So? To the software Engineers?

[Guest Chris M]

Advertisement (gone after registration)

Now that the experienced software engineers (HERE) on this forum have spoken in the past about the X1's capability's, ..WHAT can realistically be added or done to The X1's First FW update, in it (current) status? Please only knowledgeable people need reply:)

 

thanks, CM

[Guest badbob]

Since the X1 is inherently slower than its nearest rivals, the only way to make it faster is to tweak the software algorithms for faster performance. There are a few competent software engineers out there who've worked with these things, who could improve the performance some. But you probably won't get a satisfactory answer because the actual code is under lock and key, and hackers who can trace the execution of the firmware are *way* more specialized than the software engineers, and they aren't looking at X1's. Unless you get one of those one-in-a-million hackers who just happens to be an expert in camera software, and has a thing for the X1. Then that would get tricky, because decompiling the code and talking about how it works is probably illegal.

[Guest badbob]

Just adding another comment to the above. If I had the actual code, I'll take a wild guess that it's some "object-oriented" language like C++, but highly developed with many levels of structure. Meaning fat and slow. So I doubt anyone would want to go in and try to tweak any of the individual components, meaning they would have to get lucky and find replacements that perform better, for some of the higher-level component objects. Still not an optimistic prognosis.

[pico]

My experience from the old days of small architectures, coding in assembler with inline machine code to obviate bottle-necks was the way to go. But today's X1 probably has 100 times the power of our whole machine room (two PDP/11-70 and a disk farm of maybe 500mb). Man, it's almost hard to believe.

 

I'm so old that my donor card's "Use By" date has expired.

[h00ligan]

LOL pico, that was great. On the upshot, old never means irrelevant - it means you have a lot to teach.

[RichardM8]

 

Be careful Edward, this thread is for knowledgeable people only.

 

It assumes the Leica had some cheese-makers on their payroll working on the X1 and now we're looking for new real programmers to do it right...

[h00ligan]

Advertisement (gone after registration)

oops, soooorry!

[grober]

My experience from the old days of small architectures, coding in assembler with inline machine code to obviate bottle-necks was the way to go. But today's X1 probably has 100 times the power of our whole machine room (two PDP/11-70 and a disk farm of maybe 500mb). Man, it's almost hard to believe.

 

I'm so old that my donor card's "Use By" date has expired.

 

DEC-ies of the world Unite! Throw off the scam called "Windows" and return to those golden days when we all drank Ken Olsen's koolaid. I was there too, bro . . . and am still working as a result!

 

BTW: I miss my M7!

 

-g

[RichardM8]

This thread is gonna be hilarious...

[Guest Chris M]

Now that the experienced software engineers (HERE) on this forum have spoken in the past about the X1's capability's, ..WHAT can realistically be added or done to The X1's First FW update, in it (current) status? Please only knowledgeable people need reply:)

 

thanks, CM

 

RichardX1, I did not ask about auto focus speed being increased, please stay on topic, thanks:)

 

chris m.

[h00ligan]

In terms of that which first party developers can add, I see no reason they couldn't enhance the manual focus design to automate scale switching.

[Guest badbob]

My experience from the old days of small architectures, coding in assembler with inline machine code to obviate bottle-necks was the way to go. But today's X1 probably has 100 times the power of our whole machine room (two PDP/11-70 and a disk farm of maybe 500mb).

 

Yes, the X1 has 100 times the power, but 20 times slower software, and 20 times less capable I/O throughput. Multiplying 20 times 20 gives 400 times degradation compared to your example (very approximate). The sad thing is that I/O is a very big problem with small devices.

 

In 1979 I watched a payroll laser printer at Firestone's main computer center print pages so quickly that you could grab one end and run fast and not outrun the printer. Today's fastest for a small office is still several times slower than that 1979 printer.

 

Same deal with media - I have a 1TB (1000 GB) pocket disk drive now, which has a sustained best-case write speed of 20mb per second. So if you want performance, you don't buy the latest small device - you go looking for the much more expensive much larger device.

[h00ligan]

Part of that has to do with affordability. Has memjet hit the scene yet? It's small, very fast, but I Haven't heard of it being picked up...

 

World’s Fastest Color Printer - 60 Pages Per Minute (VIDEO) - published more than 3 years ago.

 

So the thing is then, should the leica x1 at it's price point have better IO? Can it be achieved with the technology used. Do we feel that the software engineering is the limitation?

[pico]

DEC-ies of the world Unite! Throw off the scam called "Windows" and return to those golden days when we all drank Ken Olsen's koolaid. I was there too, bro . . . and am still working as a result!

 

BTW: I miss my M7!

 

-g

 

RT-11 forever! And ever, and ever. VMS, OpenVMS. Hell, I learned WindoZe internals because earlier they were kinda like VMS. Not kinda enough, though.

 

WindoZe changed my life. No, it wasn't a sex change operation. It just FELT that way. Without anesthetics!

[Guest badbob]

So the thing is then, should the leica x1 at it's price point have better IO? Can it be achieved with the technology used. Do we feel that the software engineering is the limitation?

 

My feeling is that the big camera co's like Canon and Nikon are under pressure with their big items to optimize the throughput, so I'm sure they tweak the software as well as the hardware. And given their margins and units sold, they can afford to. Leica I presume doesn't do that, unless they do some tweaking for the S2 and possibly in some limited cases for the M9. Unlikely they'll do any tweaking of the X1.

 

If there's a new model of the X1 and it doesn't have any greater I/O demands than the X1, then they might justify a software improvement as part of the new camera project. If the I/O demands are greater, then they will have to upgrade the software just to keep performance the same (no net gain).

 

Wouldn't it be something if Leica were known someday for their superior camera firmware as well as hardware? The programmers who can do it are out there, but you have to go get them - they won't walk in the front door and apply.

[davidhunternyc]

Terrific discussion guys. I don't know anything about software engineering and so I find this fascinating. Hopefully Leica will be tuning in here too.

[h00ligan]

Yes Bob, I agree - it would be wonderful! They way the industry seems to be moving it will be necessary too. What I think I don't understand is if they hire the best engineers in lens /camera design - why are they skimping on the electronics. I think partially it is because they have been able to rely on Panasonic for the majority of that work until recently - at least as you said for the majority of the leica branded cameras sold.

 

MY personal thought is the largest issue with the camera is the buffer/controller they chose, that will really be the hard line in the sand so to speak. IO increases would be welcome - and maybe it is an issue of the engineers at Leica simply saying - hey, this throughput is good enough, not being concerned or aware of overall impact.

 

I would love it for someone to do a tear down so we could see which controller they chose and other such technical elements. It would answer a lot of questions, namely what can be optimized with firmware.

 

For now I'll stay positive that it is a result of being green and they will be able to further optimize the tools the create.

 

That said I really don't understand why they couldn't have created a swapping scaled focus mechanism so when you hit a designed marker in the master manual scale mode you can pass directly to a more detailed view of the next scale, or allow it to be selectable through a certain button depress under certain conditions. Or more puzzlingly, why they didn't add a manual focus ring around the lens.. perhaps because it would have to be focus by wire and they don't have any experience in that realm?

[Guest Chris M]

Ed, great response!. I too know that budget constraints as a whole with Leica are very tight. They seem to be on track now with where they stand (Financially) as a company. Hopefully, when Leica produces future camera's they will understand that they need to be able to offer little surprises, ie; FW Updates for the price point that they (LEICA) sells there camera's for. Allways leave a little wiggle room for a (great fix) or option. thats good businesses practice if you don't quite get it write the first time around. IMHO

 

cm

[Nicoleica]

DEC-ies of the world Unite! Throw off the scam called "Windows" and return to those golden days when we all drank Ken Olsen's koolaid. I was there too, bro . . . and am still working as a result!

 

BTW: I miss my M7!

 

-g

 

Oh golly. I must be more ancient that any of you. It was punchcards and printer loops when I started programming. Not a VDU in sight, hand-punches and card sorters. Still, we did write compact and efficient code, and we would always have an endless supply of confetti.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[RichardM8]

The amount of apparent insight, knowledge and assumptions here is staggering.

 

Maybe I look at this the wrong way but without any factual knowledge about what, how and why the X1 is like it is today, it's 'kinda' difficult to draw any valuable conclusions. Regardless if you have coded the calculations for the Roman Colosseum or not.

 

The general consensus here seems to be/become that Leica doesn't have (enough) skilled software engineers. Yet, the current 3 Solms models (X1, M9, S2) rank among the highest or are #1 for IQ in their respective segments. This is not just a glass/sensor thing. Image processing is a decisive factor for a cameras' output. This is the work of Leicas' software engineers.

 

As for the X1; the Ricoh GXR A12 is probably the closest technical relative to the X1 at present (97% sure the same sensor, prime lens compact with contrast AF). The X1 beats the A12 by a large margin on IQ, AF speed, AF accuracy and technical file quality. IMO apples vs. apples gives more insight than apples vs. oranges...

 

 

 

But this belies the fact that there are some obvious flaws and omissions in the current X1 firmware. And most of them are rightfully mentioned here too. The possibilty to shoot DNG only is about the most obvious one. On my wish list are more selectable values in the ISO menu scale like 320, 640, 1000 etc. The camera knows them in auto ISO mode so they can be put into the user menu too.

 

But this is nothing unique. It took Canons' massive software department many years and 3 or 4 entire generations of their SLRs to get the selected ISO in the viewfinder...

 

My two cents on the subject.

 

Richard.