Panasonic Lumix DMC-GX7 First Impressions Review

none of these shots were designed for comparison of image qualities, original or "print" mode, because DPReview don't have a definition of image quality when they designed these tests.

use of camera advertised ISO makes the comparison impossible (because it means different things on different cameras) and use of picture height makes it impossible to compare directly across different aspect ratios.

these are fundamental flaws though I agree that accurate exposure control means bloody effort and may be turned down for cost. the correction of PH can be done easily.

Sep 3, 2013*permalink

zodiacfml

I'm so excited to see the new studio tool despite using gprs in a small island as intrnet connection. Anyway, I love the new Print size option and the flat DOF. It was a revelation. Like, in terms of noise and high ISO, the modern cameras came close to each other. Yet, in detail and noise, FF is still significantly better even at print size resolution. Can't wait to see the Nikon's, Ricoh GR, and RX1R .

One more thing, DPR please include a Sigma Merrill camera and other print size options such as 15MP and 20MP.

Aug 31, 2013*permalink

Red G8R

For me the most impressive thing here is the Sony RX100 II. It stands up well against the others despite the smallest sensor.

Aug 31, 2013permalink

tjbates

Anyone know the specifics of silent mode (beyond what's already been written above) Specifically, I'd lke to know if the silent mode produces higher resolutioin images than the silent mode on my GH2. The GH2 can shoot 4MP in 4:3 ratio but only in the SH burst mode - 2-40 shots per burst. It is truely silent.

Aug 31, 2013permalink

rsongusa

I, too, would like to know the specifics of the silent mode. Can I shoot RAW in silent mode?

Sep 7, 2013permalink

Demon Cleaner

The electronic shutter shoots full resolution RAW images. The only limitation when compared to the mechanical shutter is that it can't be used above ISO3200 and for exposures of longer than 1 second.

Sep 13, 2013permalink

Julio Tegner

The test shots show it comparable to the Pen E-P5 and NEX-6.

But why did they make a $1000 camera body with no audio in? That alone makes it a loser unless you have no intention of shooting video with decent sound.

Aug 27, 2013permalink

John Driggers

Because they make a different camera optimised for video-buy that one if video is your primary or most important concern. If stills is your most important concern, this is your camera. That's why they made a $1000 camera body without audio in and and that's why I bought it, 'cause I give f-all about video. If I did, I'd buy the GH3.

Oct 16, 2013permalink

JJ Rodin

Where are the new "higher resolution" test sites that old studio scene had ? The old lady with the small vertical skiggly lines (needed 16mp or beyond to resolve) & details of face, the male head, the fine detail in feathers, etc ?

The skin tones tests are good addition, but need some/MORE resolution test sites!!!!!!

No specific Dynamic range areas as well !!

Needs Improving quite a bit IMO !!!

Aug 26, 2013permalink

Combatmedic870

Man....the GX7 is completely on par with the NEX6 at all iso's and a TAD bit ahead of the EP-5 at 6400 and 12800.

Thats pretty dam impressive!

Aug 26, 2013permalink

basically the analog ISO of 4/3" won't go beyond 800 or 1600. higher ISOs are numerically amplified and noise reduction may be applied differently.

Aug 26, 2013permalink

Andy Turner

Do we have any idea on when the full review is likely to appear on DPReview? Thanks

Aug 23, 2013permalink

Misa

To me it looks much more like a direct translation drom Lumix L1 to Micro Four/Third and I think is very good, I wish I could buy one right now :)

Aug 22, 2013permalink

looks the designer of L1 is not fired.

Aug 22, 2013permalink

lxcellent

This is an ugly camera. I know that you should not judge a book by its cover, but why not make it look more like the classic rangefinder like they did in the LX1. Shoot. Follow Fuji's lead on this.

...and yes, the way a tool looks DOES make a difference!

Aug 20, 2013permalink

JeffreyG

Have you seen the camera yet? Personally, I think it looks very similar to the Fujifilm X series cameras.

Aug 20, 2013permalink

> Fujifilm X series cameras.

which are ugly.

Aug 21, 2013permalink

well that's totally subjective, personally I think this is the best looking camera out there

Aug 21, 2013permalink

there are (subconscious) objective reasons behind subjective judgements, like what we call sexy has much to do with reproductive capability even you always take contraceptives.

those who think retro camera designs are good have their reasonings rooted in misknowledge/misunderstanding.

Aug 22, 2013*permalink

um, so liking which camera will boost my reproductive capability the most?

Aug 22, 2013permalink

straylightrun

yabokkie, it's time put down those evolutionary psychology books...

Aug 22, 2013permalink

xMichaelx

What happened to DPreview's "Block user" functionality?

There's little point in coming here if trolls (lxcellent, in this case) can't be blocked.

Aug 23, 2013permalink

sadwitch

I agree there should be a function to block messages. If not there's always the tedious way of starving them for attention.

Aug 25, 2013permalink

tomservo33

Looks great to me, except the eyecup bulge....but if it feels anything like the GX1, which is by far the most comfortable mirrorless I ever held, it will be a great shooting camera and perfect for those us with large hands who can't safely grip many of OLY's offerings while climbing up a small cliff to grab a shot of some birds nest or baby mammals!

Aug 29, 2013permalink

while alot of people probably love the EVF, I don't like the eyecup bulge either, I'm hoping it can be detached.

Sep 1, 2013permalink

LarryK

I like an "Ugly" camera, doesn't draw attention. Besides, I look though it, not at it. Others will have to suffer.

Sep 30, 2013permalink

pgphoto_ca

For me, the E-P5 has a slight better ISO over the GX7 at 800 ISO. approx 1/2 stop diff.

Image quality is prime for me. I just hope Panasonic will improve a bit the ISO before they release the GX7...

go go go Pana !

Aug 20, 2013permalink

OngNikon

And you had detected the difference. And it is going to be game changer for you!!

Aug 20, 2013permalink

tomservo33

From the test images on Imaging Resource, The P5 seems to have a little better resolution and sharpness, at low iso, or perhaps less anti-aliasing or NR than the GX7, side by side (but that could be lens based or some setting. I was looking at eyelashes on the mannequin, so not sure if that is relevant, since the leaves on the trees on outdoor house shots look identical, perhaps a lighting response?

Aug 29, 2013permalink

Hugo808

Where's the robot? How can I choose a new camera without it's little metal smile?

Aug 19, 2013permalink

Higuel

I do miss the drawing of the old lady drawing!!!!!! SO MUCH easier to see the definition and detail between different cameras or evaporating with increasing ISO!

Miss also the box with the coloured stuff inside!!!
Those should stay for sure!!!

One very welcome thing is the colourful circle target!!!! :D

Aug 19, 2013permalink

M Jesper

People seem to really care about 'their' test chart. :)

Aug 18, 2013permalink

when DPReview don't?

Aug 18, 2013permalink

KZMike

For me it is more about having more cameras [aka reviews] that I'm to use for comparative purposes. . . How does one compare the GX7 with the Oly OM-D EM 5? This change virtually means starting from square one and has little meaning until the 'bank' of tests has enough history for comparative purposes. It also renders the previous bank of test useless to use for comparison with the new cameras with the new studio comparison.

Some things are best left along, especially if it is 'not broken'

Aug 20, 2013permalink

Timmbits

@kzmike:
the EP5 is the same sensor - it is right there in the choices

Aug 23, 2013*permalink

ThePartaker

What strikes me about these tests is how well the RX100 II holds up as it only has a 1" sensor and is truly pocket-able. It looks just about usable up to 1600 - will you be doing a full review of it soon? What do others think?

Aug 18, 2013permalink

The RX100 II's high ISO shots do hold up very well. Hopefully the low light autofocus speed and accuracy are also good.

Aug 21, 2013*permalink

unfortunately, accd'g to what I've read, low light AF is still not great on the RX100 II

Sep 1, 2013permalink

rxbot

Pocket lint has a review with raw and jpeg samples. Their conclusion is E-P-5 wins on IQ but G7 wins on package because the Pen5 does not have the built in EVF so adding a finder adds a lot to the package cost.Auto focus speed both cameras are blazing fast. Personally I will wait till the next versions of XE-1 and Nex6 come out as I like 16MP APS-C sensors.

Aug 18, 2013*permalink

Pocket lint has nothing to substantiate their claims. IT is only to their eyes it is so, if I am correct. So it doesn't mean too much to me. Others, with their eyes have said the opposite. Those who did noise test etc in general say that they performequally with a slight advantage to the GX7 at higher ISO (really high ISO that is).

Also we need to see how well Dynamic range etc performs which we cannot assess here. Maay be EP5 etc indeed have clearly better Dynamic range and is the better sensor.

The best Panasonic sensor ever when it comes to DR, the GH1 scores 11,6 eV. G6 scored 11,5 btw. So I think a score of 12 to 12,5 is what we can expect. With the changes they made they should be able to get close to that.

GH1 scored 772 Low Light ISO. 895 ISO was the score of EP5. If this is the best Pana sensor ever, it should be possible to come close to that.

But okey...I'll await DxO.

Aug 18, 2013permalink

Pentax use Sony sensors to "out perform" Sony cameras by a good deal. obviously Oly failed to do the same.

Aug 18, 2013*permalink

rallyfan

AF speed comment is very interesting!

Aug 19, 2013permalink

though 70D is only the first budget DSLR with an old DIGIC processor, the dual-pixel AF is at least better than A99 (impression).

Aug 21, 2013*permalink

From the DPR photos, the high iso is very close, the ep5 may have the edge. Good to hear that the autofocus is fast. Hopefully it works as well in low light and with 4.3 fps tracking as claimed.

Aug 21, 2013permalink

Robert Morris

Yes, I think their reviews are worth about what the name suggest "Pocket Lint" and not one furball more.

Sep 16, 2013permalink

Roger Nordin

From what I have been reading the GX7 sensor gives lower high ISO noise at the expense of earlier highlight clipping vs the E-M5. Seems to me the sensors are very close, just that Panasonic vs Olympus used different approaches to how to tune the sensor. Doesn't surprise me either, my GF3 and GX1 has very abrupt highlight clipping and limited highlight recovery compared to my OM-D E-M5, which has just amazing highlight recovery and resistance to clipping. I am so amazed I sold my Canon APS-C stuff, I just didn't see a point sticking with it at all.

Oct 25, 2013permalink

Octane

One slight problem I see is that all of these elements are made from synthetic dies and materials. While they are colorful, they are no what nature offers in terms of color. Yet the majority of subjects we photograph are nature and of course people (natural skin). A printed photo of a person is a very reduced representation of what real human skin is like in terms of colors, shades, how light gets broken, absorbed and reflected to create it's unique look.

Of course I'm aware it's impossible to have a real person be part of a test scene that has to be 100% constant over years. OTOH I know from owning about 20 different digital cameras over the years that some cameras produce wonderful looking photos from artificial colored scenes, they do poorly when it comes to resolving very fine differences in colors and shades which is key for natural looking scenes, especially when it comes to portraits (and landscapes).

Again, I don't really have a solution, but something to think about.

Aug 17, 2013permalink

lester11

The Paul Smith watch told me everything I wanted to know (for my particular taste in snaps). How the letters were sharpened (or not), how the dial ticks were resolved (or not), how the particular glittery blue of the face was smeared (or not), how the second hand suffered aliasing (or not), how the glass reflections affected the render. Every time I saw a better rendering, I bought the camera and dumped the previous (smile) or lived with the difference 'cos it was smaller than the bank balance... Can you not tuck it into a blank space somewhere, please?

Aug 17, 2013permalink

ET2

No, because this scene is so much larger than the old one, the watch would be tiny. It won't look anything like the last scene. Just look at the Queen of hearts in previous scene and compare how much smaller the cards look in this scene.

Aug 18, 2013permalink

Tan68

It is a bit more like a resolution chart... Will any deviation of resolution in the corners create distracting arguments.. if these differences might be due to different lenses...? I know the same lens in different mount is used but they are still physically different lenses and may have some different characteristics...

The different lighting levels is a good idea and I know there are dark swatches, but I might miss the little box of shaded puffy balls and/or thread. [well, the parsley in the corners does do just as well so the friendly looking puffs are gone..]

Resizing the images to match at ~8MP is a good addition. It should help some people understand if the 24MP APS-C sensors really are noisier than 16MP...

This change offers more than the last change. Still, I hate to see the robot and batteries go. I remember him from my first digital camera purchase... Maybe add a picture of the robot somewhere. Just for a little fun and to retain touch with the old :^)

Aug 17, 2013*permalink

it's more than a simple sharp black-white resolution target, for which software can make good guess and generate artificial image that may match the target well.

Aug 17, 2013permalink

I have compared the exposes of GX7, NEX6 and EP5. In lowlightmode there is not difference. bar two small exceptions out of 21 (7 shots per cam from ISO 100 to 25600 times 3) they all use the exact same shutterspeed.

In Daylight mode somehting is different. From ISO 200 to ISO 3200 the NEX uses 1/3 less shutterspeed, so 1/3 of astop. compared to the others.
ISO6400 Oly falls behind it seems. It is 1/2 a stop slower. Shutterspeed is 0,0004 s instead of 0,0003 on the other two.

The strange thing is that ISO12800 and ISO25600, the EP5 uses the exact same shutterspeed. Where GX7 goes down to half the shutterspeed, as can be expected when you move up 1 stop. To makeit clear, the Oly file at ISo 12800 has 2,5 times longer exposure than the GX7. The Sony stick to ISO 0,00025 s which is logical. It is 1/4000s its max shutterspeed. when we go up to ISO 25600 both Panny and Oly use the same shutterspeed. Strange but it means that Oly again uses 2,5 times more time GX7.

Explanation?

Aug 17, 2013permalink

you may have assumed that all were shot at exactly the same lighting which may not be true.

actually in a well controlled test, lighting should be adjusted carefully to compensate slight differences among shutter speeds and aperture sizes (less than 1/3 stops of normal control).

Aug 17, 2013permalink

I surely do assume that all things are kept equal. If you want to compare, you have to. If you want to show how good an individual cam can perform, you should not.

I am convinced dpreview ekpt everything the same.

Aug 17, 2013permalink

I only want the exposure on the sensor be the same.

for GX7 vs NEX-6,
many say their image qualities look the same.
let's assume everything the same except shutter speed, then
NEX-6 got 0.32 stops underexposed and 0.17 stops smaller area (11.1%) because a flaw in the test.

so NEX-6 is 0.49 stops better than GX7, while the area ratio suggests that NEX-6 gets 0.70 stops more light.
then per unit area (mm2) performance of GX7 is better.
(assume both of them cook raw data the same, too)

Aug 17, 2013*permalink

so if Pana had made an APS-C sensor for Sony,
it would have beaten NEX-6 by 0.2 stops.

Aug 17, 2013permalink

? I think we'll need to await throrough testing to say anythign about total IQ.

Aug 17, 2013permalink

Steen Bay

@Jorginho - Did you also check the f-stop? DPR usually uses f/6.3 on mFT and f/8 on APS-C. If that's the case here too, then Nex-6 gets 2/3 stop less exposure if the shutter speed and lighting is the same.

Aug 17, 2013permalink

http://www.dpreview.com/forums/post/51998946

Yes of course. Look here to see the result in a table I made. All were at f5.6.

Aug 17, 2013permalink

@Yabokkie....What you seem to forget is that the NEX6 files also are smaller than the GX7 EP5 files. Everything is smaller in them by a clearly visible amount. This means that noise will appears smaller and finer too. When you make them th same size, the noise of the NEX6 will becoem a bit more visible..

Aug 18, 2013permalink

there may be other reasons.

one is although they are both 16MP sensors, DPReview uses a smaller portion of a 3:2 sensor for the main part of the scene within a 4:3 frame. what left over (left and right of the 3:2 sensor) are mostly empty with some simple figures.

this is a fatal flaw that DPReview are determined to have.

btw, I'm not saying the two are the same. I only wanted to say how we could interpret and what we might get from these tests, and one of the considerations is exposure variation as you pointed out.

Aug 18, 2013*permalink

also from the sensor point of view, we better test at the same shutter speeds, though a 1/3 stop difference won't make visible difference.

Aug 21, 2013permalink

Joemuma

We are always looking at the shadowy box with thread spools. It shows the low light performance of the sensors and makes it easily comparable as the noise reduction of the cameras are removing the details. You yourself used this for the "Image quality test" in the A99 review. - Would miss this box! (you too?)

Aug 17, 2013permalink

The problem is that to evenly-light a large, flat target like this, the light has to be fairly straight-on, so the box we built didn't end up with much shadow.

Aug 19, 2013permalink

scotbot

Excellent, especially for lens reviews.

Aug 17, 2013permalink

dscottsatx

This new studio test would be a lot better if it wasn't so shallow.

Aug 16, 2013permalink

Stu 5

The fact it is so shallow is what makes it so good. Testing is meant to be done on shallow test charts in the first place.

Aug 17, 2013permalink

> is what makes it so good

the depth of field is not so bad an issue.
actually it may well be welcome if seen from another angle.

my understanding is that we need a flat target because of
f-number and focusing problems.

in ideal, if f/5.6 is used for 4/3", f/7.1 should be used for APS-C and f/11 for 35mm format but that's not what we get. then we give up and look for a flat target.

Aug 18, 2013*permalink

fcna72

How can I download full size JPG of the scene other than ISO 125? Whatever the ISO setting selected (and visiable), the downloaded file is the ISO125 P1030034. Is this intended (and what am I doing wrong) or a bug?

Aug 16, 2013permalink

That would be a bug - we'll look into it.

Aug 17, 2013permalink

fcna72

Thanks Richard

Aug 17, 2013permalink

fcna72

PS: issue seems to be related to firefox browser. Working fine with IE

Aug 17, 2013permalink

a fundamental bug the tool has, or the test procedure has, is that it fits the scene to image height which make it impossible to compare sensors of different aspect ratios.

Aug 17, 2013*permalink

108

I must say this rx100 second is impressive. Not bigger than a s95, would fit in the belt bag and ever ready . Photography at its simplest.

Aug 16, 2013*permalink

108

In the low light comparison with gx7, ep5, nex6 and rx100II, all in auto wb except not specified for the rx, , ep5 jpegs come constantly much too warm/yellowish and too saturated. Main reason I don't go over to Pana is the Oly jpeg engine/colors etc....so much for it. I have this kind of problem with E620, but to that extent..and I don't shoot raw, and don't like Lightroom

Aug 16, 2013permalink

peevee1

Probably "keep warm color" is ON.

Aug 17, 2013permalink

108

thanks

Aug 18, 2013permalink

Dimitris Servis

I think you miss an area with shadows that can show the resolution and dynamic range like the box of colorful fluff and threads.

Also, I really wonder, do you shoot at the same aperture? What distance? How do you make sure everything is in the dof.

Aug 16, 2013permalink

Dimitris Servis

Where's the 1000 drachma bill? Grexit has happened and you put it back in circulation?

Aug 16, 2013permalink

Imagefoundry

some feedback:

1. incandescent lighting needs to get a lot more even
2. little gear icons (showing camera settings) are only present for 2 leftmost magnified samples. Two samples on the right don't show their settings...
3. I wish there were a few bright, colorful objects in primary colors (RGB/CMY, but not flat) to show effects of channel clipping

finally, everything is shot with the maximum aperture?? makes no sense

Aug 16, 2013permalink

HelloToe

1. The whole point of the low-light shot is that it's NOT even. Some of us take pictures outside of perfectly lit studios, you know. That's why objects are repeated, so you can compare how they look in direct light vs. shadow. If anything, I'd say it needs to be darker, and more from the side so that objects cast shadows on each other.

2. For me, the gear icon is only missing for the RX100. They said they're planning to add more settings info in the future, though.

3. Check the paint tubes? There's definitely a real shortage of 'color with details/texture', though. You've got those nice patches of greenery, but for any other colors, all you've got to go on for texture & details is the thread.

Aug 16, 2013*permalink

goblin

Older test scene was good enough, and most of all - had all the old non CaNikon cameras which you guys will never reshoot on the new one. This new scene is a break.

Couldn't the new light simulations be added to the existing scene ? And where are my bottles !!! :D

Aug 16, 2013permalink

The old scene was too small - we had to shoot too close up, which ended up mis-representing cameras such as the RX100. Its 3D nature also meant we had to stop the lenses down beyond their sharpest point - the new chart fixes both of these problems.

There are plenty of non-Canon or Nikon products in the re-shoot list.

Aug 16, 2013permalink

goblin

Thanks

Aug 16, 2013permalink

kutya1977

Thanks! Couldn't you keep the existing stuff though and rearrange the scene to be bigger, possibly with additions?
Using a scene that is (hopefully …) still under construction makes future comparisons harder.

Oct 29, 2013permalink

David25

Please please add back the blue watch, some cameras could resolve the black hour marks and some could not!

Aug 16, 2013permalink

they got some spirals in different colors,
may be not enough?

Aug 16, 2013permalink

HelloToe

If you put the blue watch in this scene, in all likelihood *none* of them would be able to resolve the details of the watch. Remember that this new scene is much larger (like 6 feet across), and shot from farther away.

Aug 16, 2013permalink

petr marek

Comparsion with older cameras is crucial. Older test scene was good enough, I think better (especially shaded things in box) and not so flat.

Aug 16, 2013permalink

ET2

"Older test scene was good enough"

No, it wasn't. That's why it was replaced.

The older scene was too small, requiring the cameras to be too close to the scene. That introduced focusing and DOF issues.

Another problem was that some lenses (like RX100 lens) didn't perform optimally as not all lenses are optimized for close focusing performance. That meant the older scene was not representative of real world performance.

RX100 easily out resolves cameras like XZ2 in real life performance, but the older scene wouldn't tell you that.

Aug 16, 2013*permalink

@ET2 - exactly.

Aug 16, 2013permalink

M Jesper

Also "shaded things in box" has now been replaced by a fully darkened scene if you press that lightbulb. Showing all kinds of different shadow levels and detail. :)

Aug 17, 2013permalink

Eric Glam

Comparing the RAW files very closely (ISO 3200, ISO 6400), it looks to me like the NEX-6 and GX7 have exactly the same IQ.

Aug 16, 2013permalink