Reading mode: Light Dark

Big Metabones smart adapter firmware updates

Go to last post
Robert

Thanks for the nudge.

The announcement raises several issues that are worth noting.

I was always under the impression that the really good AF performance with EF Lenses on M4/3 bodies was matched by their adapters on Sony A7 series MkII bodies. Now here is an announcement that E mount bodies are playing catch up.

It is not clear on whether the performance of EF lenses on MkI Sony 7 series bodies is now acceptable. The comments only talk about "Sony E bodies".

Therefore I remain puzzled on whether A7 II bodies have been doing AF better all along and now the other Sony bodies are working better. Or whether the A7 II bodies are now working as well as the M4/3 bodies and nothing much is said for the more stuttering preformance of earlier A7 I bodies. But I guess they are lumped in with the generic reference to "E mount".

As I have no access to an A7 II body I suppose I have every ability to upload the software on to my MkIV adapter and see how it performs with EF lenses on my A7R MkI and NEX6.

I will do so as soon as I can fit it in.

Interestingly there is also a confused message about the firmware being the same for M4/3 but I think that this is just the uploading and installation procedure. Again I should check to see if there are any extra goodies for the M4/3 mount crowd who are already smirking like that Cheshire Cat with their superior AF perfomance from Canon EF lenses.

After some "chat" about the Sigma electronic EF to E adapter it is noticeable that dpr has expressed some caveats as to how well it works. Sigma have a lot of experience in the EF mount and its operating system but I suggest that making compatibilty with the twists and turns of an electronic interface without full co-operation from the oem owner is not exactly a walk in the park.

Lastly that wonderful lens - the Canon EF 135mm f2.0 is still a recalcitrant for full compatibility. Pity.
 
Shortly after I posted this, dpreview posted an update reporting severe issues with the new firmware; after trying an updated adapter on an A7R II, they "do not recommend you even bother trying".

That said, those reservations probably are in no way valid for the Micro Four Thirds version, and it is unclear how other Sony bodies (those without the ability to drive adapted lenses' AF in on-sensor phase detection mode) behave with the new firmware.
 
From what I'm reading elswhere in a Micro Four Thirds context, contrary to the issues found with the E-Mount adapter on the Sony A7R II, firmware 2.3 seems to impose no problems.

First reports speak of a slightly accelerated phase detection AF with the E-M1 and that a 2.2 issue which could make a CDAF OM-D freeze at infinity focus with specific (older) lenses seems to be fixed.

(Interestingly, for firmware 2.2 some people had reported that at least with specific lenses, the 400mm f/5.6 L being one of them if I remember correctly, CDAF with an E-M10 II had already become both faster and more precise than PDAF was with the E-M1 at that time, something I would have deemed physically impossible until then.)

Someone said the 135 L would "run slightly better" now with the E-M1, whatever that means.

Oh, and someone reported that the Tamron 150-600 would work now, after none of his MFT bodies used to function anymore as soon as that lens was attached.

Which actually is a message that could very well be just what I needed to eventually buy a smart adapter (plus the Tamron lens) myself...
 
Last edited:
@Tom Caldwell

The first fast AF (for adapted lens on mirrorless) was in last July on Panasonic, when the SB ULTRA 0.71x for MFT was launched.

Next comes the E-M1 and A7RII in September. Back then, they were the only cameras engineered to autofocus quickly with DSLR lenses (classic Four Thirds lenses and Sony/Minolta A-Mount lenses respectively) using phase-detect AF. Now, they are joined by A7II and A6300. There are certainly other cameras featuring PDAF, but only these four use PDAF to drive DSLR lenses to get the same AF speed as a DSLR (and better than CDAF by a good margin). Olympus decided that there was no point for an E-7 because E-M1 was fully capable of doing that job. Olympus CDAF for E-M5/10 also came at the same time (September).

It took several fw iterations for A7RII to get to the point where it was more-or-less on par with a genuine Canon in AF speed. (CDAF on a Panasonic will always lag behind a Canon DSLR in AF speed with the same lens.) But E-M1 never got the necessary fine-tuning to attain speed-parity, until now. But E-M1 AF-C still doesn't work.

The rest of the Sony line-up, A7S, A6000, etc., had always had very slow AF, which took seconds to lock. It was only until now that they get fast CDAF, using a method similar to that used for Panasonic and Olympus (other than E-M1).

But CDAF is slower than dedicated PDAF. The fiasco today from the E-mount side is primarily owing to A7RII, A7II and A6300 users who didn't know about or were unwilling to press the necessary buttons to get back the old PDAF behaviour.

Metabones supports this huge range of cameras, from JVC, AltaCine F5, Blackmagic to the brands I have already mentioned above. Each firmware release has to cater to every camera (not to mention the diversity on the lens side). There was some impedance mismatch between the expectations of A7RII, A7II and A6300 users and what they get, and their reaction has been very vocal. But as far as I can gather, no camera and no lens have gotten any worse, and many of them got numerous small improvements. (As the community has noted, on the MFT side we have fixed a few problems.) Some new features might have failed to meet expectations, but all you have to do is to opt-out but not use them.

--
Bo-Ming / conurus
 
Last edited:
Thanks Bo Ming,

But can you specifically advise if there has been any recent improvement in the performance of the Sony A7R MkI body with EF lenses. I have no interest in the MkII series bodies. Last time I upgraded my MkIV adapter (about when the M4/3 v1.8 firmware was released) the AF results with my stock of lenses was still quite variable whilst my Panasonic bodies were working very well right across the board with only the perennial bad boys: the 135/2.0 and 50/1.8 II being not up to scratch (and probably a few others outside my ken).

So well does M4/3 work with Metabones adapters to Canon EF that I just cannot be bothered wrestling with the Sony A7R interface and EF lenses where some work acceptably in AF and others not so well. This would not be so bad if the A7R were a nicer camera to use for manual focus.

With M4/3 I can get fast accurate AF with touch up MF if desired. This is adapting lenses in luxury ..... :)
 
Yes, this v50 thing is the big improvement to AF on A7R Mark I. The performance is similar to or slightly below Panasonic now.
 
I would suggest using DMF + focus peaking on Sony to deal with the misfocus issue - you can tell right away if it mis-focused or not. It is an issue with native lenses too, but adapted lens clearly exacerbated the issue.

E-M1 now tries to achieve higher PDAF speed by a learning algorithm, the same as the one used for A7RII. It takes a couple of locks to train the adapter. You may help the learning process by starting with easy targets with no drastic distance changes. But all in all, it sounds like the learning function is too aggressive and backfired, so some fine tuning on the part of the firmware might be necessary in the next version.

--
Bo-Ming / conurus
 
Last edited:
cyberstudio said:
I would suggest using DMF + focus peaking on Sony to deal with the misfocus issue - you can tell right away if it mis-focused or not. It is an issue with native lenses too, but adapted lens clearly exacerbated the issue.

E-M1 now tries to achieve higher PDAF speed by a learning algorithm, the same as the one used for A7RII. It takes a couple of locks to train the adapter. You may help the learning process by starting with easy targets with no drastic distance changes. But all in all, it sounds like the learning function is too aggressive and backfired, so some fine tuning on the part of the firmware might be necessary in the next version.
Click to expand...
 
cyberstudio said:
Yes, this v50 thing is the big improvement to AF on A7R Mark I. The performance is similar to or slightly below Panasonic now.
Click to expand...
Hooray!

I might even nearly start to like my A7R (grin).

Like all things, put it aside for long enough and it will become useful again :)

One of the best things I never did was probably not disposing of my Canon EF lenses when I wandered off to other brand camera bodies. One of the worst things I probably did was going overboard investing in them in the first place.

I relished Ricoh then the prospective ability to re-use my Canon EF lenses and retire the 5D pushed me to Sony then the compact size of the GM1 as a systems camera made M4/3 attractive. The ability to use EF lenses to acceptable levels of oem perfomance made the M4/3 system replete and now if I can use my A7R without being literally thrown against its cranky interface then I will be doubly happy.

But now in retrospect I would not ever be able to afford (mentally or otherwise) such a significant investment in high quality lenses. But at the time I thought dslr camera bodies were "forever". Nothing is forever.
 
ahaslett said:
cyberstudio said:
I would suggest using DMF + focus peaking on Sony to deal with the misfocus issue - you can tell right away if it mis-focused or not. It is an issue with native lenses too, but adapted lens clearly exacerbated the issue.

E-M1 now tries to achieve higher PDAF speed by a learning algorithm, the same as the one used for A7RII. It takes a couple of locks to train the adapter. You may help the learning process by starting with easy targets with no drastic distance changes. But all in all, it sounds like the learning function is too aggressive and backfired, so some fine tuning on the part of the firmware might be necessary in the next version.
Click to expand...
Click to expand...
 
Last night I upgraded my EF-FE MkIV to v0.50 and my EF-M4/3 Ultra and Smart Adapters to v2.3.

It was dark and indoors with available light so not a real test.

I only used a limited number of lenses simply because they were handy:

The (cheap) Yongnuo EF 35/2.0, the Canon EF 100/2.8L Macro, and the Sigma 4.5/2.8 Circular Fisheye.

This must be the oddest test bunch in existence:

The Youngnuo build looks like a Canon 50/1.8 II clone and acts like one - I could not get it to register wider than f5.6 and it did the usual crazy cat focus attempts before giving up - much the same as expected - 100% fail. It must have cloned 50/1.8 II firmware inside it.

The Canon I thought to be one of the most compatible of the EF lenses to be adapted. But it was not working that well on the A7R at all - fudging all over the place but sometimes locking on if there was enough light. It did work a little better when I changed from pinpoint to single area focus. I will have to work on it some more but at present version 5.0 firmware is not looking so good. But I am not giving up until I have tried some more lenses and in better light. The same lens on the same subjects with the Panasonic GX7 was focusing well, the mind works tricks, but maybe a tad slower than the version 1.8 of the firmware I was using before (acceptable nevertheless and it simply showed up just how bad the A7R really was). On the other hand the Olympus E-M1 has turned into a speed demon. Focusing so fast and accurately that I hardly noticed it doing it!

The Sigma is a special case because it has such a deep dof. But as far as I could see it was behaving much the same as the Canon lens.

I will see if I can manage a little wider experimentation today in proper daylight.

(Update: Yikes! The auto focus capability of the E-M1-EF 100/2.8 combination in daylight is astounding! Bo-Ming: you are a genius)

[Edit] I have now tried the A7R with the EF 100mm f2.8L Macro with Metabones IV adapter - v5.0 firmware in daylight - indoors & available light. I am happy to advise that the combination is now focusing very quickly and accurately - I am happy! For the first time since I bought that body it seems to working the way it should. Now to get used to that awkward interface .... I don't know what was happening last night but it seems that it's learning algorithm does not do much learning in low light. Outside - more distant objects - having some trouble with accuracy on foliage but hard objects a bit better - I will have to investigate this more.

--
Tom Caldwell
 
Last edited:
Bo, it's nice to see you are still involved in the project that you started way back when Contax N lenses were first being adapted to Canon.

I hope the partnership with Metebones has been prosperous for you for the magic you've done.

--
I don't have any AF lenses, so if I want a picture, I have to do more than squeeze a button.
I just bough my first camera, Best camera EVER!!!1
Interested in Adapting lenses? head here: http://www.dpreview.com/forums/1065
My shots:
http://www.flickr.com/photos/lightshow-photography/
My lenses:
http://www.fredmiranda.com/forum/viewprofile.php?Action=viewprofile&username=LightShow
You can't depend on your eyes when your imagination is out of focus.
-Mark Twain
Reputation is the shadow. Character is the tree
-Abraham Lincoln
####Where's my FF NEX-7 ?????
 
Last edited:
Nothing is meant to be easy and the A7R is going to work but seems to need some thoughtful use.

The "difficult" EF 135mm f2.0 now seems to be working at quite acceptable auto focus speeds but takes some fits of its former self from time to time.

The accuracy seems to vary. At flexible spot (small) it is quick but not so accurate, but it might be better in some situations than others - I can also try medium and large - so many lenses/combinations .... The "center" focus point is more accurate at close defined objects but on the fuzzy side for two handy but reasonably distant wallabies.

I think that when I work the changes and get some normal type activity I will be better informed.

In the meantime it is the Olympus E-M1 that has become the star turn - it is Purrfect with the once cranky Canon EF 135mm f2.0L lens. Spot on quick AF that does not have to be "managed" like that of the A7R. Right at the moment in the midst of my experimenting phase the images form the lowly 4/3 sensor would blow those from the FF sensor A7R away. (Sorry guys but I hope I can do better if I work harder at it :) ).

The same lens tried on the Panasonic GX7 continues its multiple re-tries habit, but when it does lock properly it is tack sharp. This lens is now a standout on the E-M1 body.

The Canon EF 70-200 f2.8L OIS.

Sony A7R Metabones IV v0.50 firmware - acceptable AF speed and accuracy - I would not call it as tack sharp accurate as might be expected of a camera with a FF sensor, but there is tweaking time available. But the fact that tweaking time might be necessary is the worry.

Panasonic GX7 Metabones Smart Adapter v2.3 firmware. Needs retries mostly but then is tack sharp. Worthy for use, but that once only chance for a prize winning shot of a moving subject might be lost.

Olympus E-M1 Metabones Ultra FR Adapter v2.3 firmware - fast accurate (oem level?) AF tack sharp. The easy winner.

Just for completeness the Sigma 4.5mm f2.8 circular fish-eye works fine on the A7R but the vignette is huge with the image a little circle in the middle of a vast sea of black - it is designed for the aps-c image circle and actually seems to perform at its best with a Smart Adapter in 1:1 crop on a M4/3 camera. Just the right amount of 100% frame on a circle image. Better looking than its native aps-c in fact.

I must away and do something useful with my life but I expect the Metabones new firmware and its performance on various lenses is going to be the usual story of variations of success. For the moment the v2.3 firmware has made the E-M1 the standout performer and maybe v0.51 firmware might iron out the precision of the AF accuracy on the A7R at least - my observation seems to be that it is just a litle on the fuzzy side of perfect. Funnily enough even MF does not seem to improve it - it might be a factor of the default extreme magnification of playback - in which case it is my mea culpa.

Don't know really but the E-M1 has improved very obviously and if anything the GX7 is not quite as good as it was with the earlier firmware.

Images captured are too experimental as of yet to be worth showing - and of course I have already forgotten which lens took which shot apart form the very obvious ones - maybe the EXIF would help ....

I have not tried "green" mode.
 
All images with Sony A7R MkI with Metabones EF-FE adapter MkIV v5.0 firmware Canon EF 100mm f2.8L Macro OIS Lens (ooc jpg)

Here I had to MF first then AF as the AF had trouble locking on otherwise
Here I had to MF first then AF as the AF had trouble locking on otherwise

Back to front - usually the idea is to AF then finite adjust by MF



AF only - difficult light - lots of water reflection - maybe a bit dark, but highlights blown otherwise
AF only - difficult light - lots of water reflection - maybe a bit dark, but highlights blown otherwise



AF worked fine - first hit
AF worked fine - first hit



AF worked fine - small crop to adjust horizon.
AF worked fine - small crop to adjust horizon.

I need to test a few more lenses to be better assured.

The images posted here are just to demonstrate that "sharp" can be had with AF and that when it did work the speed was reasonably acceptable. But it was not completely reliable.

Occasionally some help was needed to nudge AF by first doing a quick MF as most Canon lenses always allow a MF override. At times the lens hunted and some shots, mainly with other lenses were caught after AF but were not in focus. This particular lens was reasonably happy. The adapter/camera is working better together than the last time I tried it (v4.3) but it is still a lot more work than the Metabones EF to M4/3 adapters. They work very well with Panasonic bodies but near enough to perfect with the Olympus E-M1.

--
Tom Caldwell
 
Tom Caldwell said:
Last night I upgraded my EF-FE MkIV to v0.50 and my EF-M4/3 Ultra and Smart Adapters to v2.3.

It was dark and indoors with available light so not a real test.

I only used a limited number of lenses simply because they were handy:

The (cheap) Yongnuo EF 35/2.0, the Canon EF 100/2.8L Macro, and the Sigma 4.5/2.8 Circular Fisheye.

This must be the oddest test bunch in existence:

The Youngnuo build looks like a Canon 50/1.8 II clone and acts like one - I could not get it to register wider than f5.6 and it did the usual crazy cat focus attempts before giving up - much the same as expected - 100% fail. It must have cloned 50/1.8 II firmware inside it.

The Canon I thought to be one of the most compatible of the EF lenses to be adapted. But it was not working that well on the A7R at all - fudging all over the place but sometimes locking on if there was enough light. It did work a little better when I changed from pinpoint to single area focus. I will have to work on it some more but at present version 5.0 firmware is not looking so good. But I am not giving up until I have tried some more lenses and in better light. The same lens on the same subjects with the Panasonic GX7 was focusing well, the mind works tricks, but maybe a tad slower than the version 1.8 of the firmware I was using before (acceptable nevertheless and it simply showed up just how bad the A7R really was). On the other hand the Olympus E-M1 has turned into a speed demon. Focusing so fast and accurately that I hardly noticed it doing it!

The Sigma is a special case because it has such a deep dof. But as far as I could see it was behaving much the same as the Canon lens.

I will see if I can manage a little wider experimentation today in proper daylight.

(Update: Yikes! The auto focus capability of the E-M1-EF 100/2.8 combination in daylight is astounding! Bo-Ming: you are a genius)

[Edit] I have now tried the A7R with the EF 100mm f2.8L Macro with Metabones IV adapter - v5.0 firmware in daylight - indoors & available light. I am happy to advise that the combination is now focusing very quickly and accurately - I am happy! For the first time since I bought that body it seems to working the way it should. Now to get used to that awkward interface .... I don't know what was happening last night but it seems that it's learning algorithm does not do much learning in low light. Outside - more distant objects - having some trouble with accuracy on foliage but hard objects a bit better - I will have to investigate this more.
Click to expand...
 
Brian

Don't know - I thought I was on the latest at v2.3 and now v2.4 is available. I was tardily hanging about on v1.8 so the performance improvement from there to v2.3 was immediately noticeable. Gone from good to really truly good on the E-M1 but I thought that the GX7 performance had regressed if anything, but the mind can play tricks.

I only have the E-M1 with its PDAF and cannot check the Olympus CDAF used by other bodies.

I found that the Canon EF 100mm f2.8L Macro was startling in how well it is working on the E-M1 and it puts the improved performance of the Sony AR7 MkI to some serious shame. If the E-M1 was not available for direct comparison I would be quite happy with the A7R MkI despite its stumbles. Occasional hunting, occasional refusal to focus, when it "achieves focus" it is sometimes not-sharp. Overall 50%+ of the images caught had some sort of focus stumble. But my success rate was improving - I don't know if this was because I was going through my own learning process - the adpater was "learning" as Bo-Ming suggested - or my tricks of using flexible spot (small) (I don't think so) - or my resolution of a hunting problem using MF first by the focus ring and clutch as Canon lenses allow, helped. Note that in the latter case the camera remained in AF mode and the idea was to approximate focus manually then try AF - which worked splendidly and could be repeated. The Yongnuo EF lens is built after the style of the Canon EF 50/1.8 II and has a tiny focus ring and no clutch - it proved to be a complete basket case. Considering that I have so many EF lenses it is a wonder that the Yongnuo was even tested - it just happened to be handy .....

Our lens stocks don't synchronise and I have not really had a chance to try my other lenses to see how they perform. The 135mm f2.0 has improved to be useful on the E-M1 and the Yongnuo 35mm f2.0 was and remains the recalcitrant that proves that the rest are "pretty good". Sigma lenses seem to work fine. All these are basically M4/3 ruminations. I think that the A7R MkI at v0.50 on the MkIV adapter is full of promise, but just a little rough at the edges yet. A significant coup for Bo-Ming for a system already fading into obsolescence oblivion since the MkII arrival. I think it might take a v5.1 or 5.2 to smooth over the rough edges of v5.0 at which point just maybe the MkI series will start to provide a reasonable shadow of the performance of EF lenses that the M4/3 guys were given by v1.8 of that firmware.
 
I was pretty happy with v2.2 on my EM1 and Bo-Ming seemed concerned about how my 300mm f/4 worked on v2.3. Maybe v2.4 has been tweaked. Hard to believe that it can also do CAF.

The performance on the A7R is miraculous, especially as AF on the A7R seems to be a best-efforts function with native lenses. Immediately bought a used 35mm f/2 before prices in GBP start to drift upwards.

Andrew
 
ahaslett said:
I was pretty happy with v2.2 on my EM1 and Bo-Ming seemed concerned about how my 300mm f/4 worked on v2.3. Maybe v2.4 has been tweaked. Hard to believe that it can also do CAF.

The performance on the A7R is miraculous, especially as AF on the A7R seems to be a best-efforts function with native lenses. Immediately bought a used 35mm f/2 before prices in GBP start to drift upwards.

Andrew
Click to expand...
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top