Jump to content

The Sony A7 thread [Merged]


dmclalla

Recommended Posts

Advertisement (gone after registration)

Thanks Jim for the link to your blog, very interesting reading. I hadn't fully appreciated how ISO-less the A7R is. I've just done a very quick check (set exposure to give almost ETTR at 3200, then drop the ISO) and I'm happy that I can at least recover 4 stops in Lightroom - at 200 results pushed 4 stops are visually identical to the 3200 one. At 100 I get an unacceptable cast, but fully accept your point that I don't need to be spending time tweaking up the ISO in camera if I want to more quickly get the image. I'll be working my way through your blog, thanks :)

 

It's a pleasure. Here, and here are a couple of links to pictures with the a7R at various amounts of push in Lightroom.

 

Especially if you have a background with film photography, it's very strange exploring exposure methods that take advantage of a camera's ISO-lessness. Don't hesitate to ask questions here. I'll be glad to help. There are others on this forum who have been exploring the issue with the M9, which has a bit of strange behavior in that its performance actually gets worse if you turn up the ISO dial past 640.

 

Here's a link to one of the M9 threads. Here's a link to another, earlier thread.

 

Jim

Link to post
Share on other sites

  • Replies 4.9k
  • Created
  • Last Reply
Given that the a7R is pretty darned close to ISO-less, I don't understand why you need to access the ISO settings so often. Or are you opposed to adjusting Exposure in post?

 

Jim

 

Whilst I am familiar with this concept from the Phase One backs, I own a P20 and they exhibit this same behaviour, the disadvantage is the lack of a clear preview on the rear screen. With the P20 this is especially important for focus checking (no live preview) . In these earlier backs the results of post pushing, especially using Capture One in the current V7 form which has improved the processing, is that the software does a great deal better than the firmware. Visually Jim, on my less than perfect screen, your results show no difference with either method, perhaps you could confirm this? In that case to obtain a decent rear view preview I would see no harm in ramping the iso in camera although it appears the auto implementation leaves something to be desired.

Can I also give a nod to the blog? Your work rate is exemplary!!

Link to post
Share on other sites

Visually Jim, on my less than perfect screen, your results show no difference with either method, perhaps you could confirm this? In that case to obtain a decent rear view preview I would see no harm in ramping the iso in camera although it appears the auto implementation leaves something to be desired.

 

Chris, if you stay away from the ISOs where the camera gives up on analog gain, and just starts bit shifting, and leaving the strangely behaving M9 out of the mix, you're right that there's no harm in increasing the ISO. That is, until you start to clip the highlights. Clipping in the camera happens suddenly; one minute you're increasing the count one for each n photons, and the next minute the count is stuck at full scale. However the Exposure control in Lightroom PV 2012 works more like film, introducing a gradual "shoulder" that is much more pleasing. Note that adding an Exposure adjustment layer in Photoshop doesn't get you the PV 2012 algorithm.

 

Sony has two modes for the preview and the EVF. One tries to anticipate what the image will look like when captured, and will get darker as you "underexpose". The other keeps the image bright no matter what. Unfortunately, when you use the latter mode, you lose the live histogram. I find the best thing is to keep the histogram and only "underexpose" one to three stops.

 

Just like UniWB, with the unfortunate side effect of a greenish finder image, using the camera in a way that the manufacturer never intended does have a few problems, but I think it's worth it.

 

Can I also give a nod to the blog? Your work rate is exemplary!!

 

Thanks. It's a disease.

 

Jim

Link to post
Share on other sites

Jim

 

Agree that "underexposing" does give highlight protection and the EVF does trump live view which has just arrived in MFD, at $35,0000 . Just a shame that give a decent, and it really isn't that anyway, frame rate they mucked about with the RAW file chucking out perfectly good data. A menu option could be forthcoming? Trouble is not enough are likely to complain.

I read little about software preferences, I lean to Capture One, and wonder if Sony have given enough data out so all are turning in good conversions given the jpeg engine yet again we need to turn to RAW for even web use it seems? Some of those graphs from RAW digger make the data look like a pushed about jpeg file!!

 

Personally I'm still uncomfortable on my fence over 7/7r, I'm leaning to the r for the file and willing to compromise on wide rangefinder glass as more exceptions that prove the rule seem to come along and I do shoot LCC files on the P20 so it's not unfamiliar territory.

Link to post
Share on other sites

Just a shame that give a decent, and it really isn't that anyway, frame rate they mucked about with the RAW file chucking out perfectly good data. A menu option could be forthcoming? Trouble is not enough are likely to complain.

 

Chris, I don't like the idea of lossy compression of raw files. I'm so concerned that I won't pick the lossless compressed option in the camera setup menu (assuming there is one) unless I know the algorithm they're using. That said, in two months of experience with the a7R and a year of making images with the RX-1 (which uses the same compression algorithm) I've not found any harm to the images from Sony's lossy compression.

 

I lean to Capture One, and wonder if Sony have given enough data out so all are turning in good conversions given the jpeg engine yet again we need to turn to RAW for even web use it seems? Some of those graphs from RAW digger make the data look like a pushed about jpeg file!!

 

The Sony compression algorithm has been published, although (I think) not by Sony. I can find a link if you're interested. It consists of moderate tone compression followed by delta modulation in chunks. WRT raw vs JPEG for web use, I always shoot raw no matter what the camera (phones expected). With Lr, it doesn't really matter to the workflow. I imagine that C1 is similar in that regard.

 

Personally I'm still uncomfortable on my fence over 7/7r, I'm leaning to the r for the file and willing to compromise on wide rangefinder glass as more exceptions that prove the rule seem to come along and I do shoot LCC files on the P20 so it's not unfamiliar territory.

 

If you can live with 24 megapixels and are willing to use native glass, the a7 looks like a winner to me. For handheld use with lenses shorter than 90mm, I think the a7R is fine. On a tripod, when you care about maximum possible sharpness -- and if you don't, why did you buy the R -- you have to take precautions to avoid the effects of the a7R shutter shock, and there's no fix yet for really long lenses.

 

No easy choices here, except the first one.

 

Jim

Link to post
Share on other sites

Jim

 

Thanks for all your work. Way over my head though.

 

Can you please summarize again max ISO for the Monochrom and for the M240. Plus can you throw in the RX-1 for good measure. I use LR5.x and often push up exposure a few stops to get my desired results.

 

I sent my a7r back due to blurred images with my R lenses, WA included. Perhaps the a7 would have been the better choice.

Link to post
Share on other sites

Can you please summarize again max ISO for the Monochrom...

 

I've not tested the Monochrom. I repeat my offer to test one if someone will provide me with a loaner. Based on the fact that the underlying sensor is the same as that in the M9 and that the sensitivity is increased twofold by virtue of the removal of the CFA, I would expect that you wouldn't want to turn up the ISO knob beyond 1250.

 

and for the M240.

 

Because of the "green shadows" problem, you don't want to push the M240 more than two or three stops at low ISOs -- say, 200 through 800. In the very deep shadows, however, pushing in post yields lower noise than cranking up the ISO to 3200. So you're damned if you do and damned if you don't. I'd opt for avoiding the green shadows and accepting the noise, and turn the ISO up to 3200 and pushing a bit in post, then I'd stop.

 

Plus can you throw in the RX-1 for good measure.

 

The RX-1 is only a 13-bit camera in actuality. It uses digital gain past ISO 3200, so you don't want to use 6400 and up. The unity gain ISO of the RX-1 is between 300 and 400. There's no need to go more than two stops above that, or 1600.

 

I sent my a7r back due to blurred images with my R lenses, WA included. Perhaps the a7 would have been the better choice.

 

Not having access to one, I don't know how the a7 does with WA lenses. I imagine that it would do fine with teles, and, with the electronic first curtain, be far less prone to rocking and rolling.

 

Jim

Link to post
Share on other sites

I've not tested the Monochrom. I repeat my offer to test one if someone will provide me with a loaner. Based on the fact that the underlying sensor is the same as that in the M9 and that the sensitivity is increased twofold by virtue of the removal of the CFA, I would expect that you wouldn't want to turn up the ISO knob beyond 1250.

 

 

That rule rather ignores the nature of the noise you get with the Monochrom sensor. There's quite a lot written about this on the Monochrom thread - in short, at higher ISO, the Monochrom gives the appearance of grain the the shadows, rather than blobby noise you see with a colour sensor.

 

In practical terms, I leave my Monochrom with Auto ISO, and I set the shutter for blur, and aperture for depth of field. ISO isn't an issue (for me).

Link to post
Share on other sites

That rule rather ignores the nature of the noise you get with the Monochrom sensor. There's quite a lot written about this on the Monochrom thread - in short, at higher ISO, the Monochrom gives the appearance of grain the the shadows, rather than blobby noise you see with a colour sensor.

 

I was assuming that the intention was to minimize the noise. If you like the noise, then go for it.

 

Jim

Link to post
Share on other sites

I was assuming that the intention was to minimize the noise. If you like the noise, then go for it.

 

Jim

 

 

I think that's a little simplistic, in terms of black and white photography - Tri-X built its reputation on the basis of what would be "noise" by that definition.

Link to post
Share on other sites

Jim,

have you seen this ? :-

 

LumoLabs -- Shutter-induced blur for Pentax K-7

 

The sensor cleaning suspension comes under suspicion and other interesting work, not A7R but relevant.

 

I had not seen it. Very thorough work. I like the way the attachment proves what I'd surmised earlier, but had not demonstrated: that the shutter's contribution to the camera's rotation is more damaging than the translation that it causes. Thanks! I wish I had some of the equipment they used to make the measurements.

 

BTW, tomorrow I will post a direct comparison of the a7R/Leica 135mm f/3.4 APO-Telyt and the D800E/Zeiss 135mm f/2 APO Sonnar when photographing an oscilloscope with a grounded vertical input through a folded optical path of 55 feet, which yields a 1-sensel phosphor dot size. The Nikon pictures will be made with the mirror locked up. Foreshadowing: the Sony shakes more in both portrait and landscape orientation, but the Nikon shakes enough that you probably have to use at least 1/250 second (or a very long shutter speed) to get all the resolution the lens can deliver. 36MP makes it tough!

 

I used unmodified RRS L-brackets for both cameras. The comparison isn't completely apples to apples since the Zeiss lens wieghs so much more than the Leica one. However, I can't put the Leica lens on the D800E, and I don't want to risk damaging the lens mount of the Sony by putting the Zeiss lens on it.

 

I've already posted images of the Sony/Leica performance and the Nikon/Zeiss performance separately.

 

Jim

Link to post
Share on other sites

I've not tested the Monochrom. I repeat my offer to test one if someone will provide me with a loaner. Based on the fact that the underlying sensor is the same as that in the M9 and that the sensitivity is increased twofold by virtue of the removal of the CFA, I would expect that you wouldn't want to turn up the ISO knob beyond 1250.

Jim

 

Monochrom @ iso 5000:

https://secure.flickr.com/photos/93600820@N04/11898226316/

 

Monochrom @ iso 10000:

https://secure.flickr.com/photos/93600820@N04/11293646993/

https://secure.flickr.com/photos/93600820@N04/11293523315/

Link to post
Share on other sites

 

Jay, the issue here is not whether or not you can get good IQ from the MM at ISO 10000, but whether you'd get even better IQ from setting the camera at ISO 1250, "underexposing" by three stops, and doing a three-stop push in post.

 

Jim

Link to post
Share on other sites

Jay, the issue here is not whether or not you can get good IQ from the MM at ISO 10000, but whether you'd get even better IQ from setting the camera at ISO 1250, "underexposing" by three stops, and doing a three-stop push in post.

 

Jim

 

I understand the premise behind the question but, judging by the superb quality of Jay's images, it seems like nothing more than an intellectual exercise. I mean, how much "better" do we usefully need?

Link to post
Share on other sites

Jay, the issue here is not whether or not you can get good IQ from the MM at ISO 10000, but whether you'd get even better IQ from setting the camera at ISO 1250, "underexposing" by three stops, and doing a three-stop push in post.

Jim

 

Not to my eye.

 

See here:

http://www.l-camera-forum.com/leica-forum/leica-m9-forum/318679-monochrom-high-iso-test.html#post2615983

Link to post
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...