Sony A7r II – Buffer Test with compressed and uncompressed RAW
Sony A7rII firmware 2.0 is out (Download here for Windows or Mac). I checked the first tests done by the community and so far that’s what people noticed when shooting uncompressed RAW images:
1) You get 80+ megabyte files
2) Buffer size is much less (You can shoot around 10 frames of RAW until you reach the limit on a fast card).
3) Image quality difference is only visible in extreme conditions (when you push files at +5EV). Artifacts in high contrast situation are gone.
More test links:
Sony A7R II Uncompressed RAW with Adobe DNG Converter (Photographylife).
Sony a7r II uncompressed raw file support in Camera Raw and Lightroom (Adobe).
First impression by Ricky Ban.
Bulletproof Guide to firmware upgrade by Brian Smith.
First comment by Diglloyd.
Comments on SonyAlphaForum.
For German readers -> [shoplink 41753 ebay]Batis 85mm in Stock at Foto Siegl eBay[/shoplink].
A7rII store list:
In USA at Amazon, BHphoto, Adorama, SonyStore US, FocusCamera, Uniquephoto.
In Europe at Amazon DE, Amazon UK, Wex UK and ParkCameras.
In Asia at Digitalrev, CameraPro.
92 Responses
I found the same. Though I’d question why anyone would shoot uncompressed RAW w/continuous shooting enabled.
That said, I think the author meant 80mb and not megapixels 🙂
Just to give you a usecase : timelapses are sometimes shot in continuous in order to keep the gap between frames as small as possible and give smoother movement of clouds, stars or whatever you are shooting. But it only makes sense if you manage to get an exposure that is longer than the time it takes to write the file to the card.
Sorry to be flippant, but you’ll need to explain that to me.
Isn’t timelapse with the smallest gap possible called ‘video’?
This is new to me as well, and I’m pretty well educated when it comes to time lapse.
Shooting in continous is just the easiest way to minimize the delta of interval and exposure time. It is not like this is a new idea made up by me. See
http://forum.timescapes.org/phpBB3/viewtopic.php?f=40&t=1871
on the topic of intervals.
That could be one way of describing a video on a more abstract level and the product of taking timelapse stills is of course a video. Usually however videos play at the speed that events are perceived. This dictates the exposure times of a video frame to be in certain boundaries that do not apply to timelapses.
I am not saying that every timelapse looks better with small time gaps between frames (there are a lot of cool miniature timelapses taken with tilt-shift lenses that use the gaps to convey a certain artificial look). But if you aim at having smooth motion it is imperative.
I understood the idea of timelapse was to show movement of an otherwise slow moving event. By definition that means the gaps between frames are less relevant because so little has moved.
Certainly, for moving starfields and night effects the exposure time will be longer, but that still has no need for small gaps between frames.
In scenes you often have more than one type of movement. See the first and second shot of (did not have enough ND Filters on me)
https://www.youtube.com/watch?v=VLqxQqzpfpU
Most of the time it just looks nervous when the gap between frames is too big.
I’ll accept that there are separate issues for timelapse that don’t concern general photography. However, I think it far fetched to suggest that you would want to use this camera at full resolution to create timelapse video.
Even allowing for the buffer, you would need huge memory cards and produce a humongous video file! Plus, once it was complete, you wouldn’t have a screen available to show it on.
Umm, I do plan on using this very Camera in that way(should be able to afford it after new years). But I think you and I have a fundamentally different understanding of what a timelapse should be. If I spend a whole evening / night out in the field and burn through 2-3 sdcards I am usually very happy to get 10-30 seconds of final video content from it. As I implied in the first post buffer does not concern me for shooting timelapse content because I am always aiming at exposure times longer than sdcard clear time with this approach.
But why are 11+7 compressed files for these kind of works not sufficient? Hello, jpg looks plenty good for the job.
Yes.
No time to read it all, but the only instance of the word ‘continuous’ in that article is in a section that suggests that clicking away in continuous mode might be ‘overkill’. I have shot thousands and thousands of time lapse frames, and I’ve never done it without an intervalometer. One frame per second is the fastest interval I’d want anyway.
This, by no means, invalidates my argument, that less gap between frames gives a smoother motion in the resulting video. It is the same reason videographers try to keep exposure times no shorter than 1/50th of a second for 24p clips – The reason why they always want built in ND Filters in their cams.
Edit: And in order to save you some time: This is the excerpt I was referring to on the link I gave to timescapes forum:
“Intervals (time between the start of each shot) where do I start?
-The easiest ‘one size fits all’ way around intervals is to get them as ‘tight’ as possible. “
Sorry, but this is bad advice, and you’re confusing things. If I were to shoot fast-moving cars, a short interval is indeed warranted. But the equation changes once the movement in front of the lens slows down. I can easily put a 10-second interval between two exposures of slow-moving clouds without the final video looking jerky. For Milky Way shoots I generally use 30-second intervals. I’m currently creating a time lapse video for a construction project and I use 60-second intervals sometimes, because otherwise I’ll be looking at clips for 30 seconds without hardly any movement.
Videographers use slow shutterspeeds to allow for smooth transitions between frames, but that effect can be achieved with any interval, though generally it’s more of an issue with faster moving subjects which would warrant a shorter interval.
What stops you from raising the exposure time of your construction site example to the interval time? I am not confusing anything here. And yes sports broadcasting tends to use shorter exposure times to avoid eliptic balls and similar effects, but if you want to have smooth motion it is best done if exposure time and interval almost match.
I am still not saying that this is the only way to shoot a timelapse! It is just a possibility you can choose.
You mean pushing 60-second exposures with a 60-second interval? I haven’t tried it yet, but I doubt there’d be any movement left other than the building going up. I would still like to see the workers moving.
This is not even to mention the pain in the butt it would be to make this happen with ND filters in broad daylight… 🙂
I notice I replied to the wrong comment. Apologies.
A tool was not created to limit use. That’s more like a toy.
Sorry, but that’s a very odd answer and completely wrong in almost all cases.
You generally find that a tool is made for a very specific purpose; it’s why you have a claw hammer, a brick hammer, a sledgehammer, a pin hammer, etc.
Not all timelapses mean a video will be produced. Doing a timelapse of the movement of the stars for star trails where you stack the resultant images to create the final image requires minimum time between each shot. If you didnt do this you would have gaps in the trails. And before you ask … No you dont get the same quality of image if you use long exposures even with LENR enabled.
I’ve done plenty of timelapse astrophotography and none at 5fps. However, I’d add that the buffer would top out in continuous shooting whether a shooter is using compressed or uncompressed RAW. Which doesn’t make for a very convincing argument imo.
That said, I’d also add that most timelapse remotes use single frame and can be finetuned to avoid hitting the buffer. Which usually doesn’t pose a problem in the end.
Sort of to be expected. However, I suspect that for most “action photography”, compressed RAW is not to much of an issue given the type of shots. For occasions when you want the utmost IQ then I think those will be slower, more considered shots. A bit of a nuisance switching modes of course.
Should Sony produce a “lossless compressed” RAW then I assume we’d see performance somewhere in the middle of these two results.
They can also buy camera using cf or xqd cards
I suggest using a UHS-I U3 card (the guy in the mirrorlessons video uses a U1 card). U3 cards has a minimum write speed that is three times faster (30MB/s vs 10MB/s) than that U1 card.
Though it does not improve buffer size, it should help with buffer cleanout time. The information I have seen is that the A7RII uses UHS-I bus internally, so there is no point in buying UHS-II cards.
As A7R II compressed files are about 42MB, and the first test shows that about 1FPS can be achieved after buffer saturation, it must already be writing out to the card at greater than 30MB/s (I actually think it was a bit less than 1FPS, but the test didn’t continue long enough to be sure). In any event, way higher than 10MB/s.
I remember someone here tested various memory cards and the actual write speed of the A7r II. Not one card above 35.27 MB/s. So that’d take a long time to clean the buffer of 80 MB files yeah.
A buffer of 10 photos should take a minimum of 22 seconds to clear, likely longer.
Great that Sony have given us the option though 🙂
I’ve just looked up the spec of the card in question (it’s no longer available), and its rates at just 20MB/s for write speed and if that is true then it really ought not to sustain more than 0.5FPS after buffer saturation on compressed RAW and less than 0.25FPS on uncompressed.
imaging resources tests buffer clearing time.
Your statement doesn’t take into account already fast U1 cards that also comply with U3 specs. An example would be Sandisk Extreme Pro 95mb/s cards. I have a few of these, of many sizes and they all write as fast as each other (tested on SD slot and lexar USB 3.0 reader) and they are labelled either U1 or U3 depending on the time when I bought them.
Therefore, it is the card itself that matters, as it has been known for a while, and not the class/classification that guarantees that performance alone.
You should read up on that. A U1 card cannot comply with the U3 standard, they would be U3 in that case. Read here before jumping to more home made conclusions: https://www.sdcard.org/consumers/speed/speed_class/
Speed Classes and UHS Speed Classes are all well documented certificated levels of MINIMUM SPEEDS. What the card manufacturers print on their cards, like 95MB/s is merely what can be achieved in optimal theoretical cases.
Home-made conclusions are pretty accurate re the real world. I have encountered various speeds on U1 and U3 cards, with writes ranging from 1kbps to 95 Mbps, both on “U3”. It’s called mislabelling, but tends to happen less with reputable companies. And a manufacturer that’s not too bothered about labelling, is not going to be too bothered if he writes “Sandisk Extreme Pro” on it, either. Other factors involved are block size, file size, random or sequential etc, so how you format the SD Card, and how you delete images can create a difference, in the real world.
Do you even sure if your card reader is UHS compliant?
I feel that my photos of wooden slices looks sharper (better) without compression, even seldom fiber of cloth, which was trapped within wood material, is visible on black backgroud – without additional artifacts 🙂 I am happy with the new mode. I found out also that lag between two photos in series are smaller, it means my camera handles quicker processing of taken shot, before getting ready to next one. It is very nice side effect 🙂 I sometimes need to retake photo as quick as possible, but I usually don’t take more than 3-5 shots in row. Of course, I use U3 cards only as was previously written. I bevelive that asking for loseless compression is smart but it need to be implemented within camera processor. I am still very surprised that they managed to provide uncompressed writing mode, i believe it is so much implemented into hardware that it may be not possible to override it with firmware modification.
Thanks, I’ll keep shooting compressed war and throw away all my pictures like I did before…
Points 1,2, & 3:
“Well no sh*t Sherlock!”
Yup, exactly zero surprises there.
Diglloyd: “Always a hassle as I prefer to sidestep the Sony root kit approach to updating camera firmware. Nikon does it right. Why do other vendors do it *so wrong*.?”
Seriously? Did I just go back in time to 2005?
Admin, why do you persist in giving this fool attention? I’m even angry with myself for clicking on that site.
The Diglloyd seems to be egomaniac, obnoxious and pedantic in this statement… Crazy whah !
Though Sony’s Updater still sucks. Why do I need a application installed and rummage around for that stupid USB cable for an firmware update? On my A77 I only had to copy the file to the SD card and start up the camera. Done. No hassle. No further problem sources.
“On my A77 I only had to copy the file to the SD card and start up the camera.”
Really? I own an A77 and always remember having to update via USB cable and computer.
No mention of being able to update by SD card here: http://www.sony.co.uk/support/en/content/cnt-dwnl/prd-dime/sony-slta77-firmware-update-ver107-win/SLT-A77
Moreover, I don’t recall any application being installed with any firmware update, merely a self-contained executable.
Well, he means the root access required to update the firmware diglloyd. com/blog/2014/20141111_1302-Sony-security-risks. html . Do you think it to be a non-issue?
Probably. As usual, he warbles on for a while without saying very much and offers absolutely no evidence of any problems caused to either him or any other user.
He is right, everybody else just loads the FW onto a card, stuff it into the camera and off you go on the update process. here you have to hook the camera to the computer, go through a lengthy install process that takes a LONG time relative to anybody else. It’s an arcane tethered way of doing it.
What is the filesize if i shoot uncopressed RAW and convert it to a DNG?
Basically, for almost all use, this is a non-issue.
First, irrespective of the camera model, if you need to shoot strings longer than this then you have serious deficiencies in your photography and technique.
Second, even if you do believe such shooting serves a useful purpose, do it more than a few times on an assignment and you’ll have such a huge catalogue of shots it will take you a week just to do a simple review.
If I cover a whole day (i.e. 14 hour) wedding then I might shoot up to 1,000 frames and I regard that as possibly excessive. Fire off like this sample video and you could quadruple that!
The simple fact is if you are looking to make best use of all that resolution it is a different type of photography than journalistic ‘gun and run’. The two things have separate and distinct needs, so even thinking about it as an issue is ridiculous.
As one simple example, if I’m shooting a Cheetah chasing a Gazelle across the savanna to a possible kill then I most definitely want a bigger buffer so that I can shoot several seconds before the buffer fills. Generalized statements like “if you need to shoot strings longer than this then you have serious deficiencies in your photography and technique” just shows your unfamiliarity with any type of photography other than what you do…
Get a video 🙂
Not even close to the same thing, single frames from a video, are WAY too blurry for serious still extraction on an animal running at 60MPH. And then there’s the fact that even 4K video results in only an 8 megapixel still frame.
Then get a sports camera. – The right tool for the right job.
You are completely missing the point, the original poster said “if you need to shoot strings longer than this then you have serious deficiencies in your photography and technique” – a sports camera does exactly what the poster says is a “serious deficiency in your photography” so by his logic, sports cameras are for people with serious deficiencies.
To be fair, people can and have shot plenty of action wildlife using single frame camera’s over the years(think film). And so the notion that a cheetah chasing down it’s prey could only be captured with a high speed sports camera won’t likely present a convincing argument. What it does show however, is the convenience and benefits that come with the use of higher performing equipment/camera etc. 🙂
Now photographers do not need to spend additional time and money waiting out the cheetah and getting the right shot thanks to higher FPS, burst and better AF though. So it brings the cost of getting that shot down.
No, he said if you need this AND lossless raw… I don’t see why the 11+7 wouldn’t splendidly work on this safari use case…
Is the A7r2 fast enough for this considering you’ll probably need a non-native long telephoto lens?
You’d be surprised how close you can get to action like this in some parts of Africa. But yes a faster frame rate would be desirable but 5 FPS can certainly work. It was just the first example that came into my head as to why somebody would want longer bursts in response to the original post.
I see, I would love to visit Africa one day!
I was just wondering if the A7r2 would work to shoot fast action, yes, my question is unrelated to the buffer size, but you’ve mentioned a cheetah 🙂 Theoretically an OSPDAF system should better track moving objects since it doesn’t need to switch between the PDAF and the sensor, but IRL I’m a little frustrated with the AF performance…
It’s not the best camera for the job for sure but it certainly would work 🙂 In my example you are panning horizontally so the subject to camera distance isn’t changing that much and the light is generally bright so it can keep up. With more complex movement and low light it would likely have some issues. The first time I ever went I used an EOS 10D and got many great shots of running animals, the a7R II has better AF and tracking capability then that camera had and about the same FPS so yes it is very possible, but, like I said, not the best. A D4s or 1Dx would be better but then of course you are getting lots fewer pixels too. A D810 has about the same FPS and makes a superb safari camera though 🙂
I would disagree with you in this example.
If you have sufficient experience of the particular animals then you will be in a much better position to gauge the relevant parts of the chase and shoot accordingly. To suggest that you need to fire off from start to finish to make sure you get the shot just confirms my point about lack of technique.
The art is in choosing the right moment. If you think otherwise I’d have loved to see how you got on with 35mm…
How people once got on with film is irrelevant. Everything has changed. If a wedding photographer shoots as little as film rolls allowed, they would not get hired anymore. If press photographers took a day to turnaround their photos to the wire, they would never be called again. Just because some photographers were able to produce quality photos at 5fps with 36 exposures per roll doesn’t make current photographers less able because they now have the option to shoot 14 fps. The hit rate goes up for photographers of great skill just as well.
I wouldn’t disagree in essentials, but was answering the point that it was a necessity to have the ability to run sequences of far greater than 10 shots.
I still maintain that if this is your only safety net then you need to go back and learn your art!
Yes, but we are not discussing high long burst rate, but long burst rates with lossless raw. Question is, do you actually need this. The safari photos have lots of artifacts from 11+7 compression? I doubt it. And as was the point in the beginning, you can have either or now. Not enough? You need lossless safari shots for the art exhibition at night with high contrast? The point is, some people always find something to complain.
I am all for timing and generally don’t shoot in machine gun mode but in my example, there are many moments in time during a sequence like that that can give you awesome shots with both animals in the air stretched out as they are running, it’s not just the final seconds. Regardless, the point is that there are many situations and many ways to use a camera and to say that if you should never need anything more than this or you are somehow deficient is just not a fair assessment.
Perhaps I’m being a little harsh, but certainly not to the degree of the criticism of the buffer filling after 10 shots!
There are so many ways to circumvent it, I just don’t get why it was even tested.
I suppose until we get a camera that can take 250mp shots, with immediate AF that never misses, at 50 fps for as long as we want, can automatically adjust the framing until artistically perfect, in any light with crystal clarity, and costs no more than $50, there’ll always be something to complain about…
True Dat! 🙂
So what software is everyone using on a mac to see these huge files. I’m on a mac 10.11.1 and I can see the thumbnail but the file is garbled or black after loading into capture one, aperture, photos, intensify…etc….
Adobe CC can handle the new files just fine.
CS6 works fine too.
I put the new files into Capture one last night and they were unable to be read. Adobe PS seemed fine.
Interestingly, I put the files into Photomatix, and they were all purple.
Same result here with Photomatix but if you just convert the files in the bracket to TIF first using Adobe, then you can still use Photomatix.
So.. considering i have adobe LR / PS, Capture One, and Photomatix at my disposal (on my MBP w 16gb of ram, running 10.11)…
I’ve been converting to DNG to manage in lightroom. For paid gigs, I keep the ARW on an external drive. I’ve been doing research telling me to stop converting to DNG so I can get the most out of ARW in Capture one / Photomatix, but i honestly use these sporadically. What you gain in uncompressed ARW seems minimal at best / only in specific use cases.
Anyone have thoughts (or links) on workflow / managing these files with these programs? Im thinking that, if hard drive space is no object, I should shoot in uncompressed and archive the originals to an external drive, and then convert those to DNG too keep the size down and store on the usb3.0 flash drives that i carry with my MBP and manage the workflow in lightroom / PS. then, if im super compelled to edit a photo in the other programs i can tap into the original ARW.
I agree that ill never shoot in burst mode so that isnt an issue. And even with the larger file size, a 64gb CD card will get me through any event im covering. So its really just file management and maximizing my ability to post process with the best results.
“Im thinking that, if hard drive space is no object, I should shoot in uncompressed and archive the originals to an external drive, and then convert those to DNG too keep the size down and store on the usb3.0 flash drives that i carry with my MBP and manage the workflow in lightroom / PS. then, if im super compelled to edit a photo in the other programs i can tap into the original ARW.”
Seems like a reasonable approach if you are concerned about file size on your flash drives.
https://photographylife.com/sony-a7r-ii-uncompressed-raw-with-adobe-dng-converter
Many people including myself have said that Sony should have larger battery.
But newer much more powerfull battery technology is just around the corner.
Sony should introdeuce as soon possible..a new tecnology battery of identicle size to replace a7 series batteries in current cameras….without the necessity of getting a new camera.
An interesting new product
Amberella has just anounced a 64mp, 11mm x11mm sensor and processor specifically for drone cameras.
Very low power consumption of 1 watt is particularly usefull in lightweight drones.
Not sure I see the relevance to the topic at hand – the new firmware???
If there was a current article posted on Soy batteries and latest sensor developments I would have posted it there insted , but the only current chioces this morning were this or the Lieca camera post.
I like to share current thoghts while they are fresh and do not know when a better post will be available.
I get that but I’m just thinking if everybody did that then there would be no logic or continuity to any topic that pops up on this site – of course I’m making the assumption that there is any logic to any of this rumor site stuff to begin with 😉
This was not intended to offend you or anyone..if so just ignore it.
Certainly it cannot be as bad as pop up adds 🙂 ;<) &*}
Not offended at all, just trying to figure out why people sometimes put completely unrelated things into threads, you clarified why you do it 🙂
Indeed, and history tells us that Barry will now be posting about battery technology at every opportunity for the next three months.
This comes up time & time again, why on earth don’t Sony or a third part manufacturer simply make a grip, one that takes an adult sized battery.
A7II could use a FW update too, what gives? More aspect ratios and uncompressed raw please. Not a lot to ask.
Not bad at all. How long does it take to recover?
Write speeds are quite acceptable on a 95MB/s card.
ie, I was able to shoot single frames (in sequence) at/or around 1fps with the occasional bust at 3fps without ever hitting the buffer. Which is usually enough to shot time laps and various types of landscape shooting(sunsets etc).
Doesn’t A7Rii max out at around 30-35MB/s write? I don’t think any faster card would make a different before or after this firmware update.
How has it affected posterization? Particularly in gradients of blue like sky during sunset. I’ve had trouble with that even in RAWs.
Next super duper killer feature in the upcoming A7iii: lossless compressed RAW files. Like everyone in the industry had from the beginning. ROLLEYES
Price: $2199.00
And the really mind blowing feature in that camera is firmware that is uneatable via an SD cart rather than having to tether the camera to a computer. 😀
It’ll be via wi-fi from the Sony App store…….. Have you tried gaining access to the Sony App sore via camera wi-fi, it appears to be some form of “enhanced-intimidation” cooked up by their “black opps” division.
FYI Sony, Apps should be free, and signing up via camera wi-fi, entering password and hooking up with CC details makes the CIA envious. Just Sayin’
Coming from Canon, I am just so amazed at how well and quick Sony listened to it’s customers who bought the A7RII. Right on Sony! The 5DSr does 15 frames raw and A7RII does 10 Raw uncompressed, who cares!! 10 shots is all you should need and ALL the data is in there for all the pixel peepers with microscopes. Now lets all go make some awesome content instead of whine.
I think everyone wanted lossless compression, not 80MB files.
Please folks – do a bit of research before repeating what others have already stated a zillion times.
The reason why it is not compress (and looseless) is most probably a HW limitation. The BionZ X most likely have embedded HW support for the 11+7 compression scheme, but no support for other compression algoritms. So while it IS possible to read the data “raw” the current generation of BionZ X cannot compress it.
Sony are stupid. If they can find a way to do it compressed they probably will eventually. Other issues like the terrible battery life is much more important (I’ve have 6 for my A7R2 now :)).
I did talk to the folks at Photomatix today and they were not aware of the problem with the uncompressed files and their software. I have provided them RAW files so that they can work it out and issue an update.