Considering 4:3, but have some questions about IQ

Bhima78 wrote:

samtheman2014 wrote:

Bhima78 wrote:

Nirurin wrote:

So at the moment I have my toe dipped into the sony APSC ecosystem, with an a6000 and the kit lens + the 55-210 zoom. I've had a few issues with it though, and I have the opportunity to sell it all and exchange for another system.

Figured it would be better to do it now, before I buy any more lenses!

One issue I had with the sony system is that there are very few lenses, and a lot of the ones I like have compromises. For example, they have some fairly decent zooms, but they're all pretty slow in aperture. And the only macro is 30mm, though it's also very cheap.

I had been under the impression that micro 4:3 had more lenses at better prices, with the downside being the smaller sensor size. However today I spent some time on DXOMark to look through the 'best' lenses... Here are some of the scores -

Sony 55-210 = 13

Sony 18-200 = 13

Panasonic 14-140 = 9

Olympus 14-150 = 10

The sony 55210 and the panasonic are both very cheap, with the Olympus being the same price as the sony 18-200 (twice the price of the others...) and yet the sony lenses are leaps and bounds better image quality?

Though saying this, the 'kit' lenses for panasonic seem to get marginally higher marks compared to sony's, so maybe it evens out on average over the line...

This is turning into some kind of 4/3 bashing, which isn't my intention. I'm partly wondering if I am misinterpreting the results of these tests somehow. I really want the 4:3 lineup to be worth the hassle of changing systems haha.

Both systems are small, I guess I want to be able to get good image quality for my money.

Any tips or thoughts are welcome, sorry for the ramble, late night

DXO sharpness numbers do not, in any way work cross platform. The only way you can even come close to determining how one lens/camera combo could fare against the other is if both systems have the EXACT same lens that you can normalize the results to.

Luckily for you, you have an A6000. And there is a lens on that system that is the EXACT same lens on m43's: the Sigma 60mm f2.8. The lens is literally a carbon copy of itself for both systems, the difference is the mount alone. Now look at the DXO sharpness numbers for both systems and then see how the lenses for each system compare to the sharpness of the Sigma 60mm on each of their respective systems. You will find plenty of m43's lenses come close and beat the Sigma 60mm. On Sony's side, you won't find any that beat it, and very few that actually get close to the sharpness of that lens.

Here are a few real world examples as to why you cannot use DXO sharpness numbers across platforms.

Real world scenario 1: Canon 6D with Canon 17-40mm F4: Sharpness score: 14MP /// Olympus E-M1 with Olympus 12-40mm f2.8: Sharpness score: 9MP

Now, here are actual real world photos with those setups:

http://www.43rumors.com/full-frame-vs-micro-43-revisited-with-pro-olympus-lens-guest-post-by-chris-corradino/

That test is comically badly done , the fact that you need to post such poor examples as "proof" rather says it all

Not sure what is comical about it. The guy took a shot with a 6D, then for the fun of it, he decided to see how his new EM1 with the 12-40 would fare against his older setup. Took the photo and there are the results.

I will save us both some time , look at the previous times you or others have posted the joke test in this forum and read the responses nothing has changed

Now some real world corner performance at infinity (really important for landscape lenses): Nikon performance in the corners at f2.0 (scroll to the bottom of the link to see the Nikon 24mm f1.4 at f2.0): http://www.cameralabs.com/reviews/Nikon_Nikkor_AF-S_24mm_f1-8G_ED/sharpness.shtml

Panasonic 12mm f1.4 corner performance at near infinity:http://www.cameralabs.com/reviews/Panasonic_Leica_DG_Summilux_12mm_f1-4_H-X012/sharpness.shtml

A FF lens at F/2.0 is the same as shooting a 12mm F/1.0 lens on m43 , same AOV, same DOF and same total light gathered . Speaking as FF and m43 system user I would love to see an example where the very extreme corners of an 24mm @ F/2 on FF or 12mm @ F/1.0 is really important for landscape" do post one I would love to see it

Night sky astro photography focused to infinity with aperture wide open could be very useful to keep ISO's lower and not get star trails.

Then why don't you show me an actual photo where the extreme corner crop highlighted in the photo below matters . Though if it the extreme corners were that important , just shoot a tad wider and crop off the extremes you can afford to do that when you have a 36mp file . That is pretty much what is done with most m43 lenses including the so called PRO 7-14mm F2.8 and 12-40mm F2.8 which have huge corrections going on though I assume that the Pany 12mm will be optically corrected.

This is how much is cropped from the 7-14mm even at F5.6 { F11FF} !

This is some 100% edge detail showing how much distortion is corrected and how much is chopped off .

Just for a visual aid this is the area being considered on the 24mm at F1.8 , I darkened the rest of the image so that the " really important" part of the image could be easier seen Perhaps you could show me a m43 12mm at f/0.9 with better results.

I do not think that there is a 12mm F/0.9 lens available though if the extreme corner performance of the voightlander 10.5mm F/0.95 wide open are anything to go by I would not hold out much hope

Here, since you are continuing to play the "where is the 12mm f0.9 lens", I'll even move the goalposts for you so I can try and fit the facts to your narrative:

Yellow is APS-C crop corner //// Orange is FF crop corner

Even the FX corner at F4 is softer than the real world result at infinity of the Panasonic 12mm f1.4 AT f1.4:

PL12 at F1.4 and F2

No need to ask about an f0.9 lens when, even at APS-C crop focused near infinity, f1.8 on these lenses isn't very good.

Again you are comparing different subjects taken at different times at different exposures with different formats , the sampled area on the m43 shot is also further from the edge and of a far larger component in the picture .Now once again you seem adamant that on a landscape shot on FF at F/1.8 { m43 @ F/0.9} the extreme corners are vital so I am sure you can link to numerous samples showing this from both/either though bearing in mind the sampled area in the above D810 makes up roughly 1/500th of the overall image I really do not have high hopes for you.

Serguei Palto wrote:

samtheman2014 wrote:

Serguei Palto wrote:

samtheman2014 wrote:

Serguei Palto wrote:

The correctly measured properties of a lens can not depend on a camera body or the other equipment used with the lens.

A lens test in isolation is pretty useless , as far as I am aware we all use a lens mounted on a camera to take our photos . So the only test that makes any sense is indeed a "system" test. Lenses for m43 need to have double the resolution just to compete with a FF cameras assuming the same MP count , given that we are stuck at 16/20mp and FF has had 36mp for a number of years and more recently 42 and 50mp the comparative results are hardly surprising , well except to some members of the m43 fan club.

Invalid point of view.

If you know true properties of a lens then you know the best possible performance on any camera including that which will be produced in future.

The true properties of the lens are like a passport of a human. The data in the passport do not depend on a policemen who checks them, but provide a valuable information on a human.

With true data in the lens "passport" you can easily compare the optical quality of a lens independently on a camera platform it is produced for. One can also easily predict the best performance of an arbitrary system (camera+lens) with an arbitrary number of Mps on a sensor.

There is a concept of the ideal lens (which, of course, is independent on a system platform), and the true optical quality of a lens must be estimated on the basis of how large is the difference between the ideal and a real lens.

Contrary, if you measured just a performance of a particular system (body+lens) you can tell nothing on the true lens performance. As soon as a new camera bodies will be produced, all the data you measured for a particular body-lens system become just a garbage. A bright example are the data measured on the camera with the shutter shock.

The power of the concept of the interchangeable lens cameras is that lenses you bought can be used in future cameras. It is why only the true lens properties are valuable.

Again I take photos by mounting lens on camera as I suspect do you , DXO updates all it's lens tests to new bodies in each format{ eventually } so you can see how a given lens will work on your respective camera. Photography is camera + lens

I am not interested in DXO updates after the lens is bought. I have to know how a lens will perform on all the cameras (even on those which will be produced in future) before I bought it. And because the correct testing can be done, why not to do the job properly?

The test is done properly and the results posted relevant to how the camera +lens  combination  will be used, Other than increases in MP or tinkering with the AA filter  what magic future camera do you expect to radically change. Have we now established that you do indeed use your lens attached to a camera ?

For me "DXO updates.." means that DXO tryes to correct previous errors.

By DXO updates the cameras I specifically mean that when a camera comes out DXO will post the results of how the lens will perform on that camera

ahaslett wrote:

You can sometimes get the EM1 and 12-40 as a kit. Other than the Panny 12-32, I wouldn't buy a real "kit" lens.

Can't advise on Panny bodies, since I need PDAF.

The 12-40mm works very well with both the GX7 and GX8. I chose it over the 12-35mm for the extra reach, close-focus ability and the snap-back manual focus ring. It was a way to simplify my travel kit (fewer primes and lens changes) and not compromise too much. It's an excellent, very flexible lens.

To the OP… As others have mentioned, it's best to consider lenses before a camera body.

We in the M4/3 universe are lucky to have such a wide choice of lenses and body sizes, abilities and styles. Add in the ability to adapt lenses from nearly every other camera system and it can be more than a little mind-boggling… and cash vacuuming.

Nirurin wrote:

To the OP… As others have mentioned, it's best to consider lenses before a camera body.

We in the M4/3 universe are lucky to have such a wide choice of lenses and body sizes, abilities and styles. Add in the ability to adapt lenses from nearly every other camera system and it can be more than a little mind-boggling… and cash vacuuming.

I'm actually liking the look of the panasonic 14-140 ii. Seems to be regarded as a very good lens (for a zoom anyway), it's not expensive, and it's smaller and lighter than my current 55-210 (and as far as I can see, better image quality). So I can have one lens that does better than the two lenses I use right now, and have space in my bag for a dedicated macro or a telephoto.

I'll also have the 12-32 kit lens still, if I get the GX80, as it will only cost me £50 and gives me the option of a pocket camera.

I think these will do me pretty well to start with.

(The views on the quality of the 14-140ii I'll have to test for myself, as I'm not sure how accurate they are. Certainly people who own it seem to think it's very good, where most owners of the 55-210 just think it's "good for the price". DXO gives them both the same rating when put onto roughly equivalent cameras, but considering how DXO works that could mean anything lol)

Yes, the 14-140mm II is very good indeed. In good light it stacks up pretty well against the 12-40mm. My only suggestion would be to add a fast prime, like the 15mm or 20mm, for low light.

To take it a step further, a 7-14mm f4, coupled with the 14-140mm II + fast prime, gives you tremendous range and excellent image quality in a compact kit.

My first M/43 set was the GH2, 7-14mm, 14-140mm I and 20mm. That kit served me very well. For nature shooting, I added a 100-300mmand an adapted Nikon 60mm f2.8D macro.

Nirurin wrote:

To the OP… As others have mentioned, it's best to consider lenses before a camera body.

We in the M4/3 universe are lucky to have such a wide choice of lenses and body sizes, abilities and styles. Add in the ability to adapt lenses from nearly every other camera system and it can be more than a little mind-boggling… and cash vacuuming.

I'm actually liking the look of the panasonic 14-140 ii. Seems to be regarded as a very good lens (for a zoom anyway), it's not expensive, and it's smaller and lighter than my current 55-210 (and as far as I can see, better image quality). So I can have one lens that does better than the two lenses I use right now, and have space in my bag for a dedicated macro or a telephoto.

I'll also have the 12-32 kit lens still, if I get the GX80, as it will only cost me £50 and gives me the option of a pocket camera.

I think these will do me pretty well to start with.

(The views on the quality of the 14-140ii I'll have to test for myself, as I'm not sure how accurate they are. Certainly people who own it seem to think it's very good, where most owners of the 55-210 just think it's "good for the price". DXO gives them both the same rating when put onto roughly equivalent cameras, but considering how DXO works that could mean anything lol)

I can't imagine buying into m43 or Sony APS-C for the slower zooms when an LX100, RX100, FZ1000, or any of the RX10's will beat everything pound for pound.

samtheman2014 wrote:

Again you are comparing different subjects taken at different times at different exposures with different formats , the sampled area on the m43 shot is also further from the edge and of a far larger component in the picture .Now once again you seem adamant that on a landscape shot on FF at F/1.8 { m43 @ F/0.9} the extreme corners are vital so I am sure you can link to numerous samples showing this from both/either though bearing in mind the sampled area in the above D810 makes up roughly 1/500th of the overall image I really do not have high hopes for you.

The different subjects at different times thing is valid in terms of judging overall IQ... it is absolutely NOT necessary when determining sharpness in a lens. The m43's area is MUCH more to the extreme of the corner than the APS-C crop that I have also put in my example, surprised you didn't notice this.

In terms of shots that could be useful wide open at 24mm: low-light landscape photography, especially landscapes that include stars. You can reduce your ISO significantly if your lens wide open can produce good sharpness in the corners while making sure you don't have star trails (if you don't want them). Yes, you could photostack/HDR/etc., but that is less flexible than just getting the shot because you are there, and there happens to be a nice shot (and you may not always be carrying a tripod with you).

Other shots could be larger groups of people in low light, or even environmental portraits in general. m43's double DoF advantage combined with solid corner sharpness  wide open means you are again, reducing significantly the ISO to get the shot.

Again you move the goalposts... basically stating that, 100% crops on a D810 are also MUCH more magnified. I understand this, but the point is: If you aren't taking advantage of those 36MP because your shot is slightly blurry, and you are still shooting at higher F-stops and higher ISOs, what advantage are you gaining here? The lenses at 100% just are not as sharp, and require quite a significant amount of stopping down before you take advantage of the 36MP sensor.

All camera systems have compromises. If you want extreme sharpness for your FF camera, you either stop down like a sonuva, or you buy really good, expensive glass.

FingerPainter wrote:

Astrotripper wrote:

FingerPainter wrote:

The Sony 18-200m on a6000 is sharper, faster and less expensive than M.Zuiko 14-150 II on E-M1.

The difference in marginal. And the Olympus on the other hand, is weather sealed. And Panasonic 14-140 is better. No matter how you slice it, zooms on a6000 don't look good compared to Micro 4/3.

Panasonic 7-14 > Sony E 10-18

E-M5II + 7-14 f/2.8 PRO: $2,100. GX8 + Pany 7-14 f/4: $1,900. a6000 + 10-18 f/4: $$1,300

Sure, but for that extra money you get:

• weather-sealing (not with Panny 7-14, though)
• IBIS
• solid magnesium alloy body
• better EVF
• higher flash-sync
• higher max shutter speed
• electronic shutter mode
• Mic input
• 4K video (Panasonic)
• High res mode (Olympus)

Olympus 12-40/2.8 > Sony E 16-70/4

E-M1 + Olympus 12-40 f/2.8 PRO: $1,800. GX8 + 12-35mm f/2.8: $2,000. a6000 + Sony 16-70: $1,550.

Sure, but E-M10 II + 12-40 is ~$1,450 and a6300 + 16-70 is $2,000. The former is more of a competitor to a6000, and the latter is more of a competition to GX8 and E-M1. Although a6300 is probably more of a competitor to GH4, as both are heavily video centric.

Olympus 14-42 kit lens = Sony 16-70 kit lens

E-M10II + 14-42: $700. G7+14-42: $600. a6000 + 16-50: $700.

And GX80/85 + 12-32 is $800. And this is the comparison that makes sense, as all those cameras are roughly in the same league.

The above should not be true, considering the difference in sensor size and resolution.

Most camera makers have lenses at different quality levels. This is true for Oly, Pany and Sony. You are comparing top quality Oly and Pany lenses to less-than-top quality Sony lenses.

The problem is that apart from some FE lenses, there appears to be nothing other than "less-than-top" quality from Sony. Sigma showed us what kind of quality can be achieved on a6000. And at a reasonable price at that. None of the Sony lenses tested by DxO approach that level of quality (still talking in terms of DxO scores). I find this weird.

As you can see, from the price comparisons for the lenses you cited, to get higher sharpness on an m43 body, you need to pay a higher price

Sure, that's because you made sure to choose more expensive bodies The lens prices by themselves don't vary that greatly

, and whe you get similar quality you pay a simialr price. I didn't bother to check weights, but I wouldn't be surprised to see that the better m43 lenses weighed more than the poorer Sony lenses.

I'm sure they are, especially Olympus ones. They go all out with their PRO line and don't seem to care about things like that. Different design priorities, I guess. It's one of the reasons I'm not really interested in them.

Where is this 5-6 P-Mpix advantage that a6000 should have over 16mp MFT camera? It's nowhere to be found.

You shouldn't expect to see consumer grade lenses perform as well as "PRO" lenses.

Sure, but Sigma 60/2.8 is not exactly a "pro" lens. I'm not familiar with Sony's marketing, can any of their APS-C designed lenses be considered "pro"? To be honest, I would hesitate to use this term at all when talking about those systems, prosumer maybe fits better (when I write PRO in relation to Olympus lenses, I only mean their official designation, as means to distinguish that line of products).

The real issue here is that Sony has fewer top-grade lenses for the wider focal lengths on APS-C.

The scores are mostly within 1 P-Mpix of each other, which I don't think is enough to declare one a definitive winner.

If they score about the same, but the Sony is several hundred dollars less expensive, isn't the Sony a value winner?

No, because for more money, you get added value in other areas, as pointed out above. If you choose the top Sony body, you will not pay less.

However, I d agree that most of the earlier Sony E-mount zooms were underwhelming

And the problem is that it's almost all there is. Sony stopped making lenses for APS-C E-mount almost three years ago. All you can hope for now is more (and better) FE lenses.

You see almost the same thing in Canikon. Mearly all of their best lenses are made to fit the FF mount.

Yes, that's why I think Fuji is the only serious APS-C system on the market. Everything else is just an afterthought.

However, this is not a problem, in that you can use these lenses on the APS-C bodies if yo want to.

Sure. The problem here is the prices for FE lenses. Most are really expensive. Looking at B&H, there are only 3 FE lenses below $700 mark. And that includes the kit lens. So yeah, you can do that, but that's a weird argument for someone who just tried to make price the major advantage of going that route.

Who said the purpose of Sony APS-C system was "a compact system that delivers high IQ". I don't think that was ever its purpose. It was a smaller, lighter system than the (dSLR) competition.

Isn't that the same? DSLR like performance in smaller, maybe even cheaper package? Sure, you get smaller kit, you get DSLR level performance, but you don't get DSLR level image quality (again, still talking DxO).

Why not just go for A7? The first generation was not much bigger than a6000. And now you will truly have an advantage in image quality. And there are some lenses that make it a very nice, compact kit, even by APS-C standards.

Price?

Sony's APS-C seems like an unnecessary compromise that you take because you can't afford to go for FE.

Perhaps it is. Or perhaps it is a system that is small and inexpensive.

Micro 4/3 can be even smaller and less expensive, so...

It makes up for it by being cheap and offering nice feature set and good AF system. But it doesn't offer the IQ advantage that one would expect to get. That's my whole argument.

Your expectations seem to be based on sensor difference without reference to size and price.

Because that's the main argument against MFT. That bigger, higher res sensor automatically means major advantage over smaller MFT cameras. Because equivalence. If DxO is any indication of reality, than that is not the case here.

And sorry, but robust AF does not make a6000 a competitor for cameras like E-M1, GH4 or GX8. Not the same league.

If they are not in the same league, why would expect the lower league camera to provide better IQ?

You know very well that it's not about IQ. Older, entry level E-PL5 delivers the same image quality as the newest E-M5 II or E-M1, why didn't you use it in your price comparisons? Do they play in the same league? Is Nikon D3300 in the same league as D7200? IQ is almost identical after all.

Nirurin wrote:

Cimarron wrote:

TN Args wrote:

Nirurin wrote:

...today I spent some time on DXOMark to look through the 'best' lenses... Here are some of the scores -

Sony 55-210 = 13

Sony 18-200 = 13

Panasonic 14-140 = 9

Olympus 14-150 = 10

The sony 55210 and the panasonic are both very cheap, with the Olympus being the same price as the sony 18-200 (twice the price of the others...) and yet the sony lenses are leaps and bounds better image quality?

You will find the DxO lens scores go up and down when you select a body with more or less MP. Try it. The Panasonic in your list as a 9, scores 13 on an Olympus E-PL5 body. So it is not the lenses that are being scored alone. So you can only really compare lens vs lens on the same body.

Unfortunately DxO have still not got any lenses tested on the new 20MP micro four thirds bodies, which would up their scores still further.

Personally I ignore DxO like the plague for a number of reasons, and I advise you do the same unless you want to get very technical.

Agreed. DxO lens ratings are totally bogus because they rely in large part on the camera the lenses are tested on. They are best ignored. Sites like Lenstip are much more reliable.

However, DxO sharpness ratings can be useful if you know how to interpret them. Their "virtual megapixel" sharpness ratings, or whatever they call them, represent the efficiency of a lens in delivering the resolution of the sensor. That's why most APS-C and full-frame lenses have higher ratings: They're being tested on cameras with more megapixels.

For example, that Sony 55-210 that has a 13 Mpx sharpness rating is only delivering 54% of the Sony sensor's 24 megapixels. But the Olympus 14-150, with a 10 Mpx rating, is delivering 62.5% of the M43 camera's megapixels of resolution.

One of the highest rated M43 lenses, the Oly 75mm f/1.8, gets a 13 sharpness rating at DxO, which means it's delivering a whopping 81.25% of the M43 camera's megapixels. For the Sony 55-210 to have the same 13 rating is patently absurd.

I've owned both. Everyone agrees that the Oly 75 is one of the sharpest lenses ever made, in any format, and the Sony 55-210 was blown out of the water by the lowly Oly 40-150 f/4-5.6 zoom.

I've had both systems and I ended up selling my a6000 and lenses because my Oly E-M5 and lenses just made better photos. And I did LOTS of side-by-side comparisons before coming to that decision. I actually wanted the Sony to win, because I loved the AF system and great sensor, but the lens selection was the deciding factor.

Have you had any experience with the EM10ii at all? do you have any advice on deciding between the EM10ii, and the GX85?

I have to admit to liking the looks of the Olympus bodies, and the IBIS is tempting, but I do lose out on a few features like the 4k post-focus (though without using it extensively I won't know if that ends up being a gimmick, and that normal focus bracketting would be better for me for macro etc)

No, sorry, I have only an Oly e-m5 and Panny GX8. Of the two you mentioned, I'd personally choose the GX85 because of its 5-axis IBIS, DFD focusing and 4k photo mode. And I prefer the rangefinder body style.

Dxomarks seems to be a good comparison within the same types of cameras, but not between different systems.

For instance, the Olympus 75 1.8 is a high end lens that is better than just about any lens out there for any camera, but if you compare its DXOmarkd score to a good Nikon or Canon lens it would score half.

If you take that same lens and go to another site, it is at the top of the charts in the 80s and the Canons and Nikons are in the 40s & 60s

If you look at image-resource, lenstip, photo.de, ephotozine,etc... You will see the m43 lenses are comparable or better than the big names for larger cameras.

Most people don't print huge wall size prints, and if you were to print a 3 meter/yard x 4 meter/yard image, you would probably see a considerable difference between a 16mp & 24mp image.

For 8x10s or 11x14 or computer screens, they will be identical because you are only using about 8mp.

Arizona Sunset wrote:

I can't imagine buying into m43 or Sony APS-C for the slower zooms when an LX100, RX100, FZ1000, or any of the RX10's will beat everything pound for pound.

Beating everything pound by pound - maybe,

beating everything buck by buck - maybe,

beating everything in terms of IQ - not that sure