I kinda feel bad at bringing the Sigma to this conversation, but it deals directly with synn's issue.
Thank you for that sample. It's absolutely brilliant and yes, if Sigma ever get around to making a full frame version of this sensor and an interchangable lens version of the DP3m, they are gonna hit it right out of the park.
Having said this, I'm still surprised the Sd1 wouldn't work for you.
Edmund: Yes, even on Sigma site its about $1700
When a Phase back produces good images it makes superb images, I have done 44" wide from 1/4 of the frame of my P45+ with no problems, and there is huge DR.
Yes, MF does great at this.
but many people in this forum have had at least this level in quality for 10 years or so.
I think your key word is quality. I would disagree on people seeing this level of quality.
I also think this varies between people in how quality is perceived, and there is the rub on measuring such a mix of esthetic and technical info. Meaning you are right in one sense, but I would disagree in another sense of looking at this file. This comes from something Jerome pointed out....The noise in shadows are not misinformation, they are simply luminosity from the true data. This is NOT seen easily in 10 years of files.
Sigma sabotaged the SD1 by overpricing it; I hope at some point they will make larger sensors and provide programming help so their usage spreads.
I hope you're right and they be a bit more quick about it! :-)
The one thing which everyone here prefers to forget is that Sigma is a lens manufacturer, and they "donate" a really good lens with each DP camera; An equivalent lens for an M series Leica or a dSLR or an MF camera would cost several times the price of the whole DP package.
Good point!
Erik:
just an oddball camera with a small market share that needs a lot of development effort.
Maybe Oddball, but with proper foundation and potential to dominate. Yes, a lot of dev needed. FF at the minimum, and they should partner up as they need the leverage.
Hulyss:
Just open the files and then only batch export ProPhoto Tiff 16bit and you are done
(even at 16 bit TIFF you lose data compared to the original raw file).
I'm surprised you'd accept this.
This I agree with Doug, and not JUST because you can't do 16bit TIF and get "the job" done, but more so in principal of the raw info as an archival data(for the "photographer", not production content producer) vs cooked.
I too think of all the possibles Sigma can do with a FF sensor, a true 40+mp, but they are moving very slow, and making baby steps. They need rocket fuel! And yes, they should be on the forefront helping other Dev apps with their RAW file type.
As far as the file on first impression.....I think it is a WOW! Look at the white area of the eyes. I have not see such true information in this area on so many camera types. There is NO color guess work pixels to be found!!! Lips, no color bleed guessing. It looks like someone took a Scotchgaurd pad on the glass and removed a thin layer of haze "film" of misinformation. Jerome is rather harsh on this to even compare to a Point and shoot.
Have you worked with a Phase file in Capture One v7??
It's incredibly sharp at 100% and it uprez's phenomenally well.
Loss of sharpness due to Bayer pattern used to be a big deal. That's just not the case anymore. I'd rate it as 10-15% depending on the subject.
The idea of Sigma advertising the SD1 as a 46mp system is a bad joke.
I agree with Doug. Yet I wouldn't dismiss or underplay(not that Doug is doing that) the different "level" or rather approach to clarity the Foveon produces.
If you let the $ difference get in your way, it is easy to try and dismiss. but if you look at the file and result, you will be hard pressed to dismiss it.
Yes, 46mp is simple dishonest advertising. It is the worst point of the commercial market game to play.
Phase One supports nearly all commonly used professional cameras. Canon, Nikon, Sony, Samsung, Leica, Mamiya Leaf, Olympus, Panosonic.
I can't speak for them, but I'm quite sure the reason this camera isn't supported is because it has a very unusal sensor which would require a lot of custom work to support, and there are very few pros using it. Very poor return on investment for their time.
Very true about C1, and I shouldn't has stated it that way. I was thinking more that they have their own priority before others, yet they HAVE extended C1 to meet so many others. I hope SigmaFoveon is in the works, A7R too :-) But very poor return on investment perhaps you are referring to when they fist launched it. It is now $1699! If it had a Canon or Nikon mount, I would have easily had one in my arsenal.
We are between people who know what we are talking about, that for sure. To not mess around, some of us are very intrigued by this foveon technology. It is not mature, some said this is the future of photography
I maybe one of those people, but I always have a back up(funny how what I use as the main I refer to as a backup) :-)
We all seek the perfect Camera/software line for our specialized or generalist work but it do not exist. Even when we put large amount of money, deep in us, even if we do not say it, we are not THAT happy sometimes.
Yes, even if we are, when you see such greatness from something so unexpected into such an open large and able market, it can make a lot of waves.
Jerome :
I am sorry, but in my opinion this is a bit overexposed on the skin (can't recover all highlights), does not react very well to processing, shadows are very noisy (but it is only luminance noise) and some bands appear on the top if try to process the file. As to the colors, I suppose that this woman wears make-up, so it is difficult to judge: it looks like make-up foundation color, not skin color.
I mean: it is very, very good for a point and shoot and at present price it is almost a steal. But is shows clear limits and I am not so sure what you are trying to prove with that picture.
It is a raw file. Are you saying the highlight a touch on her forehead and under her right eye are blown out and cannot "pull back" in the dev? I highly doubt this. Hulyss, please test.
Not reacting very well to processing? (this maybe a 16bit vs 12 or 14bit issue I would guess. I know the MF files are meaty and have mass info to push around. But this you can't tell unless you have the RAW!
Why would you try and process this cooked file? That is not a apples to apple comparison! Hulyss saying to do this in my opinion is something he was referring to for his own needs, and not something to use as a baseline of measure!
Yes she wears makeup, this file is screaming that at you. Its the first thing I noticed. You can actually see the cake layer vs thin areas. Look at how clear the lips are with the rouge! The best part is, no matter how far you zoom in, the color information is not a blurry mess. it stops where the reality of the subject stops...No interpolated info!
With all this said, you can't ignore how our brain functions. It in of itself fills gaps and information and interprets color differently in situations we are faced with. If you watch this video it has some great examples of this:
http://www.youtube.com/watch?v=SzZVzzgTxys To compare this to a point and shoot is not just wrong, but dismissing a great deal of information.
I think you maybe looking at certain numbers and charts to make up for a calculation. I have used a lot of gear and seen many files from many sources, MF and others. There is no doubt in my mind that IIIIFFFFF(big "IF") Foveon can get a FF or more, and at least double the resolution, AND get the dev/raw ironed out, it will be the next industry shifter.
Until then(and further on), we will have these conversations.
Indeed, and it also shows that there is hardly any color separation between the channel responses. There is more overlap than separation, which also shows when examining the Raw data (which almost looks like a monochrome RGB image). There is significant mathematical separation and amplification required to produce (saturated) color, which also explains the relatively poor high ISO performance of Foveon sensors. Also skin tone color in the shadows is pretty poor.
It is almost surprising how a color image can be calculated from that source data, but it also demonstrates that with clever post-processing almost any Raw data can be made more acceptable. So I'm still not sold on the skin color from e.g. Leaf backs being caused by the sensor. It's more likely a combination of CFA filter choices and an IR filter to match, and a Demosaicing that favors skin tones. The latter is just software doing its job.
Cheers,
Bart
I think its a mistake to compare things looking at the TIF file, as I think Doug would agree. We should be looking at the RAW. Without that, we can't conclude things.
Look at the improvements C1 made from earlier versions to the way v7 processes. Sigma is still at the first stages.
BTW: Skin tone in the shadows is great to me. This is often where sensors add misinformation.
Regarding the Leaf back... Thats why I think Foveon is a great sensor technology, it takes the sensor out of the equation. But you not liking the colors your subjective preference.
The interesting question with Foveon is how much software improvement can be done, compared to all the work which has already been done with Bayer
yes Edmund!
Seeing Bernards pix, I'm kinda surprised I don't have one of these cams! I don't have any point and shoot sized camera yet
