Pen-F : the best Olympus ever for image quality ?

Jorginho wrote:

According to who? Not according to dpreview, not according to DxO (1/3 of a stop at best), not according to mirrorless comparison (extensive review that only gives Fuji a lead beyond ISO6400) and where DR of the Em1.2 is visibly better according to them (and noise is worse). not according to Gordon Laing (although that is JPG and not so extensive, but still), not according Petapixel (albeit only measured DR, where the EM1.2 wins from Fuji..again...bnut slightly).

You can't pick and chose DXo when it suits you. DXo clearly shows about 1 stop more DR on the Fuji.

? DxO never ever tested any Fuji. So you are quite right that I cannot pick and chose since there is nothing to pick and chose there.

There is. We were talking about modern APS-C. That's Sony A6500. It's also pretty easy to see that having copper-wise design for a sensor of the new Xpro2 sounds a bit too familiar with the Sony sensor for ASP-C A6500 they have.

So yes, you can compare. Modern APS-C - Sony A6500 and/or Fuji. Take your pick.

As for mirroless- they clearly mentioned in their comparison that the shadows could lift more on the Fuji- DR range is total DR. For DXo, ok, 1/3rd of a stop better ISO, though other places have mentioned more, but let's go by Dxo.

Let's look at their conclusion to keep it simple: "In terms of image quality, many will assume that a larger sensor is better, but the E-M1 II proves that Micro Four Thirds can hold its own even in the company of the best APS-C sensors on the market. While high ISO performance remains an advantage for the APS-C X-Trans sensor, particularly past 3200, the Olympus camera doesn’t pull any punches when it comes to dynamic range."

I looked at their view and examination. They did mention in that test they were talking about DR that the shadow range on the fuji was better. Well guess what? Total DR = highlights + shadow range. I have seen this mistake being constantly done by many review sites.

That said to be clear, I do think the mirrorless comparison does have value.

Holding its own against the best APS-c sensors on the market....doesn't sound like your assertion which was and is "The new Sony/Fuji sensor pulls ahead of EM1.2.

You are picking and choosing again. If you picked DXo why are you ignoring the A6500 performance? That's modern APS-C no?

The ISO difference is about one stop where it normally has been between m43rds and APS-C. And the DR of best APS-C (or even old APS-C like the sensor in Pentax K-5) is much better still."

Or does it...?

Read above.

So you get 1/3rd better stop ISO on Fuji, ~1 stop better color sensitivity, ~1 stop better DR. That's Dxo.

? No? Where? There is no Dxolabs score. We do have photphotons who do something with Dxo scores....And here is what they found...

Sorry, I am looking at the Sony A6500 stats, which is roughly what the current Fuji sensor is doing.

First table...Fuji is not better at al at DR

But I have a feeling you are talking about something else may be. Can you give me a link to DXO tests of the Fuji?

Please read the clarification above. Even if Fuji is behind, the point of modern APS-C still stands due to DXo's Sony documented performance.

The mirrorless comparison website as good as it was missed the point that if you have more shadow range, well, that's more total DR range.

No. They note more information in the brighter areas on the OMD and more information in the darker areas with the Fuji.

Not quite what I remember. I'll double check later.

Also: DR of K5 at base ISO is much better. And that is the only thing. It is worse elsewhere..

Actually the K5 still does better DR across the range for the most part, but the print value gives better DR.

Only the printvalue matters really. Note that if you use screenvalue D810 has the same colour sensitivity as the EM1m2 and K5 etc....I used D800E...the printvalue is clearly the better indicator.

I wouldn't say only the print value matters, but I will give you that overall it's important.

K5 is signifcantly better at base ISO and may be ISO200. And then it stops with about 1/5th off a stop. That is not close at all to your claim and apart from base DR it is never ever close to even 1/2 a stop.

OK, so I can stand corrected as the iso climbs on K-5. Still, for something where DR matters the most like landscapes, the K5 will do notably better. Keep in mind, that's a sensors *years* old and it still does better.

Only problem there is if you clip highlights the print won't save you. So will have to expose a bit more consciously on the Olympus and rely on more noise at high resolution to sample down to compare. But still, ok, fair point.

https://www.dxomark.com/Cameras/Compare/Side-by-side/Olympus-OM-D-E-M1-Mark-II-versus-Sony-A6500-versus-Olympus-PEN-F___1136_1127_1070

Better DR, better color sensitivity by about that.

BTW, this is the Sony A6500 sensor which is the base the Fuji sensor derives from.

True...but it is not the same. There is variation from camera to camera inspite of the same sensor due to hwo the signal is treated by the electronics.

Except several people around have noted the similar performance. It shouldn't be surprising actually. Main differences of Fuji vs Sony here are more due to the Xtrans CFA. Fuji trades a bit of color resolution, better B&W resolution.

All of this said, I can say that the PenF is the closest I have ever seen m43rds get on Fuji. I would still rank Fuji higher but I am very satisfied with the PenF raws & jpegs.

Yep it never was a big gap at all. PenF is not lacking in IQ at all.

It's still bizarre the color of PenF at higher ISO's is better than the EM1.2 in some ways. I would imagine some high ISO shots may show this a bit. But yeah, the Em1.2 sensor overall is at least better in some situations than the Pen-F's/GX8.

Not in some ways. In one way but probably in no way. COlour sensititivty...look at the chart..something went wrong here. The base ISO of PenF is ISO91. We can see a value here fro DR, noise and tonal range. But we see no value here for colour sensitivity.

Actually they explained that in their review.

"The PEN-F records the same 23.1-bit color depth at measured ISO 180 (manufacturer 400) as it does at its LOW ISO setting."

No they explain nothing here. Where is ISO91 gone? Out of the window?

I think the statement above they made is pretty straight forward.

If we take this literally the PenF has no colour range at ISO 91.

That would be a very faulty interpretation

Quite!

When we look at the GX8 sensor as a proxy we see its CS is well below EM1.2. Since the CS of PenF mysteriously is also better than all APS-c and CS is strongly related to DR (when you compare cams and look at DR you can predict the outcome in CS 9 out of 10 times).

But we are looking at the Pen-F, not the GX8.

Bith use the exact same IM279 sensor. Some difference is possible, a huge difference not. And this is a huge difference.

Well what I am considering here is what "magic sauce" Olympus could have done to keep the sensor that way. It's certainly a big difference. I know the GX8 also looks "a bit special" so I wouldn't rule out in the end they are similar, but the shot to shot I saw at higher ISO vs EM1.2 showed a clear advantage of color for PenF>

SO The measurement at ISO180 should be at ISO 91, the etc. You can do this in your head looking at the graphs and take GX8 in there too. You can easily see that now the PenF CS graph compares almost identical to the graphs fro DR...A little worse than GX8 at the lower ISO's, a little better at the higher ISO and overall worse than Em1.2 and almost all APS-c sensors.

The graphs are indeed similar to the GX8, *except* in color sensitivity. And this is a highlight that DXo mentions. From my own experience with the PenF, I can indeed say that I noticed something rather strange (for the better) at high ISO. The color sensitivity advantage is at high ISO of 1600 and up.

The few who actually own both PenF and Em1.2 note the latter being the better and they would actually be pleased if they could swap the sensors....May we can ask them how it deals with colours. Having said that: compared to my GH4 Em1.2 retains colours better at higher ISO and that has been noted in many reviews so it seems to be a trait shared by the 20 MP sensors.

I also saw a comparison through different ISO's of EM1.2 and Pen-F and, yes indeed, somehow the PenF was retained better color in the 3200+ (or was it maybe 5000+) up region. I was pretty surprised.

I wonder where. I know dpreview showed something like that with ACR beta. Otherwise most reviews conclude the best sensor is the one in Em1.2 and almost surely GH5. We'll see how that one willl score...

In the end I expected GX8/PenF performance from Em1.2. So actually I am more surprised that the Em1.2 scores so much better on all other parameters (well...MUCH only when it comes to noise).

I think this is what I saw

http://www.slrclub.com/bbs/vx2.php?id=olympus_e10_forum&category=1&no=536442

The original has more explanation. I was surprised the PenF kept color on the glitter over the EM1.2 loss of it.  Given DXo's color sensitivity result, where they measure it still better over EM1.2, it seems to go together.

But yeah, I am kind of a bit surprised by this result. Nevertheless I would certainly expect the EM1.2 from all I  have read to be 1/3rd to 1/5 stop better in some areas/situations than PenF.