Lee Saxon wrote:
But wait!!! Anyone who says 11+7 isn't enough gets shouted down and called a Sony hater and a troll! How could 12 not be enough
12bit vs 14bit, as implemented in DSLR's, matters little from all the tests and theory I have seen and read (any difference is below the noise floor I believe the theory goes). 11+7 lossy raw compression silliness - where you could only truly claim what? 7 or 8 bits when considering the entire range (not just "in this part of the range here, it's actually 14 bit"... or similar nonsense) matters more to me at least.
It's possible some would prefer the 12 vs 14bit to matter more as that is controllable (using certain settings) whereas the compression stuff forced upon users is not controllable.
Tariq Gibran wrote:
12bit vs 14bit, as implemented in DSLR's, matters little from all the tests and theory I have seen and read (any difference is below the noise floor I believe the theory goes). 11+7 lossy raw compression silliness - where you could only truly claim what? 7 or 8 bits when considering the entire range (not just "in this part of the range here, it's actually 14 bit"... or similar nonsense) matters more to me at least.
It's possible some would prefer the 12 vs 14bit to matter more as that is controllable (using certain settings) whereas the compression stuff forced upon users is not controllable. ...Show more →
Sony's 12 vs 14 bit is actually altering the ADC operation vs just changing the encoding of the raw - this is the source of the lower DR/extra noise. I don't see this difference in a DSLR 12 vs 14 bit, such as on a Nikon.
snapsy wrote:
Sony's 12 vs 14 bit is actually altering the ADC operation vs just changing the encoding of the raw - this is the source of the lower DR/extra noise. I don't see this difference in a DSLR 12 vs 14 bit, such as on a Nikon.
So it's not the same 12 vs 14bit question since Sony is doing something a little different? Got you.
snapsy wrote:
Sony's 12 vs 14 bit is actually altering the ADC operation vs just changing the encoding of the raw - this is the source of the lower DR/extra noise. I don't see this difference in a DSLR 12 vs 14 bit, such as on a Nikon.
I haven't been able to see the difference on Nikon DSLRs either. You're probably right about why - they're doing the same 14-bit ADC conversion but simply truncating what gets written out to the memory card.
But I would think that would only save a little memory card space without making much difference in processing time/power.
So why would Nikon bother having certain modes which default down to 12? Or am I remembering incorrectly and they don't?
Here's a test I just did on the A7s - 30 second exposure EFCS vs 30 second bulb EFCS vs Silent Shutter (Fully-Electronic Shutter). For some reason the black point on the second bulb exposure for the +7EV test is off - I would ignore that and just compare the noise.
snapsy wrote:
Here's a test I just did on the A7s - 30 second exposure EFCS vs 30 second bulb EFCS vs Silent Shutter (Fully-Electronic Shutter). For some reason the black point on the second bulb exposure for the +7EV test is off - I would ignore that and just compare the noise.
Your test shows worse noise in Bulb mode than mine when pushing to an extreme 7 stops. Different cameras and different exposure increase though.
Although it seems that all A7 series bodies have deficiencies with these modes, perhaps the camera used may make a difference. My test was with the A7R.
Here is a test with the new A7RII where there is almost no change between the EFCS vs Silent Shutter captures:
Keep in mind that the poster didn't push exposure.
snapsy wrote:
Here's a test I just did on the A7s - 30 second exposure EFCS vs 30 second bulb EFCS vs Silent Shutter (Fully-Electronic Shutter). For some reason the black point on the second bulb exposure for the +7EV test is off - I would ignore that and just compare the noise.
Setting the exposure all the way up to +7 EV is pretty extreme though, and the difference is negligible at +4 EV so I'm not going to stress out about this.
Still weird that Sony has to do this in Bulb mode. In the other modes, it makes sense.
snapsy wrote:
Here's a test I just did on the A7s - 30 second exposure EFCS vs 30 second bulb EFCS vs Silent Shutter (Fully-Electronic Shutter). For some reason the black point on the second bulb exposure for the +7EV test is off - I would ignore that and just compare the noise.
therealthings wrote:
Now what are we going to do with our big stoppers and singh rays for our >30s exposures. ISO 400 doesn't sound like a great solution to me.
Maybe then try not to underexpose your shots by 7 stops?
shirozina wrote:
Maybe then try not to underexpose your shots by 7 stops?
Or if you must underexpose by 7 stops-- how about stacking a couple 29s exposures?
I'm willing to wager that exposures over 30 seconds are exceedingly niche in the grand scheme of things, and a quick accounting check probably showed Sony that not giving this niche it's ideal setup probably saved the company money and hassle in the long run. (Also, I don't think Sony really gives a hoot about defending Big Stopper and Singh Ray sales.)
I'd file this 12 vs 14 bit issue under: Things it's good to be aware of so one can work within the limitations of the system.
Agree for me at least over 30 seconds is not a everyday scenerio even with a 6stop ND. I usually work with 3 and 6 stop ND. Of course we do get into longer but we have to see what percentage that really is. Also now that we know than we can figure out our work a rounds. I'm wondering though even though Sony put this in print. Will we have the exact same issue as the A7r or maybe it maybe better with the new sensor. I think this is a test that needs to be done with the new cam and see what exactly those results are. Hopefully we will have a better idea towards the end of this week since a lot of us get there camera this week. Mine is coming from Amazon on the 6th.
Folks shooting Astro type work may need to revise there workflow
Lee Saxon wrote:
But wait!!! Anyone who says 11+7 isn't enough gets shouted down and called a Sony hater and a troll! How could 12 not be enough
Because people still aren't getting this subtle point, there's a huge difference between a linear 12-bit encoding and a non-linear 11-bit encoding (used in Sony's RAW compression) which is actually preferable. The non-linear encoding uses a greater bit depth where it's needed most i.e. the shadows, and fewer bits where it makes no difference, i.e the highlights. A linear 12-bit encoding just throws away 2 bits everywhere.
Matt Grum wrote:
Because people still aren't getting this subtle point, there's a huge difference between a linear 12-bit encoding and a non-linear 11-bit encoding (used in Sony's RAW compression) which is actually preferable. The non-linear encoding uses a greater bit depth where it's needed most i.e. the shadows, and fewer bits where it makes no difference, i.e the highlights. A linear 12-bit encoding just throws away 2 bits everywhere.
Dunno about preferable, for highlight recovery at least IMO not. In the 12-bit mode Sony CRAW bit depth is alwaus <= 12, it is 12 bits for luminance values 0 - 2688 and then it is 11 bits until luminance 5184, 10 bit from there to luminance 7580. Explained here in detail.
Here's a partial histogram (reduced bit depth RAW due to Long Exposure NR) from Rawdigger in "place" where where 12-bit drops into 11 bit, data get thinner and thinner when it is at best 12 bits.
Fred Miranda wrote:
Your test shows worse noise in Bulb mode than mine. Different cameras and different exposure increase though.
Yep, it's a combination of the two. Here's another A7s raw set with the midtones separated from the shadows and a more uniform area to highlight how the exposure/tonal level affects the perception of the noise difference: https://horshack.smugmug.com/photos/i-9K9SrRt/0/O/i-9K9SrRt.jpg
tn1krr wrote:
Dunno about preferable, for highlight recovery at least IMO not. In the 12-bit mode Sony CRAW bit depth is alwaus <= 12, it is 12 bits for luminance values 0 - 2688 and then it is 11 bits until luminance 5184, 10 bit from there to luminance 7580. Explained here in detail.
The 11-bit encoding varies from 13-bit in the shadows to 9-bit in the extreme highlights, the question is whether that degree of posterisation in the highlights would ever be visible in the image. Potentially, yes, if you had a very low DR scene and stretched it out. But in general I'd much rather have more resolution in the shadows.
I'd rather have Sony's 12bit RAW with >14 stops of DR than Canon 14bit files and their rather unspectacular DR limits. If astro photography is your specialty then I agree that Sony may not be your best tool but for everyone else this is looking like another forum thread filler of no relevance to 99.9% of real world shooting situations. If you need to rescue shots underexposed by 7 stops than maybe don't blame the camera.......
Matt Grum wrote:
The 11-bit encoding varies from 13-bit in the shadows to 9-bit in the extreme highlights, the question is whether that degree of posterisation in the highlights would ever be visible in the image. Potentially, yes, if you had a very low DR scene and stretched it out. But in general I'd much rather have more resolution in the shadows.
This is only true for "non-reduced" bit depth files, for "reduced 12-bit mode" (bulb, silent, continuous, Long Exposure NR) RAW Files there is no 13 bits for shadows, it is 12 bits for shadows, then 11, 10 etc. bits as luminance grows as I explained above.
shirozina wrote:
I'd rather have Sony's 12bit RAW with >14 stops of DR than Canon 14bit files and their rather unspectacular DR limits. If astro photography is your specialty then I agree that Sony may not be your best tool but for everyone else this is looking like another forum thread filler of no relevance to 99.9% of real world shooting situations. If you need to rescue shots underexposed by 7 stops than maybe don't blame the camera.......
To my eyes Sony 12 bits RAW in Bulb or Silent shutter mode is almost identical in DR to the Canon 5DS. If you can live with Sony 12 bits files, then the DR advantage over Canon is gone or unimportant.
Fred Miranda wrote:
Your test shows worse noise in Bulb mode than mine. Different cameras and different exposure increase though.
Although it seems that all A7 series bodies have deficiencies with these modes, perhaps the camera used may make a difference. My test was with the A7R.
Here is a test with the new A7RII where there is almost no change between the EFCS vs Silent Shutter captures:
Keep in mind that the poster didn't push exposure.
This is a great post! Following up on the quick test i done last night i pushed the Exposure to +5EV and the difference between the mechanical shutter (on the left hand side) and the electronic one still is remarkably small.. I don't have the technical knowledge to know what they have done in order to obtain this results but as a user i'm seriously impressed.. :-)
