I downloaded and looked at both DNG files, and although the lighting seems like a mix of daylight and perhaps fluorescent, I still find that the pixel-level characteristics are much nicer than my 5D. Sharper, more pleasant detail. Sigh. If I were only a lot wealthier I would own 4-5 cameras. I suspect the M8 would get the majority of the workout, but as an only camera, I can't see it, and two camera systems is too expensive and makes no sense for me.
Carsten - a good lens on a 5D or 1ds2 can have stunning levels of detail too. If a good M lens was used on this pic it's likely we are looking at the qualities of the lens just as much as any quality of the M8 sensor. My own experience with M glass is that it is quite a bit better than a lot of SLR glass once you get in to the very fine details. This was evident on high res film scans so I suspect the same will be true with digital sensors.
I took a look at some of my sharpest images with the 5D, including the Canon 100 Macro, the 80 Lux stopped down, the 90 Elmarit and a few other lenses, and in none of them did I get the same fidelity in natural edge-sharpness as I see in this M8 photo. Over-sharpening has a harshness to it; in this portrait the sharpness has a very "soft" look to it, very natural and pleasing.
To be honest, it doesn't matter to me if the lens or the sensor is responsible, the point for me is that neither of them prevent these kinds of results. It does look like a very good system. If Guy is right about it being a lens in the range of 35-50 or so, then perhaps it was the 35mm Summilux Asph., apparently one of the great lenses of the M range.
Guy Mancuso wrote: that's too funny. Trying to please all the leica freaks is going to be a nightmare. They are fanactics about there beloved M film series. Shoot me if i get that bad.
Who will be the first to post a question someting like "Should I sell my M7 when my M8 comes in, or keep it?"
OK - I can see I am alone in not being 100% about the Leica look. I have learned so much from this forum and all of the wonderful contributions from Guy et alia. I take all of your judgements very seriously so I will certainly be checking out the M8 with a view to buying if the favourable consensus here consolidates after actual use. I can't wait to see Guy's trail-blazing enthusiasm when the box arrives...
I would like to see a JPEG test of the M8/28f2 against my 5D/T*35f1.4 in day and available light, although the M8's size/weight advantage has probably already won it that competition.
Brainiac,
Sean Reid has -in his DMR test- photographed the same subject with a)Canon 1DS+35/1.4L and DMR+35/1.4Leica and frienkly I could not see much of a difference if any.
Regarding noise at high iso I do not ecpect the M8 to be better or as good as 5d at1600 ISO and higher. Just pure speculation from my side.
For most peope the extra detail of 14-16Mp will not be that important esp under 11x17 print and 20x30 inch on screen (240-300, and 72-96 pixels and inch!)
However, the 12-16 bit DR is seen at any resolution. It maybe that factor that drives many of us to prefer a DMR to 5D rendering (plus C1 processing, that is non-trivial)
I don't see this so much a 'Leica look" where I think the glass makes more difference.
Those who use Leica on 5d etc have stunning images,
I know it is late in the game to say it (sorry Guy :-) but I was more skeptical BECAUSE of the enthusiam, as many of you are, before I decided to buy the DMR. It was the eidence that caused me to buy (actually less than a year ago in December) I read for WEEKS!
That said, if you go to the big thread you will see as Shirozina has experienced that even in smaller prints detail can change the image (P45 vs P45!) - increase transparency and 3-D (that "...I can put my hand right into the picture..." impression)
Regads, esp to the skeptics (keep us thinking! )
Victor
I often ask myself which is the best way to downsample without degrading image clearness and detail.
I sometimes even feel that upsizing images is causeing less IQ-loss than downsizing.
How do you guys downsize? And is there a recommandation to downsize in full factors (like 50%,25% etc.)? Just to make sure than 4 pixel can become one instead of 4 pixels have to become 1.4 (as an example)?
Well, I am hardly an expert at this because all the information I have is gleaned from looking at these discussions about Leica equipment over the web (since Leica cameras are a bit out of my price range), but I suspect that the difference in quality of edge sharpness that Carsten is talking about between Leica and Canon gear may have more to do with anti-aliasing or possibly filtering than the glass or something inherent to the sensor itself. Canon seems more concerned about eliminating color moiré and stair-step artifacts while Leica seems more concerned about the image being 'pristine' or something like that, i.e., not fooled around with.
AGeoJO wrote:
Hmm, the crop seems to be sharper than the non-cropped image. It looks like it was taken with a short tele and fairly wide open or even completely open but not sure about the FL. I hope that Phase One will work together closely with Leica and release an updated C1 before too long after the M8 is released.
The current C1 Version 3.7.5 supports the M8 now. Version 4.0 was announced at Photokina and will fully support the DNG standard.
gogopix wrote:
these can't really be full raw since they are only 10Mb files. It may be 8bit instead of 16 bit
all my DMR files are 19Mbits
so I assume this is STILL not the true IQ
it will be better! No No Victor ERROR # 1,234,567
NO! I just checked and full size image 16-bit is 10Mb. there must be some added comression going on (hopefully non lossy)
I wonder how they got another factor of 2 out?? (DMR files are all 19Mb
Victor
Victor
The word from Leica is that they found a way to get the file size of the raw M8 down to the 10mb level allowing 400 images on a 4GB card! This is the standard size, not a downsized raw file
Doing this to many files in a foreach loop is incredibly fast as the jpeg code only reads the file to the depth required.
The djpeg command decompresses the file, outputs it via the pipe (|) to cjpeg (compress jpeg) and then you direct to (>) a file. I think you can scale from 1/2 to 1/16.
If you are in the habit of handling very large numbers of files, you can save yourself a lot of time by working on the command line. Try generating a set of web jpegs from a thousand files in aperture and you will see what I mean.
For those who use Aperture, Mac OS X has been updated to 10.4.8 and there seems to be new RAW formats supported. I don't know if the Leica M8 is one of them.
t_streng: was the comparison you referred to done with a DMR versus a 1Ds?
From what I have seen the pictures straight out of the box from the 1Ds have more "pop" than the successors, like my 5D. The images out of my 5D need more tweaking than the 1Ds, just as the DMR also produces more vivid pics straight out of the box. The 1DsMkII and 5D are more dull and need more postprocessing to look alive.
open81: I would keep the M7, but hey that's me...
...just had a look at some old Hassy Superwide shots the other day and got the itch back for shooting film - looks great with so little effort... ...I almost miss the days I wasn't bogged down in front of the screen, tweaking away... ...almost.
Looks great, probably as much IQ you can get with a 10mp sensor. Though the real test would be on a tripod and a pic with more details and less OOF areas
From what I have seen the pictures straight out of the box from the 1Ds have more "pop" than the successors, like my 5D. The images out of my 5D need more tweaking than the 1Ds, just as the DMR also produces more vivid pics straight out of the box. The 1DsMkII and 5D are more dull and need more postprocessing to look alive.
Henrik - I've got to disagree with you on this one. 6-12 months ago I would have agreed, but a year later & 10,000+ clicks with the 1Ds2 my opinion has changed. I've got 1000's of images like this from the 1Ds - unedited from DPP (RAW--->JPEG):
There are many variables at work here (different light, different scene, different lenses...), but I've worked enough images from both cameras to say that the 1Ds does not have more out-of-camera pop than a 1Ds2.
In general I agree with this thread - you do match lenses to the sensor & image processing engine.
that's too funny. Trying to please all the leica freaks is going to be a nightmare. They are fanactics about there beloved M film series. Shoot me if i get that bad. 