Olympus E-M5 Mark II - love and hate at first sight

[DPC]

At the end of a day testing cameras for a shoot tomorrow, all I can say is that it's not a straightforward as I thought when I posted this morning's test.

In fact, against my GH4 and 5D MK III, the EM5 Mk2's skin tones look really good. The colours are great and the files are very easily tweaked. 

Downloading John Brawley's ungraded files is quite informative, by the way.

[DPC]

At the end of a day testing cameras for a shoot tomorrow, all I can say is that it's not a straightforward as I thought when I posted this morning's test.

In fact, against my GH4 and 5D MK III, the EM5 Mk2's skin tones look really good. The colours are great and the files are very easily tweaked. 

Downloading John Brawley's ungraded files is quite informative, by the way.

[Marcio Kabke Pinheiro]

I look at this and start to care alot less about a bit of moire and aliasing. They are giving you a gimbal at sensor level... It is incredible. This opens up far more cinematic qualities than 4K. Camera movement counts for so much.

 

 

 

​The low light performance was better than I expected. The softness is present, but worked well with the video. Only the aliasing was perturbuing, when filming the car lights.

[Flynn]

 

In an otherwise favorable review this says the continuous autofocus is not good. http://www.mirrorlessons.com/2015/02/05/olympus-om-d-e-m5-mark-ii-review-chapter-whats-new-whats-great/

[Inquisitive]

I think the real issue with this camera is not the lack of 4k or even the moire we've seen it I can live with it.  That being said is it really a 2.7 crop factor ibis on or off?  So a 7mm is an equivalent 18.9 vs 14 normally that disturbs me.  I was comfortable with the m4/3 compromise but I don't know if I want to compromise even more!  How does this compare to the GH4 4k crop?

[Cinegain]

GH4 4K is 2.3x crop. BMPCC is still the worst though at close to 3x. But that's why there's focal reducers.

In meanwhile... (FT4) UPDATED: Yes, GX7 successor is coming in May/June.

[Rinaldo]

Some disagree.... http://robinwong.blogspot.com.br/2015/03/olympus-om-d-e-m5-mark-ii-vs-e-m1-video.html

[Don Kotlos]

Some disagree.... http://robinwong.blogspot.com.br/2015/03/olympus-om-d-e-m5-mark-ii-vs-e-m1-video.html

​They do not just disagree, they provide evidence for it. That test replicates the same findings of my own test so I am pretty sure something wrong happened to Andrew's review.  

[presbytis]

I see things seem to have gotten a little hysterical on this thread regarding the E-M5 mark ii! I've had a decent look at what little material there is on the web and have made a few observations. Have people seen the new stabilisation comparison between the E-M5 mark ii and a GH4 posted on 43rumors.com recently? Pretty incredible really how good the Olympus stabilisation is in comparison. This is indeed what I think Olympus mean by cinematic quality images in their advertising blurbs - cinematic quality 'motion' in images (after all almost any mention of video in their advertising is associated with the mention of stabilisation). Apart from the stabilisation, one of the first things you notice is indeed the added crop on the E-M5. I doubt the crop factor is as much as 2.7x though, considering the GH4 was shooting 1080p and not 4K. I'm sure this added crop is a technical constraint required to give the degree of sensor shift needed for effective stabilisation of the amount of movement generated between the frames of video footage. After all, a person trying to keep still taking a still frame is a very different 'kettle-of-fish' to a video camera on the move. The oversize of the standard m43 lens image circle as opposed to the used area of the m43 sensor probably allows enough shift to compensate for the shake/movement produced when taking a still, but the added movement generated between frames of moving video probably requires this extra crop to be effective. With all the talk about moire and aliasing the other thing that surprised me with this footage was that I didn't notice any on the E-M5 side but did in the windows of the distant buildings on the GH4. Having said this though, the compression of this footage is appalling, making it difficult to say anything much. One thing I did notice though that is interesting, is the motion cadence of the E-M5 footage is extremely different from that of the GH4 even though the frame rate settings etc. are the same. Although the video from both look sharp when playing, when you pause the frame the GH4 frame looks much sharper due to the E-M5 footage containing much more motion blur. It looks like the difference between using a 180 degree and a 360 degree shutter. Perhaps this explains some of the disparity in results between different people's testing? If one test was from a tripod the frame grab from the E-M5 would appear sharp but if the frame grab was from an image that was handheld using stabilisation it may contain a bit more motion blur making it appear softer. Looking at Robin Wong's test I must say I agree with him that the E-M5 mark ii footage looks sharper and has much better dynamic range than the E-M1 footage. The other curious thing I notice (and others noticed this too in the bath towels test posted here) is that the E-M5 mark ii footage appears to have greater depth of field using the same lens at the same setting as opposed to the E-M1. Especially look at the dials on the top of the close up of the camera - what's going on here? I must say also the look of some of the videos posted on the web from the E-M5 mark ii seem really good. I actually downloaded the ungraded sample footage that John Brawley shot and cut this material into a timeline of shots I'd recently filmed for a commercial on my A7s. While the scenes were obviously completely different, I thought the material actually cut very well - it wasn't like I was noticing a big resolution hit every time I cut to the Olympus footage and that was viewing on a 4K monitor. I just think 'horses-for-courses'. I'd potentially use the E-M5 for tracking and movement shots within sequences of material shot by other cameras. I probably wouldn't mix static shots with those from another camera like the A7s within the same scene. But having said this, that's generally standard good practise anyway because different cameras produce different colour responses (especially to skin tones that we really notice), making the footage extremely difficult to grade.

[Bob Goldberg]

Hi Andrew,

 

I'm a regular member of DPReview and I was actually responsible for a couple threads talking about the E-M5 II video, generally in an unfavorable light. I'm not sure if you've had a chance to read them. One of them was in response to your first impressions.

 

I generally agree with your comments about the E-M5 II video being disappointing, but, in all fairness, I did want to make sure to ask you to clarify your comments regarding the video being worse than the E-M1, both in regard to sharpness/fine detail and moire. There was a lot of concern about this specific portion of your impressions on DPReview, and I was hoping you could provide some evidence in this regard (i.e. a comparison between these two cameras), possibly in your more comprehensive review later.

 

I don't think many people are questioning that the E-M5 II isn't up to GH3 standards (or Panasonic standards in general for video), but they are questioning that the E-M5 II isn't up to E-M1 standards.

 

Thanks

[Portlandia]

Per a quick framing test the crop factor appears to be about 2.4x. Good news is the Panasonic 15 1.7 becomes around a 35mm equivalent field of view.

[frosti7]

I'm not discovering america, but Olympus obviously use heavy Software-IS, hence the crop&lower quality (and the super stabilization)
its a good idea, but they should provide allow to turn it off

[presbytis]

I'm not discovering america, but Olympus obviously use heavy Software-IS, hence the crop&lower quality (and the super stabilization)
its a good idea, but they should provide allow to turn it off

​But the E-M5 II does allow this. The IS has three settings: completely OFF; IS mode 1, which does sensor shift and a digital stabilise with a slight additional crop; and IS mode 2, which only does the sensor shift.

[Inquisitive]

What I think Olympus should have done is documented the crop in specifications than it wouldn't be a debate.  As is we have to depend on others to tell us.I hope it is 2.4 that is ok 2.7 was crossing a line.  I know why there is a crop that being said if I need a wide shot and could turn IBIS off why should I still loose my field of view.  I'm sure there is a good reason don't feel anyone has to speculate for me and happily if it's 2.4 it's not as big a deal.

[Cinegain]

Here's a translation of a review left on Amazon.de by someone...

Long have I waited for this. Olympus finally seemed to have put the focus on video with this camera. Unfortunately the video mode lets down. Yes, finally there's 24p and a high bitrate codec... but that's rendered useless when the actual image quality just doesn't hold up, I would even go as far to say the old E-M5 did a better job. Videos aren't sharp (compared to the GH4 with same lens, the Olympus looks like upscaled 720p), lots of moiré, rolling shutter at first glance also isn't very impressive and dynamic range could've been better. Such a pity. Autofocus during video... forget it. Generally speaking continuous AF doesn't appear to be all that snappy and especially in less favorable light conditions it's searching quite a bit.

IBIS: rather brilliant, makes a gimbal stabilizer unnecessary 90% of the time, even when running it looks alright.
Photo quality: as you would expect from Olympus, nothing to be criticized here.

Overall I'm giving this 3 out of 5 stars. For someone who mainly shoots stills and regards the video mode as a nice added feature, I can really recommend you this camera. For someone who's mainly init for video... unfortunately I can not.

I'll continue using my A7S and GH4 + gimbal combo. A real shame, IBIS would have made shooting a breeze, but it's of no use when image quality has to suffer like this

[fuzzynormal]

Since I purchased this camera for specific gigs that I do often: run 'n gun small biz profiles for the web @30p, the question I'm asking myself right now is, "Should I return the M5II and go back to the original E-M5?"  As a video camera there's not a lot of difference between the two...except about half the price.

Sure, all the internet stuff raves about how the video quality is increased, and it is to an extent, but is it enough?  Maybe.  Maybe not.

Again, just wish none of this was even a consideration.  Such a shame it is.

[Cinegain]

Yeah, although it's availlable practically everywhere I look (I went into a store close by just now, they had it), the only place it's not availlable is where I pre-ordered (silly me, I thought that would put me ahead). I think it's the universe trying to tell me something. So I'm going to listen and told them they could just go ahead and cancel it...

[fuzzynormal]

Here's what's frustrating:  The video is old 7D quality.  In 2015 Oly should be more competitive than that.  

But, the stabilization is a freaking great tool.

Will decent video and Oly's 5-axis ever unite?  One day, just not today.  So, as we cry in all our collective internet tears, here's a clip that shows why we're longing for Oly to merge great video with great stabilization.  It's not much by way of a good shot, but keep in mind it's shot HANDHELD at a 600mm FFEquiv focal length.  Yes.  600mm.  Handheld.  This shot would be impossible otherwise:

SPRK-150.mp4

[Don Kotlos]

Here's what's frustrating:  The video is old 7D quality.  In 2015 Oly should be more competitive than that.  

But, the stabilization is a freaking great tool.

Will decent video and Oly's 5-axis ever unite?  One day, just not today.  So, as we cry in all our collective internet tears, here's a clip that shows why we're longing for Oly to merge great video with great stabilization.  It's not much by way of a good shot, but keep in mind it's shot HANDHELD at a 600mm FFEquiv focal length.  Yes.  600mm.  Handheld.  This shot would be impossible otherwise:

SPRK-150.mp4

​I don't know why but my E-M5ii can produce videos with much better resolution that your sample. Do you use a profile with reduced sharpening? I have seen that with the E-M5ii that will decrease the quality so '0' should be optimal. If that is not the case then I think that the stabilization algorithm might be causing this blurring. 

[fuzzynormal]

​I don't know why but my E-M5ii can produce videos with much better resolution that your sample. 

It's a 600mm shot.  That's kind of a long lens.  The photons are traveling through a lot of atmosphere in a jungle, after a rain fall, at 95° Fahrenheit.  And the posted resolution of the video is 846x468.  Yes, I turn down the sharpening all the way. 

Pixel peeping on this shot?  Not a good idea.  Point of me posting this is that I got watchable video with a handheld camera and a 600mm focal length.