GX8 v G6 long exposures at night

I took my new GX8 for its maiden run shooting Aurora's a few nights ago, and before finishing up I took a shot with both my G6 and GX8 with the same settings (2500iso, F1.8, 13 seconds). Imported and pushed the exposure by +3 (nothing else done) with the results below. Definitely less noise and more detail in the shadows (bushes on the left and the rail on the right) and things like the water and sky look more natural. The G6 seems to struggle with noise on left and right edges compared to the gx8.

It is certainly an improvement, however I do wonder if the G7 at 16mp may have been a better choice for this type of shot.

GX8

G6

According to the Imaging Resource G7 review, the GX7 and the G7 have the same sensor.

Comparing GX8, GX7 (proxy for G7) and G6 on DXO Mark , at manufacturer ISO of 2500, the dynamic range of GX8 and GX7 appears to be about the same.

Those are nice shots, the only way I can think of improving is to drop the ISO of the GX8 to ISO 1600, underexpose (don't want to increase your exposure time, nicely chosen) , then push in PP.

Eric Nepean wrote:

According to the Imaging Resource G7 review, the GX7 and the G7 have the same sensor.

Comparing GX8, GX7 (proxy for G7) and G6 on DXO Mark , at manufacturer ISO of 2500, the dynamic range of GX8 and GX7 appears to be about the same.

Those are nice shots, the only way I can think of improving is to drop the ISO of the GX8 to ISO 1600, underexpose (don't want to increase your exposure time, nicely chosen) , then push in PP.

I think the G7 has an "improved" GX7 sensor that handles higher ISO's better then the GX7. There is certainly a marked difference between the G6 and the GX8 and I'll have a lot more room to push the shadows compared to what I could do previously and they should clean up nicely.

Hopefully I can get a chance to process the shots soon, including some panoramas with the Aurora dancing above a bush fire (should test the dynamic range!).

xelnt wrote:

Eric Nepean wrote:

According to the Imaging Resource G7 review, the GX7 and the G7 have the same sensor.

Comparing GX8, GX7 (proxy for G7) and G6 on DXO Mark , at manufacturer ISO of 2500, the dynamic range of GX8 and GX7 appears to be about the same.

Those are nice shots, the only way I can think of improving is to drop the ISO of the GX8 to ISO 1600, underexpose (don't want to increase your exposure time, nicely chosen) , then push in PP.

I think the G7 has an "improved" GX7 sensor that handles higher ISO's better then the GX7. There is certainly a marked difference between the G6 and the GX8 and I'll have a lot more room to push the shadows compared to what I could do previously and they should clean up nicely.

Hopefully I can get a chance to process the shots soon, including some panoramas with the Aurora dancing above a bush fire (should test the dynamic range!).

I'm curious, are you using RAW or OOC JPEGs? Which noise reduction SW do you use?

Eric Nepean wrote:

I'm curious, are you using RAW or OOC JPEGs? Which noise reduction SW do you use?

These are RAW with no noise reduction and the exposure boosted to see how the cameras handle the shadows. Editing wise I normally use Photoshop Elements (fine for what I do) however I need to update due to needing GX8 compatibility. So I'm trying out Lightroom at the moment, however it is a complete dog when you are opening folders and RAW files off a file server. Ridiculously slow. And I hate the Creative Cloud pay per month model.

xelnt wrote:

Eric Nepean wrote:

I'm curious, are you using RAW or OOC JPEGs? Which noise reduction SW do you use?

These are RAW with no noise reduction and the exposure boosted to see how the cameras handle the shadows. Editing wise I normally use Photoshop Elements (fine for what I do) however I need to update due to needing GX8 compatibility. So I'm trying out Lightroom at the moment, however it is a complete dog when you are opening folders and RAW files off a file server. Ridiculously slow. And I hate the Creative Cloud pay per month model.

I have noticed that for high ISO images, the noise reduction in OOC JPEGs  in some cameras is hard to beat in PP, might be worth a look.

For noise reduction and raw conversion, you might want to have a look at the DXO SW. Its on sale until Dec 25th, and is reputed to have really good noise reduction, and pretty good RAW conversion. YMMV. There is atrial.  And it can work as a plugin to Lightroom, I understand.

xelnt wrote:

I took my new GX8 for its maiden run shooting Aurora's a few nights ago, and before finishing up I took a shot with both my G6 and GX8 with the same settings (2500iso, F1.8, 13 seconds). Imported and pushed the exposure by +3 (nothing else done) with the results below.

Thanks for sharing this. I was wondering how big of an improvement it is, and as expected, it's significant.

It is certainly an improvement, however I do wonder if the G7 at 16mp may have been a better choice for this type of shot.

No, it would not. It would be roughly the same as G6 (probably a bit better, but not by much).

Those 16MP sensors from Panasonic exhibit really high levels of dark current, and that's probably the reason they do not have a real Bulb mode. GX8 does not have this problem. It can make 30 minute exposures, and behaves very well at long exposures. The only other Panasonic cameras that can handle long exposures well are GH3 and GH4.

I would love to see how GX8 fares against E-M10 II, which probably has the lowest high ISO noise among m43 cameras.

Eric Nepean wrote:

I'm curious, are you using RAW or OOC JPEGs? Which noise reduction SW do you use?

These are RAW with no noise reduction and the exposure boosted to see how the cameras handle the shadows. Editing wise I normally use Photoshop Elements (fine for what I do) however I need to update due to needing GX8 compatibility. So I'm trying out Lightroom at the moment, however it is a complete dog when you are opening folders and RAW files off a file server. Ridiculously slow. And I hate the Creative Cloud pay per month model.

You know you can still buy a standalone version of Lr right? For me buying LR PLUS Photoshop Elements is still cheaper than CC, since Lr isn't updated every year (so like $75 every ~2 for the upgrade version) and I only tend to grab every other version of PSE (usually on sale for like $80-100 for PSE + Premier Elements)...

That's like $160-170 every 2 years for LR + Photoshop/Premier Elements (or less depending on how much the Lr release schedule legs), vs what, like $240 every 2 years for just LR + Photoshop? I'm not opposed to CC on principle, but they'd have to sweeten the deal or kill their standalone software for it to appeal to me value wise.

Impulses wrote:

You know you can still buy a standalone version of Lr right? For me buying LR PLUS Photoshop Elements is still cheaper than CC, since Lr isn't updated every year (so like $75 every ~2 for the upgrade version) and I only tend to grab every other version of PSE (usually on sale for like $80-100 for PSE + Premier Elements)...

That's like $160-170 every 2 years for LR + Photoshop/Premier Elements (or less depending on how much the Lr release schedule legs), vs what, like $240 every 2 years for just LR + Photoshop? I'm not opposed to CC on principle, but they'd have to sweeten the deal or kill their standalone software for it to appeal to me value wise.

Yeah I'm aware of the standalone version, however what has put me off that is the deliberate crippling of the standalone version by stripping out features that they are adding to the CC version. That said I probably don't really need new features - its not as if Elements is updated between major versions so it shouldn't bother me anyway!

I need to play with lightroom more - however if it is going to be so slow even opening folders across a network to open a file I'll be giving it a miss. It seems to open every RAW file in the folder (say 20 meg each) just to render the thumbnail - annoying.

Eric Nepean wrote:

I have noticed that for high ISO images, the noise reduction in OOC JPEGs in some cameras is hard to beat in PP, might be worth a look.

For noise reduction and raw conversion, you might want to have a look at the DXO SW. Its on sale until Dec 25th, and is reputed to have really good noise reduction, and pretty good RAW conversion. YMMV. There is atrial. And it can work as a plugin to Lightroom, I understand.

I haven't got around to DXO yet - not sure how well it goes with post processing however.

Astrotripper wrote:

xelnt wrote:

I took my new GX8 for its maiden run shooting Aurora's a few nights ago, and before finishing up I took a shot with both my G6 and GX8 with the same settings (2500iso, F1.8, 13 seconds). Imported and pushed the exposure by +3 (nothing else done) with the results below.

Thanks for sharing this. I was wondering how big of an improvement it is, and as expected, it's significant.

It is certainly an improvement, however I do wonder if the G7 at 16mp may have been a better choice for this type of shot.

No, it would not. It would be roughly the same as G6 (probably a bit better, but not by much).

Those 16MP sensors from Panasonic exhibit really high levels of dark current, and that's probably the reason they do not have a real Bulb mode. GX8 does not have this problem. It can make 30 minute exposures, and behaves very well at long exposures. The only other Panasonic cameras that can handle long exposures well are GH3 and GH4.

Interesting - thanks for that. This was the first time out under the stars with the GX8 so I haven't worked out how far I can push it - I know the limits of the G6 and now I need to figure out how far I can push the GX8.

And grab one of those rumoured 12mm F1.2 Olympus lenses...

Hi,

I am also tempted by the DXO discount. I have been using photoninja+photo mechanic for a couple of years. I think I spend too much time pulling sliders to get the colours right in photo ninja so I am looking for something else. I just stumbled across raw therapee when googling DXO reviews etc. After just two days of testing Raw therapee I am able to beat photo ninja on a couple of my more problematic pictures. Thats quite Promising. I also like the white balance defaults and controls better than the ones in Photo ninja. You may want to check it out. Raw therapee is licensed under GPL and will be free forever.

http://rawtherapee.com/

Regarding your pictures:

I was hoping that the improvement from G6 to GX8 would be even bigger. I need more fuel for my GAS 8-)

BR,

Øyvin Eikeland

PS! Sorry for the off-topic, but you started it

I understand you want to show us the differences and the differences are clear.

When I look specifically at the pics I would say it is overexposed. There is no need to see the beach and all other uninteresting things in the foreground. But with F1.8 and ISO2500 there was not much of a visible aurora to speak of clearly so we would probably see nothing if you exposed "right".

mFt sensors have not gained even 1 stop over the past 9 years when we look at noise. Not per DxO and not what we see here where G6 represents the typical mFT sensors 2008-2012. Of course this one was clearly better than the G1 but it was basically due to much better dynamic range.
The best mFT currently are almost identical to the Sony Nex3 from 2010 when it comes to noise. We have gotten us 1/2 a stop extra may be. For such photography that is simply not good enough, but in the end what hampered things the most in this case was the fact that it was merely a photographic aurora...

Jorginho wrote:

I understand you want to show us the differences and the differences are clear.

mFt sensors have not gained even 1 stop over the past 9 years when we look at noise.

So which is it?

This example really demonstrates the difference very clearly. Bump that exposure time even more, and the difference will be even greater (and amp-glow will start showing its ugly head as well).

Not per DxO and not what we see here where G6 represents the typical mFT sensors 2008-2012.

DxO does not measure dark current, which is the predominant noise source at long exposures like that. And the differences between Panasonic and Sony sensors in this regard are huge, as demonstrated here. So an E-M10 for example, will perform much better at long exposures than most Panasonic cameras except the ones I mentioned earlier. And the difference is much bigger that what DxO or other noise measurements done in "normal" conditions would let you believe.

BTW, the Nex-3 you mention, is actually one of the best APS-C cameras in terms of handling dark current, and would leave both of the Panasonics in the dust for shots of this kind.

Astrotripper wrote:

Jorginho wrote:

I understand you want to show us the differences and the differences are clear.

mFt sensors have not gained even 1 stop over the past 9 years when we look at noise.

So which is it?

All mFts "is it".

This example really demonstrates the difference very clearly. Bump that exposure time even more, and the difference will be even greater (and amp-glow will start showing its ugly head as well).

I see a difference, I do not see 1 stop difference (subjective like hellof course).

Not per DxO and not what we see here where G6 represents the typical mFT sensors 2008-2012.

DxO does not measure dark current, which is the predominant noise source at long exposures like that.

I use G1, GH2, EPL5, GH4 and have used them all for either aurora borealis and noctilucent clouds. EPl5 and GH4 (Sony vs Panny) was actually a win for the latter in the exact same situation if only slight. I think the Panny retains colours and tonality a bit better (RAW) and noise too. It seems thought that beyond 15 seconds the Sony does better.
GH2, EPl5 and GH4 are reasonably similar for low light noise but EPl5 and GH4 indeed are somewhat better. I wonder if the difference reaches significance.

That is my personal experience which remains in line with DxO.

And the differences between Panasonic and Sony sensors in this regard are huge, as demonstrated here.

Nonsense. You cannot attribute this to Panny vs Sony since one uses 2010 tech and the other 2015 tech.

So an E-M10 for example, will perform much better at long exposures than most Panasonic cameras except the ones I mentioned earlier.

Well not for 8s exposures where EPL5 does not do anything better than a Gh4 and in fact does worse.

And the difference is much bigger that what DxO or other noise measurements done in "normal" conditions would let you believe.

Beyond 15 seconds may be. At shorter exposures that is not the case.

BTW, the Nex-3 you mention, is actually one of the best APS-C cameras in terms of handling dark current, and would leave both of the Panasonics in the dust for shots of this kind.

So. How does this help mFts. A 2010 APS-c that does much better according to you....
I feel mFts can gain a lot when it comes to noise these days. May be new organic sensors will solve this, may be BSI tech or somehting else. I hope so.

Jorginho wrote:

I understand you want to show us the differences and the differences are clear.

When I look specifically at the pics I would say it is overexposed. There is no need to see the beach and all other uninteresting things in the foreground. But with F1.8 and ISO2500 there was not much of a visible aurora to speak of clearly so we would probably see nothing if you exposed "right".

I'm not sure you don't understand what I was doing here sorry - this was not about the composition of the photo or the Aurora - it is about how the cameras compared in long exposures and the shadow detail they picked up. The un-edited photo is by no means over exposed. In post I pushed the exposure by +3 to see how cleanly the shadow detail had been resolved.  And I'll disagree about foregrounds - in Aurora photography quite often it is just as much the foreground as the background that makes a photo.

xelnt wrote:

Jorginho wrote:

I understand you want to show us the differences and the differences are clear.

When I look specifically at the pics I would say it is overexposed. There is no need to see the beach and all other uninteresting things in the foreground. But with F1.8 and ISO2500 there was not much of a visible aurora to speak of clearly so we would probably see nothing if you exposed "right".

I'm not sure you don't understand what I was doing here sorry - this was not about the composition of the photo or the Aurora - it is about how the cameras compared in long exposures and the shadow detail they picked up. The un-edited photo is by no means over exposed. In post I pushed the exposure by +3 to see how cleanly the shadow detail had been resolved. And I'll disagree about foregrounds - in Aurora photography quite often it is just as much the foreground as the background that makes a photo.

I am sure it can be. But not in this case to my mind, the foreground is just a mess. Sometimes it can be very usefull, I agree.