Is the next OMD going to be an EM1 IV, not an EM1X II?

Michael Meissner

Mu-43 Top Veteran
Joined
Sep 19, 2018
Messages
843
Location
Ayer, Massachusetts, USA
No need for dual processors. Just introduce a new one with enough power and stop recycling old tech. Just look at the processors in mobile devices. They're making all the camera CPUs embarrassed.
And BTW, all of the processors in cell phones are dual/quad processors.

However, changing microprocessors can be problematical, particularly if there is a lot of assembly code specific to one line of microprocessors that would have to be rewritten. If the new processor is upwards compatible with the old processor, you can just drop in the code or do minor tweaks. With a lot of software, the people who wrote the original code may no longer be with the company (or remember the fine details of something they wrote years ago). And often the software teams are understaffed, so they may not have the time to rewrite everything.

Changing microprocessors can also mean having to retrain the staff to use the new microprocessor. My day job is supporting the GCC compiler for companies that make chips. Back in the day when I worked at Cygnus Solutions and we supported many different architectures, it would take me 1-2 months to fully come up to speed on the new microprocessor instruction set. It might take people who have only used one microprocessor in their professional work more time to learn the new architecture.

The tools used (compiler, debugger, etc.) may not be available for the new chip. Even if there are replacements, things aren't always as compatible as they could be.

There are possibly other things that affect the choice of micro-processors, including chip availability, what form factor the chip is in, what agreements the parent companies of the camera and microprocessor maker have done that controls the microprocessor.

Right now, chip availibility is an extremely volatile topic. Remember, the company making the camera has to be able to buy the chip in enough quanity to make the current products, and to be able to buy more in the future.
 
Last edited:

PakkyT

Mu-43 Hall of Famer
Joined
Jun 20, 2015
Messages
4,576
Location
Massachusetts, USA
BTW, all of the processors in cell phones are dual/quad processors.
To nitpick, one processor (CPU) but with multiple cores. The E-M1X uses two CPUs but I have no idea of those Truepix image processors are multiple core or just old fashioned one core types.


particularly if there is a lot of assembly code specific to one line of microprocessors that would have to be rewritten
No one writes code in assembly anymore. :laugh1: But yes new processor does mean new or updated code.


Changing microprocessors can also mean having to retrain the staff to use the new microprocessor.
Software engineers and programmers should be able to transition to new processors without a lot of "training" as this is what they are trained to do.


The tools used (compiler, debugger, etc.) may not be available for the new chip.
That is just silly talk. No CPU manufacturer in the world would ever release a chip than no one could actually use. All tools required are always developed with the new chips and are always available to developers sometimes even before the final versions of a new process are released.
 

Michael Meissner

Mu-43 Top Veteran
Joined
Sep 19, 2018
Messages
843
Location
Ayer, Massachusetts, USA
No one writes code in assembly anymore. :laugh1: But yes new processor does mean new or updated code.

You don't write complete programs in assembly much these days, but often times specific libraries are written in assembly to take advantage of specific CPU features like new instructions that the compiler doesn't provide support for. In addition, there is the use of extended asm to provide support to access specific instructions.

Software engineers and programmers should be able to transition to new processors without a lot of "training" as this is what they are trained to do.

Not necessarily. Remember, if the company makes the change, every member of the team has to make the change. There are likely some members of the team that just don't handle change well. You have both 'senior' programmers that are set in their ways, and new hires that would need to learn. Yes many people can learn, but it takes time. Often times when management makes schedules, there is no time in the schedule to learn new processors or even upgrade stuff.

That is just silly talk. No CPU manufacturer in the world would ever release a chip than no one could actually use. All tools required are always developed with the new chips and are always available to developers sometimes even before the final versions of a new process are released.
I meant the particular company providing the compiler may have either folded, or decided not to produce a compiler for the new chip. Big code can get tuned for specific compilers, and it can be a mess to change it. Often times, people may not realize that it is tuned for a specific compiler until they have to use something else.

As a senior GCC developer I have spent something like 36 years providing such toolchains at several different companies, so I am well aware of the amount of work involved. I and my team members have done some ports to completely new processors. More often, we have upgraded the tools to provide for new versions of the processor. In fact, IBM just formally announced the power10 that I've worked on providing the support for in the last 3 years. But the reality is the toolchain may or may not be available when the first samples are available for testing (and when a company might be thinking about moving to). And maybe GCC is available, but if you are used to a different compiler, you may have to wait for the tools to be available.
 
Joined
Aug 13, 2014
Messages
4,264
Location
Honolulu, HI
Real Name
Walter
So you are saying that the processor in the E-M1 III is old? Not any more powerful than the one in the mark II? Okay.

The EM1 III has one TruePic™ IX Quad Core Processors (copied from the webpage with an s on the end), the latest one. Nothing old about it. The EM1X has two TruePic™ VIII Dual Quad Core Processors. The EM1 II has one VIII.
 
Joined
Dec 6, 2015
Messages
1,337
Location
Brisbane, Australia
Real Name
Angus
And you guys bitch about Panny's naming conventions.

Good grief.

Microchip shortage. Sony is having problems too. A lot of things are being delayed.
New cars for example, used car prices are rising here in Australia because the chip shortage is flowing onto the new car market.
 
Joined
Dec 6, 2015
Messages
1,337
Location
Brisbane, Australia
Real Name
Angus
I love the E-M1X economics, and I bought a second one, because shooting with an E-M1 mk II(I) with grip just isn't the same. And then Olympus released the intelligent AF bird mode and the second body felt even more justified.

I don't really care which form factor the next body is in, but if it's not a plus size body, then I hope they release a revised grip with a joystick.

However, I walk about 7-8kms most days with two E-M1X bodies in my bag (My place backs onto a creek and trail with a lot of bird life). If I could shed a bit of weight by dropping the vertical grip while still keeping bird intelligent AF, that would be a win too.
 
Joined
Dec 6, 2015
Messages
1,337
Location
Brisbane, Australia
Real Name
Angus
Ah that answers my question about if the TruePic are multi core and it appears they are. So the E-M1X is working with 8 cores while the Mk3 has a newer processor it has 4 cores.
It's also why the E-M1X has dual card UHS-II card slots while the mk III doesn't.

I also believe why the intelligent AF modes didn't come to the mk III, TruePic IX is superior but one's still not fast enough.
 

fsi22

Mu-43 Regular
Joined
Oct 25, 2014
Messages
28
Ya and I just remembered that I have my first grandchild coming early next year so considering I am likely to utilize the video feature a lot more, I am in hindsight kicking myself for not jumping on that 4k capable E-M1.2 (less than 5k shutter activations and 3 OEM batteries included) when it was posted Saturday. I will dearly miss the Tilt LCD screen of the Mark 1 though. For what I typically shoot tilt is much more useful to me than articulated. But practically everything else on the Mk2 is an upgrade.

Edited to add: Just looked at the GetOlympus web site in the refurb store. Again, the new OMDS is supposed to be a leaner company able to respond quicker without the overhead of the Olympus corporation, yet this is still showing up on the website...

Get 20% off with code PHOTODAY20 through 11:59:59 PM ET on August 20. Only $799.99 with coupon. Save $200.​

Red text emphasis mine. Kind of like my comment above about USB 2 this is another embarrassing bush league error that kind makes you question this new company and if they are competent to continue the imaging brand.
A suggestion, from someone who has a 17 month old running around. Get the m1 iii, it's eye af is insanely good. I've owned the ii, iii and x. iii is leagues ahead and it works exceptional in video
_B213308.jpg
Subscribe to see EXIF info for this image (if available)

Son was throwing his head back. No problem for iii
_B080036.jpg
Subscribe to see EXIF info for this image (if available)


Toy in front, iii instantly locks on his eye and never gets confused


_B140596.jpg
Subscribe to see EXIF info for this image (if available)

This blew me away, we were at the beach, i was falling backwards and the eye af still grabbed him. That's uncropped so would've been pretty close to minimum focus distance as well.

Oh and I always use my 25 f1.2 @ 1.2 when taking pictures of my son. The eye af is seriously next level. The iii is such an underrated camera. eye af on the ii was awful, The X is too slow and not accurate enough.
 
Last edited:

Darmok N Jalad

Temba, his aperture wide
Joined
Sep 6, 2019
Messages
1,997
Location
Tanagra (not really)
In regard to processors, many times the updates are iterative, where they can either get more frequency through a new process revision, consume less power, or even both. Sometimes it’s just more cores, more cache, or more memory—all of these provide performance gains without much, if any, change to execution. If you look at the progression of x86 over the decades, software support didn’t die with the next generation of CPUs. While new instruction sets can be added and leveraged, even for old unchanged code, the CPU simply does the same job faster than before—it’s how AMD and Intel have designed things for a long time.

I would venture a guess that camera coders are like game console developers, they are working within fixed hardware, seeking ways to make the software perform the best it can within the limits of the camera hardware, which isn’t changing until the next generation. Look at how much Panasonic’s software engineers have improved the G9 from when it launched. V2.0 was a big leap, and v2.4 was another. Besides, the camera companies are not just pulling a random SOC off the shelf and cramming it into the camera, they are specifying what they want and getting it manufactured. They have to set a design target that hits acceptable manufacturing yields while also getting it to work stably inside a passively cooled chassis that is exposed to extreme environmental conditions. Considering the age of current hardware, the next generation of cameras should have quite a bit more performance to work with.
 

cedge

Mu-43 Regular
Joined
Mar 31, 2012
Messages
101
Location
Dorset, UK
My next camera is going to be a
EM1 mkii.
This is not a rumour, it is a fact.
I am always late to the party, but I will get there.
I've actually gone back to shooting a lot with my EM1.1, about to sell my Pen-F.
 

Michael Meissner

Mu-43 Top Veteran
Joined
Sep 19, 2018
Messages
843
Location
Ayer, Massachusetts, USA
In regard to processors, many times the updates are iterative, where they can either get more frequency through a new process revision, consume less power, or even both. Sometimes it’s just more cores, more cache, or more memory—all of these provide performance gains without much, if any, change to execution. If you look at the progression of x86 over the decades, software support didn’t die with the next generation of CPUs. While new instruction sets can be added and leveraged, even for old unchanged code, the CPU simply does the same job faster than before—it’s how AMD and Intel have designed things for a long time.
Though unfortunately at times specific instructions are slower or faster going from one processor to another (either another processor from the same company or a competitor like AMD). Trying to get the compiler to do all of these micro-optimizations can be rather tricky.

As an example, in many of the Intel processors, the REP; MOVS instruction pair has extreme optimization built-in and the compiler generating code to copy a series of bytes would just use REP; MOVS. On other platofrms, like AMD, there isn't the extreme optimization built-in, and the compiler will generate a series of moves and such that are faster in total execution than the REP; MOVS.

Of course that extreme optimization isn't for free in the Intel side -- it needs significant space for the microcode support and perhaps significant building block micro-ops to implement this. This means that less of that microcode space and/or silicon die area can be used for something else.

Basically everything is a trade-off.

And unfortunately, at times the chip designers may not even realize there is a slowdown until we run the benchmarks, and trace things at the micro level. And usually by that time, the essential design for the machine is frozen, and you have to live with it.

I would venture a guess that camera coders are like game console developers, they are working within fixed hardware, seeking ways to make the software perform the best it can within the limits of the camera hardware, which isn’t changing until the next generation. Look at how much Panasonic’s software engineers have improved the G9 from when it launched. V2.0 was a big leap, and v2.4 was another. Besides, the camera companies are not just pulling a random SOC off the shelf and cramming it into the camera, they are specifying what they want and getting it manufactured. They have to set a design target that hits acceptable manufacturing yields while also getting it to work stably inside a passively cooled chassis that is exposed to extreme environmental conditions. Considering the age of current hardware, the next generation of cameras should have quite a bit more performance to work with.

The GH5 mark II release said essentially that the processor used in the GH5 mark I was out of steam, and they needed to use the newer processor from the G9 (and presumably S line) to be able to support future firmware enhancements.
 

pake

Mu-43 Hall of Famer
Joined
Oct 14, 2010
Messages
2,562
Location
Finland
Real Name
Teemu
So you are saying that the processor in the E-M1 III is old? Not any more powerful than the one in the mark II? Okay.
I'm saying it didn't bring much to the table and OMDS CPUs need a major upgrade. Not something that only allows you to replace two chips with one. We need something that gives us class-leading features (such as animal eye AF, faster HHHR etc.).
 

Latest threads

Links on this page may be to our affiliates. Sales through affiliate links may benefit this site.
Mu-43 is a fan site and not associated with Olympus, Panasonic, or other manufacturers mentioned on this site.
Forum post reactions by Twemoji: https://github.com/twitter/twemoji
Forum GIFs powered by GIPHY: https://giphy.com/
Copyright © Amin Forums, LLC
Top Bottom