Hi Res question

ArizonaMike

Mu-43 Top Veteran
Joined
Jan 19, 2018
Messages
798
Location
Arizona, USA
How exactly were you processing the hires raw ORF file?
I had been using Dxo's PhotoLab3. It recognizes both the ORF and the ORI files and properly displays them. What I had done was the following:

1) Select the ORF file and crop it 1:1,
2) Select the ORI file and resize it to the size of the ORI file. I made several attempts, sometimes using PhotoLab3 to resize and sometimes using Topaz's AI Gigapixel to resize,
3) Once the ORI file was resized I cropped it 1:1 as in (1) above,
4) I then output both files us 100% quality jpgs with no further adjustments.

That gave me 2 jpgs cropped identically and the same size so they could be compared.

With the jpgs it was simpler.

1) Select the high res jpg and crop it 1:1,
2) Select the regular jpg and resize it to the same size as the 50MP jpg using PhotoLab3,
3) Crop the resized jpg 1:1,
4) Output both to jpgs, 100%, using no presets (ie, no further adjustments).

You weren't processing the ORI file by mistake instead of the ORF file, by any chance?
No, I was not.

I know the hires ORF can be processed correctly in Olympus Workspace
I will try that tomorrow and see if it makes any difference.

Have you remembered to set the camera to give you 50MP JPGs and not 25MP JPGs?
Yes. I have the 50MP jpg written to one card and the 25MP jpg written to the second card. I also have the high res raw and regular raw written out as well.
 

Machi

Mu-43 Top Veteran
Joined
May 23, 2015
Messages
680
@ArizonaMike
Try Rawtherapee. It's free, it has deconvolution sharpening and generally it's better than Olympus Workspace for recovering small details thanks to the lots of deBayer algorithms.
 

ArizonaMike

Mu-43 Top Veteran
Joined
Jan 19, 2018
Messages
798
Location
Arizona, USA
I have to admit that I must have completely forgotten all of that over the couple of years because until you told me I did not remember doing so. I do remember discovering this issue when I first go the M1.2 and mentioning it on a forum, but I did not remember finding the same results that I found this time.

As it is I had completely stopped using the high res functionality in my M1.2 and only remembered about it when I decided to buy a second camera so I would have one dedicated to birding and the other dedicated to normal walk-around functionality. When I looked at the available functionality on the PEN-F and the M4.2 I saw high res for both and that is what prompted me to start this thread. I guess from what you say that it is a duplicate of what I posted several years ago, but I was not aware of that when I started this thread.

I feel a bit foolish for having forgotten that I had posted all of that before and even worse for not remembering that there was an answer to my original question, so I guess I am grateful for you telling me. Perhaps this time I will remember ...

do you want to know what's wrong with your RAWs or no?
Well, of course. If there is any issue with the raws. This thread has mostly convinced me that I should be ignoring the raws and paying attention to the jpgs. If the 50MP jpg is fine then I assume the high res raw is as good as it should be, regardless of how it looks when processed and compared to a resized image.
 

Machi

Mu-43 Top Veteran
Joined
May 23, 2015
Messages
680
I have to admit that I must have completely forgotten all of that over the couple of years because until you told me I did not remember doing so. I do remember discovering this issue when I first go the M1.2 and mentioning it on a forum, but I did not remember finding the same results that I found this time.
....
I feel a bit foolish for having forgotten that I had posted all of that before and even worse for not remembering that there was an answer to my original question, so I guess I am grateful for you telling me. Perhaps this time I will remember ...
That happens to everyone. I know that well! :)

Well, of course. If there is any issue with the raws. This thread has mostly convinced me that I should be ignoring the raws and paying attention to the jpgs. If the 50MP jpg is fine then I assume the high res raw is as good as it should be, regardless of how it looks when processed and compared to a resized image.
If you are happy with jpegs then you can ignore problems with RAW's processing but as I wrote in the previous post - you can try Rawtherapee. It's not easiest software in the world but It can provide great results, especially in the case of hi-res files.
 

Growltiger

Mu-43 Hall of Famer
Joined
Mar 26, 2014
Messages
2,038
Location
UK
The hires raw files need special processing compared to any other raw files, so perhaps it isn't surprising that not all raw processors support them properly. Stick to those that are known to support them.
When comparing hires and enlarged normal res, it isn't fair to use fancy sharpening or enlarging tools from Topaz (Sharpen AI or Gigapixel AI). If you do that you would need to apply the same tools to the hires file too for a fair comparison.
You are not the first person to forget they asked a question before. What is really good as that you got the same answer both times.
 

RAH

Mu-43 Top Veteran
Joined
Dec 1, 2013
Messages
797
Location
New Hampshire
Real Name
Rich
I think that there is a reason why Olympus makes 50MP jpgs out of 80MP raws. From what I have read, they decided that there is no more detail to be gained by making an 80MP jpg, so whatever they do software-wise, they get the best detail for the jpg by manipulating the camera results DOWN to 50MP. I mean, there has to be a reason why they didn't make their jpgs 80MP, right?

Wrontniak has an interesting article about this topic, with sample images here:

http://www.wrotniak.net/photo/m43/em1.2-hires-2.html

At one point he says, "I will not rush to any conclusions yet, although at the moment (and based on these samples) I see no advantages of using 80MP images (SHR) over 50MP ones (HR)."

(My apologies if this article is already mentioned in the long threads already mentioned).

Also, from my experiences with manipulating just standard raw files (not hi-res) with a non-manufacturer editor, you have to do A LOT of tweaking of the settings to even match the camera's jpg images. I mean, for example, say you have a jpg with a blown-out sky that you are hoping you can fix by using the raw file. So you use say a program like Rawtherapee. It CAN often fix the problem you want fixed, like the blown-out sky.

But ALSO, you have to do all the things that the camera's jpg engine does for you - especially noise reduction, sharpening, etc, etc. So I often find myself fixing the problem (the sky) but then comparing my raw results with the jpg from the camera, just to get everything else OK.

(I use Rawtherapee; I agree with Machi that Rawtherapee is a great raw program; also good for just jpg editing too if you don't need layers).

So, once you get into raw manipulation, all bets are off when you are going to do image comparisons, IMHO. You are manipulating so many things that there are too many variables. The same is somewhat true with upsizing (and especially downsizing, although that is not at issue here) - as soon as you DO anything with an image, it makes it harder to compare things. That is why I posted a 100% crop of my jpg results WITHOUT upsizing it to compare vs the hires results.
 

ph.

Mu-43 Regular
Joined
Feb 14, 2010
Messages
55
Location
norway
Sirs,

When experimenting with the high-resolution, I found that placing my PenF on the ground was not good enough as support, but when anchored to the compact Leitz table tripod, it would give far sharper images than the normal resolution snaps. The optic I used was the Leitz Apo Macro Elmarit 100mm which should be sufficient to evade detail mush..
p.
 

ArizonaMike

Mu-43 Top Veteran
Joined
Jan 19, 2018
Messages
798
Location
Arizona, USA
When comparing hires and enlarged normal res, it isn't fair to use fancy sharpening or enlarging tools from Topaz (Sharpen AI or Gigapixel AI). If you do that you would need to apply the same tools to the hires file too for a fair comparison.
I agree and I did not use Sharpen AI. I did use Gigapixel for one test but used Dxo's PhotoLab3's normal resizing functionality for the latest tests and posts.

I can't help but think that any resizing operation, regardless of what software, will automatically apply sharpening but I don't know how to compare the high res raw and normal raw images without resizing the normal raw image so that you are looking at nearly identical crops.

One thing I purposefully did not do was sharpen the high res photos and I will try to do that today but the thing that started me off on this whole thread was the dpreview high res raw images posted in their reviews (or overviews) of the PEN-F and M5.2. Those raw images were tack sharp while mine never were, and hence my questions.

While I assume that if my high res jpgs (which are, of course, sharpened by the camera algorithms) are OK then the raws are probably the best I can get, I will get a chance to see. On another thread on this forum I was asking about comparisons between the M5.2 and the PEN-F because I was planning to order a second camera. That order is due to be delivered today and I will get a chance to try its high res functionality and see if those raw images are also less sharp than the jpgs.
 

ArizonaMike

Mu-43 Top Veteran
Joined
Jan 19, 2018
Messages
798
Location
Arizona, USA
I found that placing my PenF on the ground was not good enough as support, but when anchored to the compact Leitz table tripod, it would give far sharper images than the normal resolution snaps
I will get a chance to test this when the PEN-F arrives today. Presumably I can take high res images with both the M1.2 and the PEN-F and see if there is any noticeable difference.
 

RAH

Mu-43 Top Veteran
Joined
Dec 1, 2013
Messages
797
Location
New Hampshire
Real Name
Rich
One thing I purposefully did not do was sharpen the high res photos and I will try to do that today but the thing that started me off on this whole thread was the dpreview high res raw images posted in their reviews (or overviews) of the PEN-F and M5.2. Those raw images were tack sharp while mine never were, and hence my questions.
I just looked at the latest dpreveiw review of the E-M5III and the hi-res raw image is NOT as sharp as the jpg, at least to my eyes. If you toggle between jpg and raw with the pull-down, the raw looks absolutely blurry.
 

ArizonaMike

Mu-43 Top Veteran
Joined
Jan 19, 2018
Messages
798
Location
Arizona, USA
One additional comment before I try Rawtherapee.

I purposefully did not try to sharpen the high res raw image because I was not explicitly sharpening the resized image I was using for comparison, but just now I loaded the high res image into PhotoLab3 to see how well the raw image does sharpen if that is applied, and it did very well. It took a very long time in PL3 to do the sharpening because the image is so large, but the end result was very nice afterwards.

I have updated my copy of Olympus Workspace and I will try that as well as Rawtherapee to see how they act, and I will try ACR to see how well it does as well.
 

RAH

Mu-43 Top Veteran
Joined
Dec 1, 2013
Messages
797
Location
New Hampshire
Real Name
Rich
ArizonaMike - Have you looked at the raw hi-res samples I mentioned in the DPReview review of the E-M5III? Do you agree with me that they look blurry and much worse than the jpgs?
 

Growltiger

Mu-43 Hall of Famer
Joined
Mar 26, 2014
Messages
2,038
Location
UK
ArizonaMike - Have you looked at the raw hi-res samples I mentioned in the DPReview review of the E-M5III? Do you agree with me that they look blurry and much worse than the jpgs?
Which software was used to make them? The only software we know for sure does the same as the camera is Olympus Workspace. It is results from that that need to be used to compare with others. I will be interested to hear the results of ArizonaMike's tests with Olympus Workspace vs Rawtherapee vs ACR vs JPG from the camera..
 

ArizonaMike

Mu-43 Top Veteran
Joined
Jan 19, 2018
Messages
798
Location
Arizona, USA
ArizonaMike - Have you looked at the raw hi-res samples I mentioned in the DPReview review of the E-M5III? Do you agree with me that they look blurry and much worse than the jpgs?
I looked at the M5.2, not the M5.3 since I was not considering that camera.

The M5.2 high res raws were tack sharp and that is what got me started on this thread - why were my M1.2 high res raws not that sharp - but I will take a look at the M5.3 raws and see how they look. I will also try to compare them to the jpgs.

As I mentioned I did find that the M1.2 high res raws, blurry though they were, sharpened nicely in PhotoLab3. I need to take a look at those in both ACR and Rawtherapee. I did try Olympus Workspace but I think it requires a bit of a learning curve because I did not see any of the jpgs and I am not familiar with the adjustments.
 

RAH

Mu-43 Top Veteran
Joined
Dec 1, 2013
Messages
797
Location
New Hampshire
Real Name
Rich
I'm thinking that maybe DPreivew has changed the way they handle raw files. How much processing to do is always a question, so perhaps the way they do it has changed since the E-M5II. They don't say much about the issue.
 

RAH

Mu-43 Top Veteran
Joined
Dec 1, 2013
Messages
797
Location
New Hampshire
Real Name
Rich
I see that for the E-M5III review, under hi-res, it says "We should also note that the E-M5 III's 80MP high-res Raw files will respond well to quite a bit more sharpening than our standard studio test protocol, so we'd recommend downloading the files yourself to check out how they perform for your workflow. " I am mentioning all this DPReview stuff because I think it points out pretty clearly the need for a lot of sharpening with the raw results. Scary stuff, actually, because when working on images, at least I try to not go too far with any kind of processing, but I know that at least with noise reduction, you need to use a lot on raws, and it is looking like these hi-res raws need a lot of processing just to match the jpg results.

I think I'd almost argue that the out-of-camera hi-res jpgs give you what you need. I mean, the color is already enhanced by the hi-res process (and I don't think you lose it from the conversion to jpg)., from what i have read about the sensor shifting technique - it's not just about resolution. Everything gets enhanced. So perhaps raw isn't really all that necessary, IMHO. I mention again that there must be some reason that Olympus downsizes their jpgs from 80 to 50. As far as I know, this is not a cropping, it is a downsizing. How is a person going to do that on a raw? I mean, it's probably some special processing unique to these hi-res files, and perhaps not even something you could do to the raw even if you wanted to. Interesting stuff, I think.
 

ArizonaMike

Mu-43 Top Veteran
Joined
Jan 19, 2018
Messages
798
Location
Arizona, USA
I'm thinking that maybe DPreivew has changed the way they handle raw files. How much processing to do is always a question, so perhaps the way they do it has changed since the E-M5II. They don't say much about the issue.
I guess I don't understand what you mean by your comment.

As I understand it a raw image is simply a data file so making it available for download does not require any processing other than copying it from the flash memory to the hard drive to make it available. What processing would dpreview be doing other than linking to it for download? What am I missing about your comment?
 

RAH

Mu-43 Top Veteran
Joined
Dec 1, 2013
Messages
797
Location
New Hampshire
Real Name
Rich
Well, that is true for what you download, but I was referring to just the screen display, with the pull-downs where you can see the results from various areas of the picture, etc. and compare it vs other cameras, etc. What you are seeing on that screen are jpg or png versions of that - you cannot display a raw on the internet.

Plus, note the quote I had in the post I was typing while you were typing, about 5 minutes ago - they specifically say that raw should be sharpened.
 

ArizonaMike

Mu-43 Top Veteran
Joined
Jan 19, 2018
Messages
798
Location
Arizona, USA
"We should also note that the E-M5 III's 80MP high-res Raw files will respond well to quite a bit more sharpening than our standard studio test protocol
And this is what I saw when I processed the high res raw in both PhotoLab3 and ACR. Just a bit of sharpening make the high res raw at least as sharp as the resized original raw. I will process them with Workspace and Raw Therapy as well but don't expect to see any real difference.

To be clear, when I downloaded the PEN-F and M5.2 raws from dpreview I did not explicitly sharpen them. I loaded them into PhotoLab3 and, of course, it applied whatever default processing it uses on the raw. The M5.3 high res raw in PL3 was tack sharp without any additional sharpening and my M1.2 high res raw was not, and that was the reason for my original post in this thread. Why the difference?

All of this may be a more theoretical issue than real issue as I have no implicit problem with sharpening, provided it is kept to a minimum, and would happily sharpen high res raw images when processing. I am sure that Olympus does that sharpening in-camera when it produces the jpg so I could just do it as well. And I have, from time to time, played around with the idea that I should normally just process the Olympus jpgs unless there is some issue, and ignore the raws unless needed.

I think I'd almost argue that the out-of-camera hi-res jpgs give you what you need.
And, for the most part, I agree with you. I seldom actually need the raws as the M1.2 does a wonderful job with the jpgs. It is just that I have all of these raw converters on my system and it seems a waste to not use them ... :)
 

RAH

Mu-43 Top Veteran
Joined
Dec 1, 2013
Messages
797
Location
New Hampshire
Real Name
Rich
Forgetting hi-res raw vs jpgs for a minute, for regular images I usually only use the raw if I need to, usually finding the jpgs entirely satisfactory. I know that the "Raw Police" would not be happy with this idea, but that's too bad. If I can withstand the tough scrutany from the Manual Mode police (I use Aperture Priority), I can do the same with the Raw.

Having said that, I DO use the raw files I'd say maybe about 25% of the time, depending on the type of photography I am doing. It's great to have them available.

As I mentioned earlier, I think the argument for just using the jpgs is even stronger for hires mode. It's too bad that the camera doesn't have the option of producing a non-compressed non-raw 50MP image (for example, a TIF). I have always used Large/Fine setting for my jpgs, but I am now thinking that at least for hires I should be using super-fine, to get the absolute best image, which is the whole point of hires.

Also, I guess it might be a good idea for me to take a look at how the Olympus raw program handles hires raws. As you say, there is a learning curve and I'm not keen on learning it, but for these hires raws, maybe it knows best. For example, maybe it has an option to produce a 50mp jpg like the camera does, but allowing you to do extra processing on the 16-bit raw first. (I haven't looked).
 

Latest posts

Links on this page may be to our affiliates. Sales through affiliate links may benefit this site.
Mu-43 is a fan site and not associated with Olympus, Panasonic, or other manufacturers mentioned on this site.
Forum post reactions by Twemoji: https://github.com/twitter/twemoji
Copyright © 2009-2019 Amin Forums, LLC
Top Bottom