G9 with birds

[BrentC]

G9 vs. D500 vs. EM1 BIF Shootout-Part 2

The Ratings and Results

Shooting almost 6,500 photos with 3 camera for this test meant I had to have a quick and simple system for reviewing and rating the photos. I choose LR Classic CC (Mac) as the G9 is supported and used a 1, 2 or 3 star system. With the caps lock key on, as I hit the appropriate number key, it would advance to the next photo. Still it took 2-1/2 days though to go through all the photos. 1 star means the photo was out of focus, 2 stars means the photo looked OK at the normal viewing size on my 27” monitors and 3 stars means when the photo is zoomed in to 100%, it is critically sharp. By this I mean the eyes were in perfect focus, the feathers had great details, the feet, if visible, were sharp. Any blurred areas and the photo was rated a 2. As most of these photos were taken at distances of 20 meters or more to the birds, the DOF was not really an issue. For the few photos taken at 15 meters or less, if the AF was properly focused on the birds head and some of the wing tips were soft because of the DOF, it was given a rating of 3.

Here are the rating results on 6,442 photos listed by the best performing in Rating #3 Critically Sharp.

View attachment 611088

OK, no surprise, the D500 AF cleaned house…. did anyone actually think it wouldn’t? But the G9 with the Olympus 40-150 did very well against the comparable FL Nikon zoom, being less than 10% behind in the #3 Critically Sharp rating. I am impressed. Compare this to the EM1 with the 40-150, only 1/3 the amount of keepers vs the D500 and the zoom. This has been my long term experience for the EM1, I felt lucky if my keeper rate ever got to 20% for these types of BIF photos. This is the reason I only shot the EM1 for less than 450 photos, the results are not going to change.

The big surprise though is how bad the G9 with the 40-150 and the MC-14 did. Less than 7% keepers! So many of the photos were were just off by a bit, and thus were rated as a 2. As I mentioned in the Gear Setups above, I did not shoot the EM1 with the MC-14, but in the past I have gotten very good results with this combo for other types of non-action shooting. On the 2nd day I took about 50 photos with the G9 and MC-14 of some terns and plover shorebirds who were not flying but just walking in shallow water, and even those results were dismal. Something just can’t be right with this and I plan on going back next week to reshoot both the G9 and EM1 with the MC-14 (with cleaned contacts).

The other surprise to me is that the best AF performance was with my longest lens, the AFS 300mm f4. For comparison, this lens is just about the same size and weight as the Olympus 300mm f4, being 6mm shorter and 35 grams lighter. With the 1.5x crop factor of the D500 bringing the FL to 450mm, this was the most fun lens to shoot with. But as far as making a small kit, the 180mm f2.8 easily wins this, it is almost 1/2 the weight, thinner and 80mm shorter than 70-200 Zoom or 300mm f/4 and has a really cool sliding built-in metal hood. Heck, it is even thinner and smaller than the Olympus 40-150mm f2.8 (with its huge plastic hood), same weight though, and it becomes a 270mm f2.8, pretty close to the FL of the Olympus zoomed out, but with a much higher keeper rate. I have not used this lens for fast actions or BIF shots before, it has been mostly used for portraits, but that will change because of the impressive results I obtained in this test.

While I haven’t mentioned IQ, I will say that all 3 of these camera produce great results. I will continue to use my EM1’s even though I have a newer 20MP body in the G9. And as far as a difference between the G9 and the D500, well even at 100% crop I do not see a difference with the photos I took. Maybe in your situation you will, but for me they are equals in the IQ department at base ISO. Hope this test is useful for some of you. If you have any comments, suggestions or want to lend me a EM1 MKII to do additional tests, just let me know.

G9 40-150 @150 with MC-14. One of the very few keepers I got with this combination.
View attachment 611089

I hope you can get a hold of am EM1.2. I'd be very interested in the results.

 

[barry]

@SpecFoto, thanks for the detailed results!
Just curious, are you running firmware 4.3 on your E-M1's?

 

[SpecFoto]

@SpecFoto, thanks for the detailed results!
Just curious, are you running firmware 4.3 on your E-M1's?

Your welcome Barry, glad you found the info useful. Yes, both my EM1's are using the latest FW V4.3.

 

[Armanius]

@SpecFoto Thanks for the test! Much appreciated!

Did you fiddle with any of the AFC customization parameters in the G9?

 

[SpecFoto]

@SpecFoto Thanks for the test! Much appreciated! Did you fiddle with any of the AFC customization parameters in the G9?

As I posted in the 1st part under Photo Gear Setups, 3rd paragraph, "The G9 was set to the “I” Burst mode L (2 FPS, see G9 Issues below) with AF Custom Setting Set 4 and IBIS on". Didn't try any other setting as that is the one the manual says is for the most active subjects. Custom setting 2 however can be used for panning, so it may have also worked. I plan on trying the #2 setting when I get out there next week to see if it helps to improve the MC-14 results.

 

[Armanius]

As I posted in the 1st part under Photo Gear Setups, 3rd paragraph, "The G9 was set to the “I” Burst mode L (2 FPS, see G9 Issues below) with AF Custom Setting Set 4 and IBIS on". Didn't try any other setting as that is the one the manual says is for the most active subjects. Custom setting 2 however can be used for panning, so it may have also worked. I plan on trying the #2 setting when I get out there next week to see if it helps to improve the MC-14 results.

Thank you for the very informative testing! I really appreciate it .

 

[Ziggy]

A few shots with the PL100-400 in both good light and shadows. Shooting shutter priority 1/800 and faster and up to 1600 ISO.

This Eastern Yellow Robin came up well with a big crop but going on a quarter of the shots were fails due to preferencing the background again.

Subscribe to see EXIF info for this image (if available)

Here's a few King Parrots. IQ is good in direct light and the system worked surprisingly well in shadows close to a stationary bird - there was a very good rate of eye catchlights in focus. With these birds and the Wood Ducks moiré is quite common. Occasionally it ruins a shot. Same deal with the G85.
These are smart birds like other parrots and will make eye contact and chat with you ... in this case it's both eyes in contact.

Subscribe to see EXIF info for this image (if available)

Rump plumage detail with moiré. Someone said stopping down deals with it.

Subscribe to see EXIF info for this image (if available)

Adult male ...

Subscribe to see EXIF info for this image (if available)

Moulting male ...

Subscribe to see EXIF info for this image (if available)

And again ..

Subscribe to see EXIF info for this image (if available)

Adult male. You could swear that often they're posing...

Subscribe to see EXIF info for this image (if available)

Wood Duck ...

Subscribe to see EXIF info for this image (if available)

On the debit side, apart from the preference for the more distant lock, once again there was an occasion when a focus box, one of a cluster, was green but not focussed. Also a number of times the only way AF could be achieved was to find something nearby, get focus and then retry focus on the subject.

Added - settings:
Diamond AF spread
Usually 1/1600s but lower if the scene is predictable
Usually the 100-400 is fully extended or close to it
AFS or AFC with repeated refocusing
9 fps RAW

 

[Ziggy]

Well actually in part 1 of the post under the G9 Issues, 2nd paragraph I told how I inadvertently changed the G9 Control dial a few times for about 1/3 of the total G9 photos (1,000), affecting the aperture. The aperture went from f2.8 to between f3.5 and f5.6, but the rating numbers were pretty consistent with the overall G9 rating. It was a bright day, but I did not notice ANY AF hunting or slowness at all with the 40-150mm and the G9 during this test. But yeah, when back there next week I will try to shoot at the mid range f stops and let the ISO climb, as I want to see what noise I will find at higher ISO's.

Thanks for the interesting tests. Keep up the good work.

Bear in mind that the G9's default on Focus/Release priority in AFC is "balanced", ie. if it can't get a lock it attempts prediction. Throw a foreign lens and a TC into the mix and perhaps your result could have been expected.

In AF terms I've seen an Olympus lens behave differently on a Panasonic body from a Panasonic lens under the same shooting conditions .

 

[whumber]

Very detailed post @SpecFoto. I’m surprised that the E-M1 did so piss-poor, and I’m also surprised at the lower performance of the 70-200 VR. It must be something to do with it’s age. I have a VR II with mine, and it performs great. I can’t see why anyone would buy the latest VR III unless they’re shooting skmehig like a D850.

The G9 did great with the supposed non-DFD compatible 40-150 PRO. I truly wonder if there’s actually a difference in performance, or if Panasonic creates a profile for the Olympus lenses but is simply telling people they didn’t to try to drive sales of Panasonic lenses.

I don't find the E-M1 results so surprising, the camera takes a very long time to acquire initial focus and doesn't do well with accelerating targets. Most of the good results that get posted are subjects where there's plenty of time to get focus lock and with a subject moving at a fairly constant speed.

 

[Ziggy]

Playing around with responses to the preference for the more distant lock. Trying a couple of options: first, letting the birds fly into the frame; second, going for small apertures so the DoF includes the birds.

Re the first, it seems to me if the camera AFC is tuned to focus on what's moving, when you're panning that will be the background, and if it's textured it will get the lock. If that's right, one response will be to avoid panning and try and let the birds fly into the frame.

I've had a try of both of these and they're worth further work. This pic used both strategies. Going up to 1600 ISO allowed f16. The result was pretty grainy but controlled well enough in LR.
Wood Ducks coming into land ...

Subscribe to see EXIF info for this image (if available)

 

[Drdave944]

Do you use a tripod or have a blind? I don't have any 800 mm equivalent equipment,but can't get that close,so always have to crop.

And some comparatively easy shots of perching birds. There is pleasing colour and detail to my eye.
A Lt Pied Cormorant protests the invasion of personal space by a White Ibis:
View attachment 607351
One of the resident Australasian Darters. Plumage colour and detail is rather better in the RAW version of course.
View attachment 607352


Processing notes:
In-camera sharpening set to -5 Correction: NR set to - 5
Images are not cropped or otherwise edited. They were shot in RAW and opened in LR for export as JPEGs for screen without added compression and with standard sharpening.

Next outing I'll work with AFC and look at setting the camera to Balanced on Focus/Release priority as that attempts some prediction (if it functions like the G85).[/QUOTE]

 

[Armanius]

Looking at page 117 of the advanced manual, it seems that Balanced vs. Focus vs. Release doesn't make any difference when using the camera in AFC/AFF and burst speeds of SH1, SH2, SH1 pre, SH2 pre, or H. The manual simply says that under these settings, the auto focus logic uses "predicted focusing."

 

[panamike]

Looking at page 117 of the advanced manual, it seems that Balanced vs. Focus vs. Release doesn't make any difference when using the camera in AFC/AFF and burst speeds of SH1, SH2, SH1 pre, SH2 pre, or H. The manual simply says that under these settings, the auto focus logic uses "predicted focusing."

That puts me off a bit.

 

[Armanius]

That puts me off a bit.

Maybe I'm reading it wrong, or there are sublevels of predicted focusing that the manual doesn't discuss?

 

[Ziggy]

Do you use a tripod or have a blind? I don't have any 800 mm equivalent equipment,but can't get that close,so always have to crop.

And some comparatively easy shots of perching birds. There is pleasing colour and detail to my eye.
A Lt Pied Cormorant protests the invasion of personal space by a White Ibis:
View attachment 607351
One of the resident Australasian Darters. Plumage colour and detail is rather better in the RAW version of course.
View attachment 607352


Processing notes:
In-camera sharpening set to -5 Correction: NR set to - 5
Images are not cropped or otherwise edited. They were shot in RAW and opened in LR for export as JPEGs for screen without added compression and with standard sharpening.

Next outing I'll work with AFC and look at setting the camera to Balanced on Focus/Release priority as that attempts some prediction (if it functions like the G85).

[/QUOTE]
No drdave. Good IS and 800mm FE mean you can shoot from some distance away and handheld - though my default speed is 1/1600 so a tripod isn't needed anyway . And the 20 MP sensor provides some scope for cropping.

 

[Ziggy]

Looking at page 117 of the advanced manual, it seems that Balanced vs. Focus vs. Release doesn't make any difference when using the camera in AFC/AFF and burst speeds of SH1, SH2, SH1 pre, SH2 pre, or H. The manual simply says that under these settings, the auto focus logic uses "predicted focusing."

Yes, so how well is prediction done?
The camera is gathering data at up to 9 to 60 samples per second.

 

[Armanius]

Yes, so how well is prediction done?
The camera is gathering data at up to 9 to 60 samples per second.

Beats me. I'd love to know. Why allow user to choose if it's all the same right?

 

[Drdave944]

Do you use a tripod or have a blind? I don't have any 800 mm equivalent equipment,but can't get that close,so always have to crop.

This keeps showing up as though these were my pictures . No they are a Ziggy's Melbourne Australia.

Processing notes:
In-camera sharpening set to -5 Correction: NR set to - 5
Images are not cropped or otherwise edited. They were shot in RAW and opened in LR for export as JPEGs for screen without added compression and with standard sharpening.

Next outing I'll work with AFC and look at setting the camera to Balanced on Focus/Release priority as that attempts some prediction (if it functions like the G85).

 

[Ziggy]

Beats me. I'd love to know. Why allow user to choose if it's all the same right?

I have no trouble getting a focus, just that in the mid to distant ground there's a preference for the more distant lock.
Speculating on how prediction might work ...
You've chosen a fast frame rate so the camera prioritises that over achieving a focus before releasing.
It may aim for real-time focus say in shots 1 and 2. With a trajectory noted it starts predicting, using DFD which tells it how far to move the lens to hit the projected next point. It can then assess the degree of pixel contrast achieved cp with that expected, and make adjustments for the next lens move. So in other words it's extrapolating from each shot after the fact rather than waiting to get a focus first.
Apparently it can assess pixel contrast in less than 400ths of a second.
And in the field I've several burst sequences that in fact show softening focus that was then recovered.

 

[Gatormac2112]

Yes, so how well is prediction done?
The camera is gathering data at up to 9 to 60 samples per second.

I wonder what determines how many samples per second it takes. I'm assuming the more samples the better, maybe release only does 9, balanced does something like 30, and focus does 60. Thats just something I pulled out of thin air, but I may try focus preference to see if the hit rate is higher next time.