有人根据菜斯提供的曲线, 对所有的菜斯头做了一个排队, 结果是折反头最好! 各位是否发表一下高见?

原文出自:http://www.listquest.com/lq/view.cgi?ln=yashicaslr&mid=1986&sp=&q=85&b=1&s=1&o=0&x=34&y=15&ci=c091510654

那位大虾翻译一下工大家参考? 我没有翻译软件.

Subject: [Y/C] Mirotar 1000/5.6 ... simply the best!

For those who care.

I have translated the MTF charts of Carl Zeiss lenses for the
Contax SLR into numbers. A single number for each lens,
representing a weighted average of the modulation transfer
functions that Zeiss supplies with their lenses. I asked Zeiss
for ascii or binary files of these curves, but they couldn't
give them to me (I would have been surprised if they could
though). Hence I found myself bent over papers for some time,
sampling the curves by hand.

I took samples at 0 mm, 5 mm, 10 mm, 15 mm, and 20 mm from
the optical axis, for both the tangential and sagittal
curves. With six curves at my disposal, I ended up with 30
numbers per lens. Subsequently, I averaged these numbers,
giving equal weights to the tangential and sagittal data.
Moreover, I gave a weight of 1 to the 10-lp/mm and 20-lp/mm
curves, and a weight of 2 to the 40-lp/mm measurement. The
reason for doing this, is that 40 line pairs per mm is the
most discriminating curve among the various lenses. Every
lens has a very good 10 lp/mm performance and is capable of
perfect 10x15-cm prints. No question about that. Differences
become visible only at large magnifications, where the
40-linepairs-per-mm data come into play. Since many people
seem to be mostly interested in the utmost sharpness, this
seems a reasonable approach.

Further, al the samples along one curve had equal weight, that
is to say, 0 mm is considered as important as 5 mm, 10 mm,
15 mm, and 20 mm. In this manner, lenses with a constant
performance from the image center to the corner are favoured,
as they should. Other people prefer to give less weight to the
corners, as these are less important for the image. That may be
so, but the difference between lenses is often found in the
corners in the first place. I could even argue that the very
image center in that case also should have a lower weight, as
it corresponds only to a small percentage of the total image area,
and also because a successful photo rarely has the main subject
at the image center. No, I decided to distribute the weight
evenly across the image.

So what's the meaning of the below numbers? Not much really,
you better forget them. One cannot possibly characterize a
lens with a single number. The numbers reveal nothing about
distortion, illumination fall-off, spectral transmission, or
a practical matter like manageability (can you envision
yourself conquering Mount Everest with the Mirotar 1000/5.6 in
your photo bag?) They do, however, give an indication of the
resolution and contrast that a lens is capable of. The MTF
curves as Zeiss provides them are measured ones. Measured
numbers of real production samples. If Zeiss is too optimistic
(or pessimistic) about a specific lens, I cannot help it. Also,
there's always a chance that there's some variation in optical
quality between production lenses. If you ask Zeiss --I did--
they say, of course, that they build lenses according to very
narrow tolerances, and that you can expect to get the lens as it
is specified. The safest conclusion is probably that if you own
or intend to buy a certain lens, the numbers below indicate what
the average performance *should* be like.

So I averaged the numbers as I described above and multiplied
the outcome by 10. Note that the uncertainty introduced by the
sampling amounts to approximately 0.1, and that differences
smaller than 0.1 are not at all significant. The third digit
is given solely for the purpose of sorting.
I start with the currently available lenses, to wit, those that
are listed by Zeiss on their site, and subsequently continue
with an overview per focal length (range), where some rare and
discontinued lenses pass in review as well.
The first lists presents the current lenses at their best aperture,
best magnification ratio, and best focal length in case it's a
zoom. The last number is the score.

currently available lenses
------------------------------------------------
01   1000/5.6         f/5.6   inf           8.44
02   35-135/3.3-4.5   f/8.0   inf   70mm    7.89
03   21/2.8           f/5.6   inf           7.88
04   100/2.8          f/5.6   inf           7.83
05   100-300          f/8.0   inf  100mm    7.80
06   500/4.5          f/4.5   inf           7.60
07   60/2.8           f/5.6   10:1          7.56
08   85/1.4           f/5.6   inf           7.56
09   200/2.0          f/5.6   inf           7.56
10   28-85/3.3-4.0    f/8.0   inf   85mm    7.55
11   35-70/3.4        f/8.0   inf   50mm    7.55
12   50/1.4           f/5.6   inf           7.51
13   100/2.0          f/5.6   inf           7.51
14   85/2.8           f/5.6   inf           7.43
15   45/2.8           f/5.6   inf           7.23
16   16/2.8           f/5.6   inf           7.18
17   50/1.7           f/5.6   inf           7.08
18   35/2.8           f/5.6   inf           7.03
19   300/2.8          f/5.6   inf           7.02
20   80-200/4         f/8.0   inf  140mm    6.98
21   135/2.8          f/5.6   inf           6.88
22   35/1.4           f/5.6   inf           6.85
23   28-70/3.5-4.5    f/8.0   inf   28mm    6.62
24   28/2.8           f/5.6   inf           6.56
25   25/2.8           f/5.6   inf           6.35
26   18/4.0           f/8.0   inf           5.84
27   500/8.0          f/8.0   inf           5.56
28   300/4.0          f/8.9   inf           5.39
29   180/2.8          f/5.6   inf           5.37
30   15/3.5           f/8.0   inf           4.53

15 mm - 21 mm
------------------------------------------------
01   21/2.8           f/5.6   inf           7.88
02   16/2.8           f/5.6   inf           7.18
03   18/4.0           f/8.0   inf           5.84
04   15/3.5           f/8.0   inf           4.53

25 mm - 28 mm
------------------------------------------------
01   25/3.5           f/5.6   inf           8.18
02   28/2.0           f/5.6   inf           6.86
03   28-85/3.3-4.0    f/8.0   inf   28mm    6.71
04   28-70/3.5-4.5    f/8.0   inf   28mm    6.62
05   28/2.8           f/5.6   inf           6.56
06   25/2.8           f/5.6   inf           6.35

35 mm
------------------------------------------------
01   35-70/3.4        f/8.0   inf   35mm    7.27
02   35/2.8           f/5.6   inf           7.03
03   35/1.4           f/5.6   inf           6.85
04   35-135/3.3-4.5   f/8.0   inf   35mm    6.49

* The PC distagon 35 mm is incorrectly displayed at
  the Zeiss site and therefore not (yet) included.

45 mm - 50 mm
------------------------------------------------
01   35-70/3.4        f/8.0   inf   50mm    7.55
02   50/1.4           f/5.6   inf           7.51
03   28-85/3.3-4.0    f/8.0   inf   50mm    7.49
04   45/2.8           f/5.6   inf           7.23
05   50/1.7           f/5.6   inf           7.08
06   28-70/3.5-4.5    f/9.0   inf   50mm    6.61

70 mm - 85 mm
------------------------------------------------
01   35-135/3.3-4.5   f/8.0   inf   70mm    7.89
02   85/1.2           f/4.0   inf           7.62
03   85/1.4           f/5.6   inf           7.56
04   28-85/3.3-4.0    f/8.0   inf   85mm    7.55
05   85/2.8           f/5.6   inf           7.43
06   35-70/3.4        f/8.0   inf   70mm    7.36
07   80-200/4         f/8.0   inf   80mm    6.64
08   28-70/3.5-4.5    f/10.5  inf   70mm    5.99

100 mm
------------------------------------------------
01   100/2.8          f/5.6   inf           7.83
02   100-300          f/8.0   inf  100mm    7.80
03   100/2.0          f/5.6   inf           7.51
04   100/3.5          f/5.6   inf           7.39

135 mm -140 mm
------------------------------------------------
01   100-300          f/8.0   inf  135mm    7.67
02   135/2.0          f/5.6   inf           7.08
03   80-200/4         f/8.0   inf  140mm    6.98
04   135/2.8          f/5.6   inf           6.88
05   35-135/3.3-4.5   f/8.0   inf  135mm    6.64

180 mm - 200 mm
01   200/2.0          f/5.6   inf           7.56
02   100-300          f/8.8   inf  200mm    6.88
03   80-200/4         f/8.0   inf  200mm    6.66
04   200/3.5          f/8.0   inf           6.27
05   200/4.0          f/8.0   inf           5.97
06   180/2.8          f/5.6   inf           5.37

300 mm
------------------------------------------------
01   300/2.8          f/5.6   inf           7.02
02   100-300          f/10    inf  300mm    5.88
03   300/4            f/8.9   inf           5.39

macro
------------------------------------------------
01   100/2.8          f/5.6   inf           7.83
02   60/2.8           f/5.6   10:1          7.56
03   100/2.8          f/5.6   10:1          7.46
04   60/2.8           f/5.6   inf           6.86
05   100/2.8          f/5.6   1:1           5.62
06   60/2.8           f/5.6   2:1           5.6
07   60/2.8           f/5.6   1:1           4.82

Finally, one may wonder what's the significance of the
numbers. The performance of the Mirotar 1000/5.6 is dazzling.
I never really looked into this mirror lens before; only upon
tackling the MTF curves I realized that it has a very large
image circle. The part that is used for 35mm photography has
very, very good modulation transfer.
At the bottom end, we find the 15/3.5. Does it make this a bad
lens? Not at all. For its focal length it is presumably among the
best, it's just not as good as the other focal lengths. I have no
experience with this lens, but I do have the 18/4, which is capable
of very sharp slides. However, when employed in high-contrast
lighting conditions, one weakness becomes apparent: lateral
chromatic aberration. One encounters coloured edges around dark
subjects against a bright background. In my experience, chromatic
aberrations are also present in a modest wideangle as the 28/2.8.
As it appears, the wider the lens, the more problematic the design:

01   35/2.8           f/5.6   inf           7.03
02   28/2.8           f/5.6   inf           6.56
03   25/2.8           f/5.6   inf           6.35
04   18/4.0           f/8.0   inf           5.84
05   15/3.5           f/8.0   inf           4.53

Luckily, the newer designs 21/2.8 and 25/3.5 (not in production) do
away with these problems. These wideangles are truly exceptional, not
only for a wideangle, but also compared to the vast majority of other
lenses. This said, I should stress that the conventional wideangles
offer good value for the money and that it takes very critical
inspections to discover their weaknesses.

A similar discussion can probably be addressed to (non mirror) tele
lenses, which suffer from longitudinal chromatic aberration. Since
I don't own a tele myself, I prefer however not to say anything
about it.

Are fixed focal lengths always better than zoom lenses? Look in
the above lists and draw your own conclusions.

Are macro lenses always optimized for macro photography? Draw
your own conclusions.

Etc. etc.

Walrus

同意小菜看法。
如同 麦记 和 管机的差别。
至少我不会将在家里摆满鉴听级喇叭，功放。
我会摆一部管机还有英制小炮来听巴卡尼尼，菜芹。
有时侯失真就像鸡精一样美。
何比苦苦追求一丝一毫？

To old seagull
   周末去败了一架禄来35S回来,周日忙着试镜头.
   我总觉得菜斯设计者在设计35毫米镜头的时候,思路和别的厂家不一样,是一种轻松的玩票心态,设计的是一种"沙龙味"的镜头,注重的不是数值而是实拍的某些特质.
   所以35毫米菜斯镜头的顶级之处也就表现在它的"个性"范围内,出了这个范围,输给LNMCP很正常.
   一个最有味道的女人,不见得是个最美丽的女人,这就是35毫米的菜斯.
    
    

小蔡周末到哪儿去了? 他应是很了解康太师的, 应该有话说.

我看了photodo上菜斯100-300评分比泥糠假能70-200/2.8低, 也总有些怀疑. 好了, 在这里, 这个头排在菜斯自家的第五位, 按理说应该比泥糠假能那些好点吧?

我怀疑康太师家的人对镜头好坏的评价都是根据mtf曲线来判断的, 网上的意见倒很符合Walrus的排队.

小蔡周末到哪儿去了? 他应是很了解康太师的, 应该有话说.

[此消息已由 old seagull 进行编辑(编辑日期 11-13-2000).]

向老海鸥致敬！

镜头的评判的标准原本没有标准。关键在于如何看。
比如对于红楼梦，道学家看到了"淫"，民俗家看到了"情"，等等。

据说N的小钢炮刚开始的评分极低，N FANS提了意见，换了只头，所以。。。

我估计P的那只恐怕是没人提意见，或者是懒的提意见。。。看过阿康的片子，绝对不应该是那个分数。因此我觉得价格可能也是一个评比条件。。

老海鸥：还记得俺有G45/2？这头不错，俺同意！俺另外两支“顶级”头，可能就不会得到多数人的同意吧？看看PHOTODO上同规格的分数：
50/2.8 MACRO(1:1):
Minolta AF50/2.8Macro:4.5
Nikkor AF 60/2.8Macro:4.2

100/2.8Macro(1:1):
Minolta AF100/2.8Macro:4.5
Nikkor Af105/2.8Macro:3.9
Pentax FA100/2.8Macro:4.2

但俺看人家的N105和P100的都很好，时不时感觉他们的比俺的好。

俺怀疑的地方很多，举些例子：
1、TAMRON 200-400/5.6LD (IF)在PHOTODO上是1.2分，但俺看到了不少对这镜头的好评；
2、TOKINA 20-35/3.5-4.5II的得分是3.3，但20-35/2.8却是2.5，俺对这个结果怀疑；
3、TOKINA AT-X300/2.8、AF AT-X300/2.8、AF AT-X300/2.8II的光学结构都是一样的，一个比一个后出来，但是分数分别是：4、3.8、3.5，却是一个不如一个，而且最大差别达0.5分，这可是要命的，俺怀疑；
4、CANON EF20-35/3.5-4.5得分3.4，而17-35却是3.2，这有可能，但会误导一些人；
5、PENTAX SMC-FA28-200/3.8-5.6AL(IF)是向TAMRON OEM的，可是PENTAX的分数是1.7,而TAMRON的是2.7，差别居然有1分；
6、PENTAX FA80-200/2.8ED(IF)的分数是3.2，其他品牌同类型的分数是：
C EF80-200/2.8L:4.2
C EF 70-200/2.8L:4.1
M AF80-200/2.8G:3.9
N AF80-200/2.8D ED:3.9
N AF80-200/2.8:4
N AF-S80-200/2.8 IF-ED:4.1
S AF70-200/2.8APO EX:3.9
TAMRON SP AF70-210/2.8 LD:3.4
TOKINA AF AT-X80-200/2.8:3.4
PENTAX的分数是最低的，最大差距为1分！但好象是PENTAX的价格最高，难道PENTAX拿这个昂贵的破玩意来糊弄P用家？（如果真是这样，向阿康致敬！）
。。。。

说了这些，无非想说明PHOTODO上的东西可疑，不过在这个主题下讨论是跑题了.

这个比拼不是photodo, 大家不要搞错了啊!

话又说回来, 如果你再买一支镜头前有人告诉你, 这个镜头的photodo评分是0.5, 你会不会不信那一套, 继续掏腰包?

to: 西老, 你的三个顶级中有一个是G的45/2, 说这个镜头时顶级的, 不也是多数同志的意见吗?  

对对，PHOTODO不能和MTF等价，MTF是有其价值的。

同意！PHOTODO确实毁人不倦。那个经过“甲醛”的MTF根本就不能说明任何问题！
看了PHOTODO，决不再信MTF。

部分同意F3hp的说法，但MTF还是有用的，垃圾的是PHOTODO的打分，本来MTF就是一个比较复杂的过程，硬要将这些东西综合成一个分数，肯定要磨掉很多内容的。而且PHOTODO上老是对无限远来测试，将焦点调在无限远拍摄的几率有多大？俺估计很小的。以这样一个小几率的测试形成一个分数，的确是垃圾，但有人用这个垃圾分数将镜头排列出“世界顶级镜头”，很好玩，居然俺的三支镜头也排列在其中！但俺和人家的非“顶级”镜头比较，还没看出区别呢，有时人家的还比俺的好！所以那里的分数是垃圾，俺同意。

[此消息已由 xitek 进行编辑(编辑日期 11-11-2000).]

F3HP: 您老可能只看了他的题目吧? 您看看他的排队, 不是相当符合网上大多数同志的意见吗?

辣泡菜: 好! 辣辣F3HP大虾!

难道就没有定量比较镜头质量的方法吗？大批量生产的民用镜头真有一年生产几把小提琴那么玄吗？大师做小提琴好象是感性、技艺因素多一些，而民用镜头就是打工妹靠设计图纸、工艺装备、质量检验等一般工厂的手段加工、装配的啊？

没细看，但可以得出的结论是PHOTODO是垃圾。:-)是MTF无用论的有力佐证。