Will IR cut filters AND coding be necessary on longer lenses?

[jaapv]

Advertisement (gone after registration)

I just purchased an rd-1. I have 22 leica m lenses and have been patiently waiting to see how that 1936 uncoated summaron looks on an M8. The problem is I do not want to spend another 1500.00 on filters (e39 e48 e49 series v vi vii whatever....) I bought and sold the dmr Franken camera and I am currently enjoying the old visoflex technique with r lenses on a canon 5d. I can wait for leica to get it right. For now I look forward to the M9 and the arrival of my rd-1.

Let me get this straight: you have 22 lenses of a value between twenty and sixty thousand dollars depending on which ones. You are talking about a five thousand dollar camera body and fifteen hundred dollars is a deal-breaker ? Sorry....

Btw I'm genuinely curious: what is a Visoflex technique on a 5D?

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[ho_co]

Joseph W--

You've got a lot of information new to me in your short Mac history at http://www.leica-camera-user.com/digital-forum/10086-will-ir-cut-filters-coding-necessary-2.html#post102988.

 

They certainly had underperforming computers compared to their clones--that much I was aware of.

 

Where can I find more info on the topic? I was unaware that Xerox had sued Apple, or that IBM offered a MacOS option; and although I've worked a little with OS/2, I had no idea that it was related to Windows NT. (In fact, NPR built their DACS stations-information service on OS/2, because--I thought--it has some kind of multitasking that wasn't available elsewhere.)

 

I find this fascinating. Can you list any sources?

 

Many thanks.

 

--HC

[sean_reid]

Sean:

 

Whether the 6-bit coding is patented would be a good question to ask your sources at Leica. If it isn't patented, the ball is now in Zeiss and Cosina's court to suggest a solution.

 

I wouldn't consider the Zeiss lenses off brand. Have you seen the price of their 15mm and 85mm lenses? Their performance is also very Leica like and Zeiss has supplied lenses to Leica in the past to be part of the Leica lens line up in both the M and the R.

 

In regards to the Cosina lenses, they could probably be considered an off brand as they typically perform just behind the equivalent Leica lens. Similar to how the Sigma or Tamron lenses do not typically perform as well as Canon L or Nikon ED lenses. There are some exceptions, such as some of the Nikon branded ED zooms made by third parties.

 

The Cosina/Voitlander lenses offer incredible value and I even own one.

 

Robert

 

Hi Rob,

 

You may be a bit surprised when you see the results in the 28 mm lens tests.

 

The analogy of Canon L: Sigma = Leica M: Cosina Voigtlander is not correct. Many people are gradually coming to realize that. And I say that without even factoring in the fact that many Canon L lenses do not perform very well.

 

The Zeiss and CV lenses aren't just less expensive alternatives to the Leica lenses, they're legitimate alternatives, period.

 

The proof will be in the pudding. Which CV lenses have you worked with on the M8 so far?

 

The 28 article should be out in a few days and I think you'll find it interesting.

 

BTW, the drawing of the Zeiss lenses is not "Leica like". As a family of lenses, the ZMs have a look that is quite different from the newest Aspherical Leicas. They're strangely consistent in that respect. Moreover, Leica lenses themselves have very different looks.

 

This whole idea of "off brand" with respect to these other lenses is mistaken. It is based mostly on a prejudice, not on how these lenses actually draw.

 

Cheers,

 

Sean

[sean_reid]

But Leica would have to put the correction data into their firmware to fix the Cyan cast for Zeiss and CV lenses. I don't see any reason they would do that, at least any economic reason.

 

No they would not, please see my previous posts on this topic.

 

Sean

[sean_reid]

Yeah, I pointed out the "closed system" issue back when Leica first announced the lens coding for vignetting correction. Whether by accident or design, it was the first step to locking out CV and Zeiss (as well as people using old Canon, Konica, etc. lenses).

 

Hi Joseph,

 

It wasn't closed system initially when it was designed simply to reduce what Leica calls "artificial vignetting". With the correction that's now needed for 35 mm and wider lenses, the in-camera processing triggered by coding is not an option (for color work), it's a necessity. At that point, it does become a more closed system.

 

All,

 

What some people here may not be keeping in mind is that there is no such thing as a "best lens". Different lenses draw their images differently on the sensor. It's up to the photographer to choose what kind of drawing he or she prefers. A closed system greatly reduces that range of choice. I'm against the idea with Hasselblad too.

 

Cheers,

 

Sean

[scott kirkpatrick]

although I've worked a little with OS/2, I had no idea that it was related to Windows NT.

 

OS/2 is unrelated to NT, except that they have some overlap in time. NT was written at Microsoft by a new team, hired from DEC. OS/2 was written at first jointly by IBM and the Microsoft Windows team. After the split between the two companies, each finished their own product. The Windows and NT products evolved in parallel inside Microsoft for some years. I used to attend the WinHEC annual conferences and hear about the always promised "convergence" which didn't quite happen in Win 98 or Win 2000, but only really happened around the WinXP timeframe, when NT had disappeared.

 

scott

[carstenw]

Advertisement (gone after registration)

The original GUI layer in OS/2 2.x (2.x was used for decades in ATMs and was rock-solid. Only later did they get changed to the blue-screen capable O/S) (PM = Presentation Manager) was fully compatible with that of Windows 3.x, at the API level. In Warp (OS/2 3.0 - fully 32 bit years before Windows), the new GUI, WPS (Wok Place Shell), was fully object-oriented, including inheritance, and you could inherit from every part of the interface to specialise behaviour. It was pretty cool, and had great performance and stability way before NT was worth anything. Microsoft put the thumbscrew on IBM to kill it though, because they felt it was standing in the way of Windows 95, which was so much less advanced, but came out later than Warp.

 

Anyway, enough of that (I used to be the editor of EDM/2 (The Electronic Developer Magazine for OS/2) for a few years, and reviewed books for it for years before that. Even developers inside IBM would subscribe to EDM/2, even before we put it on the web, where it was distributed as a Help file).

[grober]

In regards to lens coding, has anybody asked Cosina or Zeiss if they would provide a coding conversion on their exisiting lenses?

 

A related issue, sort of: Zeiss answered my e-mail about "fixing" their 25mm ZM Biogon lens so it pulls up the 24mm frame in the M8's viewfinder. (Right now the M8 thinks my Biogon is a 28mm.) Zeiss is aware of the issue and will let me know when he can affect a solution.

 

I will also use this Biogon on my Bessa for film work. Of course on the Bessa, I need to use an auxilliary finder (the Zeiss one is the best).

 

After Zeiss fixes this problem THEN I'll worry about the 6-bit coding issue.

 

-g

[chrism]

Where can I find more info on the topic? I was unaware that Xerox had sued Apple, or that IBM offered a MacOS option; and although I've worked a little with OS/2, I had no idea that it was related to Windows NT. (In fact, NPR built their DACS stations-information service on OS/2, because--I thought--it has some kind of multitasking that wasn't available elsewhere.)

 

I find this fascinating. Can you list any sources?

 

Many thanks.

 

--HC

 

RoughlyDrafted has a series of articles related to this history. The most recent is here, and the page has links to previous articles if you want to read the whole story.

 

Chris

[carstenw]

By 1992, Microsoft's NT 3.0 had proved its stability and capability, and people were porting major applications developed on SGI and Sun workstations to NT, giving Microsoft a tremendous boost in image processing, CAD, 3D graphics, and scientific applications.

 

One minor addition here. Your version makes it sound like everyone was happy to move onto NT. This was not the case. Coulter, of VAX (in)fame, had designed NT, and it had all kinds of problems, such as molasses-like speed, a poorly thought-out driver architecture, no decent graphics cards, and so on. VAX was also not respected by Unix-fiends, whereas IRIX from SGI was the high-end graphics platform. The way that Microsoft got graphics apps onto the relatively lame NT platform was to buy out Softimage (where I worked before I moved to mental images), the leader of the 3D animation scene of the time, with Softimage|3D. They forced Softimage to port this to NT, and for the next-gen product, Softimage|XSI, forced the developers to use COM, which at the time was a major mistake. Softimage is still removing pieces of COM from inappropriate places, since at the time no one knew how to take advantage of it, and at what level it would work well. At an early stage even a float was wrapped in a COM object!

 

Anyway, the rest of the graphics field had to follow suit, since everyone was afraid of Microsoft and of being left behind, and once all the apps were there, the hardware came, and a couple of years later, Microsoft sold Softimage to Avid, who still own it.

 

NT started with 3.1, IIRC, and hit 3.51 before the speed was acceptable. 4.0 was when the graphics drivers were moved to the kernel-level, something Microsoft did because they couldn't figure out how else to get the performance. This is where NT's stability suffered, and it didn't really recover until Windows 2000. By then, the Mac platform was coming on strong with a "real" Unix, and Apple started writing and buying high-end graphics apps for it, so Microsoft has lost its stranglehold on that market again. SGI just makes highly parallel machines for the DoD these days, and are nearly bankrupt. Unless I missed it, and they actually went bankrupt. They never recovered after Microsoft's move.

[blakley]

Sean, you wrote:

 

It wasn't closed system initially when it was designed simply to reduce what Leica calls "artificial vignetting". With the correction that's now needed for 35 mm and wider lenses, the in-camera processing triggered by coding is not an option (for color work), it's a necessity. At that point, it does become a more closed system.

 

I take this to mean either that you don't think it was really necessary to correct for the vignetting or that you don't think it's feasible to correct for cyan corners in RAW processing or PS (am I reading you right?)

 

Certainly without lens coding we don't get lens info in the EXIF data (unless Leica implements your excellent suggestion to allow manual entry of lens data via the menus) - so automatic correction is probably out. But...

 

I don't get the feeling that manual correction would be terribly labor-intensive, though; do you think that anything more than a red circular gradient adjustment layer of the correct intensity is necessary?

[sean_reid]

I take this to mean either that you don't think it was really necessary to correct for the vignetting or that you don't think it's feasible to correct for cyan corners in RAW processing or PS (am I reading you right?)

 

Certainly without lens coding we don't get lens info in the EXIF data (unless Leica implements your excellent suggestion to allow manual entry of lens data via the menus) - so automatic correction is probably out. But...

 

I don't get the feeling that manual correction would be terribly labor-intensive, though; do you think that anything more than a red circular gradient adjustment layer of the correct intensity is necessary?

 

The "artificial vignetting" is pretty mild on the M8. One could try to correct for the cyan drift in RAW conversion or Photoshop but

 

1) He or she will have to remember what lens each frame was made with. That can be interesting for a pro shoot of 1000 frames.

 

2) The correction in RAW conversion or post (if such existed and if one knew which lens was used for each picture) would have to be done frame by frame, which is tedious.

 

Cheers,

 

Sean

[robsteve]

Hi Rob,

 

You may be a bit surprised when you see the results in the 28 mm lens tests.

 

The proof will be in the pudding. Which CV lenses have you worked with on the M8 so far?

 

The 28 article should be out in a few days and I think you'll find it interesting.

 

BTW, the drawing of the Zeiss lenses is not "Leica like". Sean

 

Sean:

 

The 28mm test will be interesting. I would guess the Leica 28mm Summicron at F2 will have better performance than any of the other 28mm lenses at F2.

 

I do not have an M8 yet, so I cannot make a comment on the CV lenses and the M8 other than what I have seen posted.

 

In regards to my comment about the Zeiss, it was probably a bit out of context. I should have said the 15mm and 85mm have high quality images and high prices just like the Leica lenses. I did not mean to infer they had the same image rendering or lens signature as a Leica lens.

 

Robert