My Antioch Falling Horseman in extremely high detail

I am not very knowledgable of photography in general, but I am a old-school IT guy. There are some benefits to using the jpg format as you and Ras point out and I understand that, but jpg format is a lot like cloning in Star Trek. Each time you make a copy, the quality degrades by 5-10%. So only the very first jpg that you save (at 90% or better) is a true representation. Any time you make a change and resave it as jpg, it looses something. This includes the resizing and stitching operations. I don't use RAW, but do always try to use PNG when I have to work on an image. Even if the original image is a public domain jpg from the internet, i convert it to PNG first and keep saving it as PNG until I'm done working on it. Then I would convert to jpg for distribution, because I still have the PNG. Anytime I want to work on the image in the future, I would go to the original PNG and never the jpg. When I read your explanation, it sounded as though you saved your work as jpg and then did additional work to that jpg, and resaved it multiple times as jpg. This can and often does degrade the image. It literally starts getting blurry. The color starts doing wierd things at the pixel level too. My point was if you are creating a perfect image, at least save the original in a cumbersome lossless format that does not degrade when modified and resaved, because jpgs rarely stay perfect for long.

Hope this makes sense.

 

Log in or Sign up to hide this ad.

PS: I didn't see that last part of your post, but converting jpg to png stops future degradation, but any flaws from the jpg format are simpy preserved. It's better as described above to start with PNG and then convert to jpg at the end. However, as you say, it will definitely increase the processing time, and older systems with less RAM and/or CPU power may choke on the file, crash or lock up.

 

That is why I shoot RAW and convert to TIFF only dropping to 8 bit JPG at the last possible moment to produce something all people and machines can handle. I save only RAW files and my finished JPG since the TIFF's can be remade from the RAW and take up so much space. The only thing JPG really has going for it is that it is understood universally but, in communication, it is not important that your language is beautiful or can express concepts like no other if you are speaking to other people and do not share a language.

Probably 98% of the coin images I have shot in the last 20 years (digital phase of my life) have never and will never be printed on paper. Of that 2%, the majority were printed 4x6" and you can not tell the difference at that size between a massive file and one that can be shown on CT. I once carried a book of my coin photos to shows for reference but now I access my Google Drive files on my phone. Prints are now made only when there is a purpose. I have considered using an old TV on the wall to display whatever I want (changing the whim at will) rather than using framed prints. At my age, the guarantee is I will not be able to keep up with inevitable changes making the way I do things obsolete. Somehow I doubt that situation will change for today's young people.

My interest in coin photography began in late 1970 when my favorite magazine (Grossbild - International Photo Technik, 1970-4, page 8) ran an article on coin photography with a beautiful shot of a Marcus Aurelius sestertius take on a Linhof 5x7 view (Grossbild was the house magazine for Linhof and I used a Linhof 4x5 at the time).


I suppose the best of my images today are as good as that one but what you had to do to make that image with 1970 technology still blows me away. I still have that magazine and my camera but neither have been used in the last 20 years during which time my digital cameras have upgraded from toys to what is today a mid level hobbyist Canon RP. Somewhere along the way it became possible to shoot better coin photos than we dreamed in 1970. For my uses, I long since passed the point of needing images to the point of taking them to prove I can. It is like mountain climbers who climb mountains because they are there not considering why they might need to be on top of that mountain.

 

Thanks for the info Moe and Doug. As Doug mentioned, TIFF may be an even better choice as I know it is supported in a lot of the focus stacking applications and appears to be more widely used based on what I've read on forums.

It would be interesting to try converting all the original JPGs to PNG or TIFF and repeat the stacking, stitching, and editing process but for now I would rather try that next time I do this stack and stitch for a different coin.

I would also shoot RAW to begin with but for large focus stacks I doubt the benefits are worth it, it makes more sense when you have one photo per side per coin. If I was doing a stack that I knew I would get printed, then I would probably shoot RAW and go through the extra hassle.

 

I tend to agree that the trouble of starting off RAW in this case would be more than it was worth in this case since you have control of the lighting and could avoid having to make large adjustments. The Canon focus bracketing software (part of DPP4 which is free with their cameras) does files as part of the conversion but I have no idea how other brands work.

I see absolutely no reason to go backwards from JPG to a better format since the damage done saving once to JPG can never be reversed. I would reshoot rather than convert. When a camera saves a file to JPG, it makes certain decisions automatically that RAW practitioners prefer to make themselves. If the camera makes better selections than the human, the result might be better but most of us like things done 'our way' and that requires RAW. I really appreciate the color controls of the 16 bit formats over 8 bit JPG's when shooting nature/landscape/people but coins might not benefit as much if you get the lighting right in the first place.

 

Hypothetically though, you could take the JPGs from the camera and convert to TIFF for all of the stacking, stitching, and editing procedures, and then at the very end convert back to JPG and end up with a better quality photo than if you had used JPG for the entire process? At least that's how I'm understanding things but might be wrong. I realise that's backwards (going from 8bit up to 16bit) but would avoid the issue of degrading JPGs each time they're saved.

If that's the case, I'd be curious how it compares to: a) using JPG for the whole process, and b) starting with RAW and then going to TIFF before finally JPG. I guess Winter is as good a time as any for these kinds of tests!

Yeah if there was lots of dynamic range like you have in a landscape photo RAW would definitely make sense but I don't think I've noticed many 8-bit limitations unless I've blown out highlights and try to dial them back in afterwards.

But I'd be curious to shoot in RAW+JPG and stack the TIFF converted RAW files and stack the JPG files and see how different the end result is. Perhaps the stacking software makes fewer 'mistakes' with TIFF as well.

 

Well done, that is a beautiful coin and great tech. Thanks for the second post.

 

Perhaps I do not understand the software that is used to combine all those images into one. JPG's only degrade noticeably when they are resaved. If it uploads all the images at once and does all the work spitting out a single image, there would be only one generation used and one round of degradation suffered. If, however, the process produces sub-stacks and saves them as JPG's, it would seem better to convert. I can not imagine a program like that saving intermediates along the way in a lossy manner unless it were required to run in a limited amount of memory. Since we are already dealing with overkill to the point that we never would make a practical use of the image, the reason to ask this question is back to the big one: We want to know. The obvious answer is to try one both ways and see what you see.