizes should be pretty equal.

Actually with identically sized images the size of the image data within a pair of TIFF image files will be exactly the same. There may be a small difference in the total size of the file due to differences in the data stored in the EXIF information. Each extra charachter stored in any field of the EXIF will add 1 byte to the file size (assuming an 8bit character encoding).

Alan

birderman wrote in post #15262879
My understanding is that resolution should be selected to suit the output method and final print size such that generally the following settings should be considered:
for screen 72 dpi for an average 15" monitor this would be an image of 1080 pixels wide
for print 300 dpi is normally considered maximum needed to achieve good results and anything higher would be considered wasting space.
therefore for 16x20 print you would need am images sized 4800x6000

Why 72 ppi?

I work with two monitors, a laptop and a 23" monitor, and both of them are 98 ppi. Apple's Retina display can be as high as 326 ppi in an iPhone or 264 ppi on an iPad.

tonylong wrote in post #15262394
David, it sounds like you might be mixing up the question about file size (in MegaBytes) with the image size/dimensions (in pixels). If you understand the difference you may want to re-word the above reply to reflect that difference.

An image can have a given resolution in pixels but the resulting jpeg file size can vary widely related to the jpeg Quality setting (governing compression). Also, two images of the same resolution in pixels and saved/converted at identical jpeg Quality settings can have significantly different jpeg file sizes because the amount of detail in the image affects the amount of compression that is possible at a given Quality setting. Take a low-ISO well-exposed shot of a clear blue sky, then one of a scene with a lot of detail, Export/convert both at, say, a quality level of 8, and look at the difference in file size, pretty substantial.

These things have nothing to do with the specs of a particular printer because the resolution will be identical.

Now if you convert/save both the images as uncompressed tiffs, the file sizes should be pretty equal.

I had hoped the first sentence would get across that file size is irrelevant when sending images to a lab for printing.

I answered in pixels because that's the important consideration when creating a file for printing. The file size is a result of the pixel dimensions required for printing, bit depth and file format. Again, irrelevant.

The fact the OP worded the question poorly should be a barrier to giving them useful information.

The 72 pixel per inch mentioned in an earlier post is another good example of dated information being perpetuated. 72 pixels per inch was the resolution of older monitors. Now LCD and LED flat screen monitors can be given a variety of resolutions during setup. If one wants an accurate on screen ruler then they have to match the pixel per inch in the software with the screen resolution setup in the device driver.

Again, not really relevant for file size, print resolution or any thing other than the size of the ruler on screen relative to the image on screen.