Sony FDR-AX100

[Dave Blackhurst]

OK, I looked at your video (Steve) on my new computer... I see some "pulsing" in the one bush as you zoom, at 4k (on a 4K 30hz screen), but not so much at 1080... still there, but not as bad... then on my 1080 60hz auxiliary screen, barely any when playing 4k, and only a tiny little bit for the 1080... certainly I only noticed it because I was concentrating on LOOKING for the flaw (I believe this would be akin to looking at Cindy Crawford and ONLY seeing a mole, but we are being picky).

I also fiddled with the YouTube speed controls, and that affected the severity of the "pulsing".

I'm not entirely convinced the problem is due to over sharpening so much as this is the result of 30 progressive frames - each time I paused, I got a nice sharp crisp image, like a still.

Let's step back a bit and consider... 30p was the ONE reservation going back to the announcement of this camera. The more I look at samples on the 4K system I just built (integrated intel graphics, on a cheap seiki 39" TV being used as a monitor), the more I see a bit of an issue with 30p. The Seiki is a 30Hz refresh, and it will also display at 24Hz - 24 is noticeably "jerky", and there is a degree of "shimmer" at 30... scrolling has a funny "motion" about it, and I have to do some more testing with video, though generally it has looked quite GOOD. I now know what I should be looking for in re 30p temporal distortion...

SO, does this make the AX100 a horrible camera? Not IMO, it makes it a camera with some potential limitations when shooting certain types of movement. It also has the XAVC S 60p 60Mbps 1080, which shouldn't have the same temporal limitations, easy enough to switch as required, I expect to use that mode quite a bit (OK, so maybe I'm making an argument for the CX900...).

I think what we are seeing is a limitation of an early implementation of "4K", the detail is there, but because that detail is so "detailed", any movement at only 30 frames per second becomes problematic. SO, perhaps 4K will need to wait for 60p to be really "great", and for now high bitrate 1080p is a "better" choice (and the AX100 has it... I would have been looking at it for that feature alone).


This raises the question of how to best reduce the "shimmer" or "pulsing" or "stutter" resulting from temporal/motion without messing up the "detail" - not sure yet what, if any, the answer is to that question...

I've been fighting to stabilize this new build for over a week before I even contemplate edit/render... I'm anticipating another set of challenges when reaching that hurdle... but after a week of tweaking, bug hunting (think I squashed the last one in the wee hours this AM), and adjusting to get a 4K system running on a "lo budget", I'm not TOO worried (yet!).

I know Ken mentioned adjusting his display to smooth things out (ain't gonna happen on my "cheap" monitor/TV, it is what it is!). But if I can make it look "good" on a cheap TV, I'm hoping for REALLY good 1080p in the end, and 4K for a few things, and I think the AX100 will deliver that, but I'm realistic that there will be limitations...

On another "front", I strolled through the local Costco, and they had exactly ONE 4K TV in 4-5 rows of 1080 HDTVs... yeah, ONE... so it may be a while before 4K becomes "standard" enough to be a worry! I think Best Buy has a little better selection (think I saw TWO last time at the small footprint local store), but still, 4K won't be "mainstream" THIS year....

Life on the bleeding edge has it's moments, this looks like one of them...

[Dave Blackhurst]

I forgot to mention that on that Costco excursion, in the few brief moments I was watching, I saw one clip with a "host" in a plaid shirt... do ya think that I saw a "few" artifacts??? Yep, that shirt had a life all it's own... and I'm sure the clip was shot with something a LOT more expensive than the AX100, no doubt edited on a expensive system and a big budget...

I see plenty of artifacts and noisy video on my HD FiOS feed - it is the RARE show that really looks "good" with consistency - watching the new "Late Night", and there's some "pulsing" in the brightness... I'm sure it's shot with state of the art gear in a brand new studio... and it looks QUITE GOOD, much better than most "HD" that seems to come in, but it's not without flaws should someone choose to be "picky"! And this is just the stuff that's supposed to be "HD".

There is an element of expectations at work here. We tend to see flaws after we notice them or someone points them out... I call this the phenomenon of "that which once seen cannot be "unseen". But if this means going through life with an "empty glass", I'll pass on that, thanks. Tech has limits and flaws, you either accept that or you don't, and wait in hopes of finding "perfect"... which may or may not ever come along...

You can enjoy (or at least make the best of) the trip or complain that you're not at your destination...

[Mark Watson]

Originally Posted by Glen Vandermolen

Timoecode? Does it have one?
Or does the counter reset every time you stop-start roll?

Already answered, no TC on this camera but I've gone a step further and made some text files using MediaInfo program, which analyzes video files and lists all sorts of properties. The XF305 has TC, so it's included here to show how MediaInfo identifies that in the video file info. I've also included some files from video files which were taken from my BMD Shuttle 2 when it was connected to the XF305. So, you can compare the data listed from the FDR-AX100 and see there's no mention of TC.

Hope this is helpful.

Mark

[Mark Watson]

While I'm at it, may as well include file data from a few other cameras I have...

Interestingly, the JVC GC-P100B does output on the HDMI during record, but I see no indications of TC from it.

Mark

[Steve Mullen]

Originally Posted by Dave Blackhurst

OK, I looked at your video (Steve) on my new computer... I see some "pulsing" in the one bush as you zoom, at 4k (on a 4K 30hz screen), but not so much at 1080... still there, but not as bad... then on my 1080 60hz auxiliary screen, barely any when playing 4k, and only a tiny little bit for the 1080...

I don't have have a 30Hz monitor so I can't comment on 30Hz. (One reason I shoot at 24p.) Computers and my HDTV are 60Hz. (It won't go faster because I don't want film to look like video.) I don't know what the Best Buy UHDTV was running at, as there was no way they were going to let me turn of all the features

What you find matches my experience. On a computer the pulsing is the worst -- from a 4K 24p YouTube source. A little less on my HDTV fed via HDMI -- from a 4K 24p YouTube source. When the HDTV is fed by AppleTV which I suspect is getting a 720p24 stream, it's not bad at all. Other people who see artifacts have big Apple monitors.

The problem is that I can't predict what anyone will see when they watch. I have no way to QC my video. And, let's suppose it suddenly shows up on a 4K TV I buy in a few years? I'm screwed because I'll have edited movies!

How bad is it? Folks in their 40's said it was terrible. Someone way over 80 saw nothing until I described it. She then said I can't "un-see it!!!" Same with the Morie on many of the buildings in the NYC video. She never saw it. But, once she saw it ONCE, she could see it whenever it appeared.

And you are correct -- it appears on live HD TV because they have no control over talent. But, I've never seen it on a primetime network show or on HBO films. And, if you watch HD shot on the BMPCC you rarely see it -- unless someone has over-sharpened the RAW video. In other words, when video is shot with pro cameras it is absent. I didn't see it from the AX1.

The comments made on Slashcom point to "over-sharpening." (In general, compression errors show-up as a loss of detail on motion. But, if they are from low compression -- that doesn't really help you.) Because each picture shows an alias in a different locations you see them only when there is motion--which why speed affects what what you see.

"This raises the question of how to best reduce the "shimmer" or "pulsing" or "stutter" resulting from temporal/motion without messing up the "detail" -- not sure yet what, if any, the answer is to that question..."

Those who are expecting my review to be really negative may surprised that after noting what I saw, I spell-out how to shoot to never have artifacts. It can be done as one of my posted videos show no artifacts.

But ultimately I realized that I didn't need to buy the first 4K camera. As you said, there is a lot to sort out. So why spend $2000 now? Since the 2014 Sony's coming out this month will support 60p, I suspect we will soon see a 60p camcorder. (Adriano was very worried about thist.)

The other issue you are dealing with -- what computer is needed. I'm not switching from a MacBook's. In a year the GPU and HDMI port issues will be sorted out. (I did know 10.9.3 would be needed. But, what else don't I know.) So I'm going to hang in with HD for another year. I think CES 2015 will have many exciting 4K products -- and many will be cheaper.

"Life on the bleeding edge has it's moments, this looks like one of them..." You are so right! I'm just not willing to bleed having had the first DV and then the first HDV camcorders. And, my sense was that Adriano was very worried because in most parts the world there is NO trial period. If he is not to lose a lot of money, he needs to sell quickly. There will always be another 4K camcorder -- maybe at IBC in September.

Bottom-line, if someone doesn't see anything with their setup they have nothing to worry about. Those who see artifacts, it is possible to shoot to avoid them. As long as it works for you.

Dave, I wish you luck with your PC. I've built one and I loved doing it -- at least until it didn't boot on first power-up.

TIP: Hammacher Schlemmer is selling the AX1 for $5000. They offer a 1 YEAR trial.

PS: Your 4K 30Hz monitor must show each frame twice or you would get sick. Just like watching a Utube 30fps stream on a computer at 60Hz. Can you tell your graphics card to send 30Hz? I'm not sure on a Mac that I can set 30Hz.

[Cliff Totten]

I have noticed this "twitter" affect too. I have seen it on BOTH the AX100 and the GH4.

It seems to be most noticeable on the very highest frequency details. (tiny and dense bush leaves are a perfect culprit)

I'm not sure, but I suspect its a very very fine granular moire combined with a fast shutter speed.

At 30p frame rate with a slow shutter of 1/30 or 1/60, high frequency details is "smudged" very VERY quickly. Imagine a 4k resolution chart being shot at a 1/30 shutter speed....even the smallest and most insignificant movement will blur the 1000+ line mark and get worse as it goes up the scale. 1500 lines on a chart at 1/60 shutter will smudge extremely fast with even the smallest movement. The 250 line mark on the scale/chart can survive the same slight movement at 1/60 shutter because those lines and much further apart.

So, raising the shutter speed can decrease the blur and sharpen the lines further and further up the scale with that same tiny motion amount.

So this is where that high frequency aliasing appears to come from? It's just my best guess. I would also suspect that all 4k cameras at fast shutter speeds will display this. (As i said, I saw that GH4 doing the exact same thing.)

To sum it up: There is a tight relationship between "motion", "resolution" and "shutter speed". They all have a big affect on each other. A drastic resolution increase also forces significant affects on the other two.

But what the heck do I know? Hahaa...I'm just spit ballin' here! ;-)

CT
.

[Ken Ross]

I'm sorry guys, much ado about nothing. As usual, Steve in on his endless mission to find something, anything, wrong with the AX100. I've seen tons of AX100 videos and rarely have I seen anyone commenting, let alone so bothered by artifacts.

The line twitter, pulsing, dancing, shimmering, whatever my buddy Steve would like to call it, is something I've seen on virtually EVERY camera that shoots some highly detailed subject with the sun hitting it just right. I've seen it on MEGABUCK professional cameras, I've seen it on live broadcasts as Dave has mentioned and I've seen it on documentaries. So Steve has discovered that a $2,000 4K camera that does so many things so well, can also manifest this issue under certain circumstances? Big whoop!

Then to call my suggestion to Adriano, that he hook up his camera to his HDTV for the purpose of determining if he sees the same kind or intensity of artifacts, 'BS', is not only insulting but demonstrates a huge lack of understanding on Steve's part in tracing down problems. Did I suggest to Adriano that he watch his footage day in and day out in this manner? NO. Did I suggest he do this a good first step in determining where his problems might lie? YES. Steve is apparently not well versed in the scientific method.

Can editing introduce artifacts? YOU BETCHA! For anyone to suggest otherwise, is again very naïve. Wasn't it Steve that was all over the AX100 a while ago for a different artifact that he later found out was his error in editing?

Can YouTube or any online streaming service introduce artifacts? YOU BETCHA! And yet there are still people blaming cameras for artifacts that are clearly being introduced by the streaming service. Hell, I've posted videos on both YouTube and Vimeo and it kills me to see not only the artifacts that can be introduced, but the general degradation in quality. The difference couldn't be clearer when you have the original file at your disposal. But to blame the issues on the camera? C'mon, pure folly.

Does this mean the AX100 is devoid of artifacts? Of course not. But golly gee, let's put things in perspective. ALL cameras have artifacts and yes, despite what Steve said, I've clearly seen them on original, unedited BMCC files. ALL cameras have artifacts of one kind or another. To then say that a camera that costs 2X or 3X the price of an AX100 has less of a given artifact (IF true), so what? Is anyone naïve enough to expect the same features and quality from the AX100 as cameras that costs double, triple or more of that of an AX100? What's amazing is that the AX100 can do better in many respects than even those cameras! The AX100 does produce a picture with more detail than a BMC 4K camera, GH4 or an AX1. That is nothing short of amazing.

As for avoiding artifacts, I don't need to read a book to know how to avoid them and I don't need to see issues overemphasized for the purpose of selling a book. I think most people here can figure it out in a very short period of time.

As for Slashcam, they had stated (despite Steve's repeated misinterpretation of their tests), the AX100 was the most resolute 4K camera they had yet tested...including the GH4.

So rather than spend one's life searching endlessly for issues that can be found in any camera, here's a novel idea, shoot real video, make real projects and enjoy what you shoot. But if you insist on looking for artifacts, they can be found on any camera at any price.

Let's make something clear here and it's very important and flies in the face of Mr. Mullen's 'logic'. When you shoot and edit a video, if it looks perfect, or nearly so, on a calibrated monitor at the correct frame rates (24p, no thank you, talk about issues!) & resolution, then that video is indeed fine. If someone chooses to watch it on a lousy monitor, in a different country at a different frame rate or on an Ipad, a cellphone or whatever, whose fault is it that the video now deviates from what we know it should look like? There are endless possibilities that are introduced at the final end point of the chain and there is little that can be done. If someone watches on a poorly calibrated monitor at a default temperature of 10,000K, is it the editor's fault or the monitor's fault that it looks too blue? But to blame this on the camera? C'mon.

I just shot this AX100, 4K video, a few days ago. This was not another series of test videos to try to induce issues, I've long since stopped with 'test videos'. This was a video that simply documented a visit we made to a location that was, IMO, worth shooting. It was not meant to show artifacts nor was it mean to HIDE artifacts. It was simply a real video of real subjects. Novel idea, huh? With that said, and my nose pressed up against a 28" Samsung 4K monitor, I see zero artifacts...and this is an edited video. From my point of view, the biggest flaw in this video was my fault. We were so surprised to see this bald eagle, I apparently paid no attention to the zebras on the eagle's head. Thus the eagle's head is over-exposed. I wanted to knock my head through the wall when I saw that. Now THAT'S an issue, not some minor twinkling in a bush and that was purely my error.

I'm sorry guys, shoot REAL video of REAL subjects with the AX100 and you'll find out how good it really is. But hey, if you're determined to look for some kind of artifacts, you might actually find them as you would with any camera. If that's what you do, may I suggest painting? ;)

I encourage anyone that thinks this camera is so laden with artifacts to download the original file (don't rely on Vimeo playback that might introduce its own artifacts). I've shown this video to 4 different people and the only remarks I've gotten were 'wow', not 'look at the twittering of that bush in the distance'. The original 4K video is available if you log in.

The AX100 at the Queens Zoo on Vimeo

[Dave Blackhurst]

@Ken -

Considering the "average" person NEVER calibrates their TV/monitor... it's crazy to expect the "end user experience" to be anything BUT "highly variable"! I've got 3 HDTV's and they all look a "little" different, even after "attempted" calibration... each brand put their own "look" into the recipe!!

I'm still not convinced I've got this Seiki dialed in either, I had to make MASSIVE changes to the factory settings to get "close". Though VIDEO looks excellent (maybe a tad dark, but really good)... stills and SOME websites have some truly strange "issues"!

AT most, I think that the potential for certain scenes to "maybe" have more potential for artifacts is "handy" to be aware of, but it's not a deal breaker, especially on a camera with a 1080 60Mbps 60p "option" at your fingertips! It's sort of like when you have a bright white "spot" on a dark background <wink> - you might have to adjust things a bit!

And when it comes to all the OTHER places in the "workflow" and display chain where things can go "off the rails", it's absolutely crazy to point at the "new kid" (camera) and say it's the one that's "defective"!

As you point out, one must have good "scientific methodology" and troubleshooting skills if one is to be on the bleeding edge, tech wise. We are fortunate that no wings fall off and nobody dies, but that doesn't mean that things always run smoothly and work as expected! Sometimes you have to "work at it".


@Steve -
As for the new computer build... it did NOT boot on the first try (bad RAM, one module out of 4, horrigble sinking feeling included), it's had some stability issues (now traced to a bad connector in the power supply cabling), 30p/Hz is what the Intel integrated graphics output, PERIOD (see a pattern with 4K?), I suppose there are dedicated cards that would do 60p/Hz for a price... but not on the "budget plan"! The Seiki TV ALSO will only produce 30p/Hz for 4K (pattern again...), and while it "should" induce headaches I suppose, it is wonderfully sharp and quite easy on my eyes for a change!

From my price shopping, I'm guessing if I wanted to spend 5-10x the budget of my ENTIRE upgrade for a monitor, I might have a 60p "monitor" option... but I wanted to not break the bank! I have a "little" 21" HD Samsung monitor alongside for reasons of a driver "issue" (without multiple monitors the driver keeps reverting to 29Hz instead of 30, which sits there blinking until manually adjusted between "blinks"), and it looks like it will be handy for displaying 1080/60p video as a comparison.

Philosophically, I expect "cheap" (relatively) gear to do a passable job, maybe not as good as "toys" that cost far more, but "good enough". I'm not seeing the "failure" of the AX100 in that respect. It is what it is, and much like the RX10, I expect it to do what I ask of it quite well. I got a good price on mine, and will sell a couple other cameras to offset, as "pan and scan" should allow it to replace at least 2 cameras. The RX10 caused a similar clean out of the equipment locker...


I don't know which "2014" Sony cams you're talking about, EVERY higher end Sony (Handycam AND Cybershot) I've got and have had, for the past 3 years or so, has shot 1080 60p, at 28Mbps.... The just announced RX100Mk3 is adding XAVC S @60Mbps (same as the CX900/AX100). The AX100 IS a 2014 camera... and I don't think based on what's released and rumoured that there are that many more "consumer priced" cameras coming down the pipe (much as I hope for an RX10Mk2 with 4K...).

Next year, I expect (depending on how the AX100 does), we'll see more 4K capture devices, if for no other reason than they HAVE to keep up with cell phone/tablet images!!

2015 and on, I suppose "60p" 4K will be the "buzz", but is there ANY stuff (camera/computer editing systems/TV's) out there RIGHT NOW at "consumer" prices that's 60p/4K??? <crickets>

SO, that's what makes the AX100 "unique", warts (it any) and all. You want to shoot 4K on a BUDGET, this (and maybe the GH4 if you've got investment in the lenses/system already) is THE GAME... or you can sit on the sidelines waiting for when we are talking about being able to buy a 4K60p camera for $500 and a 55" 4K60P TV for $500, that's fine too... you'll be sitting waiting a while!

[Steve Mullen]

Originally Posted by Cliff Totten

I have noticed this "twitter" affect too and I have seen it on BOTH the AX100 and the GH4. It seems to be most noticeable on the very highest frequency details. (tiny and dense bush leaves are a perfect culprit) I'm not sure, but I suspect its a very very fine granular moire combined with a fast shutter speed.

It could be shutter speed with some folks, but I shoot at 24p at 1/48th-1/50th. Exactly correct.

You may see it with the GH4 on some samples because it has a Sharpness control and so you never know what it has been set at. Actually, you can see it on HD content when it's been over-sharpened or rescaled without first being passed through an anti-aliasing filter.

It really doesn't matter if one calls it moire, because moire is simply a specific type of aliasing that occurs on closely spaced horizontal (typically) spaced lines.

But, this is well understood and there is no reason for me to explain it here.

What's good is that this thread has now switched from "I see nothing senior, nada" to "I see it but it doesn't matter to me." For some it may be fine to buy and use a 4K camera that can create aliasing even though other 4K cameras do not -- and this issue is not a problem with the HD cameras we own.

But, there are those of us who will not buy a camera that can create artifacts when others view their video. Hell, I won't watch my video with aliasing. We won't because there is no need to! There are other 4K cameras that have an industry wide professional recognition and there will be many more lower-cost consumer cameras! What's the rush to 4K when there are truly great HD cameras at half the price?

But that's your choice--at least everyone now knows what their choice involves. Good luck. Over and out.

[Wacharapong Chiowanich]

This thing that I have seen in the raw 24p/25p/30p 4K Sony AX100 and Panasonic GH4 footage. I'm not sure what to call it: flicker, twitter, shimmer, glimmer etc. but I guess they may all be alright as long as they tell you there's this artefact visible in the image that often occurs in conjunction of

- Motion, either the panning motion by the camera or the motion of the subject itself in front of the stationary camera

- High level of contrast of the in-focused subject(s) that is in motion in either manner mentioned above.

For some of the footage, since the shooters don't mention the shutter speed they used, I assume this artefact might be due to the shutter speed being set too high, over 1/50th or 1/60th in most cases. Now the interesting thing is in virtually all the rest of the footage where the shooters specifically mention they strictly used the 180-degree shutter rule or the shutter speeds that doubled the frame rate (1/50th, 1/60th), I have still seen this artefact IN ALL OF THEM if the scenes in the footage fall into the two criteria I mention above.

I've watched the playback of all the footage using the 4K source files on a couple of my calibrated 60Hz 1080p Samsung displays. I've tried the playback in both downsized 1/4 (1080p) and actual size (1/4 of the 4K frame) and seen the same thing.

I'm in Cliff Totten's camp and suspect this visual anomaly that seems to be unique to slow frame rate 4K image is caused by a combination of factors. The probable main culprits:

- Shutter speed. Anything over doubling the frame rate is obviously too high. Even if the 1/50th or 1/60th is used it may still not be enough. You may have to go to the absolute limit (to minimize trailing and blurring) of 1/30th for the 24p/25p and 1/48th or 1/50th for the 30p and be extra careful about every camera move you make.

- Very high delineation of in-focused subjects. On the scale of 4K, 1080p image outlines are like the same image with a slight Gossian blur applied so we see much less problem when the subjects move. This could be helped by lowering the sharpening in-camera and lowering the contrast if the controls are available.

- Too low the frame rate. I mention this because I've rarely seen this artefact in the raw footage shot with the Sony AX1, Z100 and the FS700 with the raw 4K recorder at 50p/60p. So when it comes to watching pleasant 4K I think higher frame rates are not only preferable but necessary.

Or the easiest way to get rid of this annoying thing is playing back your slow frame rate videos on a 4K HDTV set that can do frame doubling or quadrupling and hoping for the best.

[Ken Ross]

Originally Posted by Steve Mullen

For some it may be fine to buy and use a 4K camera that can create aliasing even though other 4K cameras do not -- and this issue is not a problem with the HD cameras we own.

Once again Mr. Mullen, totally false. I OWNED both the GH4 and the AX100. I have sent my GH4 back because it did not have the DETAIL that the AX100 had. Notice I said 'detail' and not 'sharpness'. There is a difference. I could see more fine detail than I could on the GH4. The GH4 is a great camera, but it presented a softer image and yet still showed artifacts.

And when you say 'it may be fine to buy and use a 4K camera that can create aliasing even though other 4K cameras do not'. Are you kidding? The GH4 CLEARLY had more artifacts (no pun intended). No, it wasn't 'bad', but it had moire the AX100 did not have or, at the very least, certainly more than the AX100 has. I have no idea where you come up with this stuff, I really don't. I compared both cameras with the full intention of keeping the one that produced the most detailed picture with the fewest artifacts. That was the AX100.

As for HD cameras, I have yet to see ANY DSLRs that did not have many times more artifacting than the AX100. None. Zero. Nada. The only one that was in the same ballpark as the AX100 was the RX10. Full sensor read out my friend, full sensor read out.

Originally Posted by Steve Mullen;1845484[/quote

But, there are those of us who will not buy a camera that can create artifacts when others view their video. Hell, I won't watch my video with aliasing. We won't because there is no need to!

Yes, that's why I gave up on shooting HD video with DSLRs even before I got into 4K. Too many artifacts. The RX10 was my salvation for HD. The AX100 is the 4K counterpart to the RX10. I have not seen people complaining about RX10 artifacts.

Originally Posted by Steve Mullen

There are other 4K cameras that have an industry wide professional recognition and there will be many more lower-cost consumer cameras!

I'm not even sure what that means. Professional Hollywood cameras? Cameras costing $50,000 or more? So what? Wow, what can I say. The AX100 already outperforms, in many respects, cameras costing far more.

Originally Posted by Steve Mullen

What's the rush to 4K when there are truly great HD cameras at half the price?

So what's the 'rush' to 4K? Well with that statement you've shown me you truly don't understand 4K or have never seen 4K on a 4K display. Additionally, you apparently still don't understand that down rez'd 4K>HD looks worlds better than the best HD cameras shooting in HD. Yes, WORLDS better. The best HD cameras are woefully short of true 1080p resolution and 4K cameras have shown that in spades. Downscaled 4K to HD lives up to the potential of what HD should have been. That's been proven time and time again.

Stick with HD Steve, you'll be happier. In the meantime, many of us will continue to shoot some amazing 4K footage.

[Ken Ross]

Originally Posted by Wacharapong Chiowanich

This thing that I have seen in the raw 24p/25p/30p 4K Sony AX100 and Panasonic GH4 footage. I'm not sure what to call it: flicker, twitter, shimmer, glimmer etc. but I guess they may all be alright as long as they tell you there's this artefact visible in the image that often occurs in conjunction of

- Motion, either the panning motion by the camera or the motion of the subject itself in front of the stationary camera

- High level of contrast of the in-focused subject(s) that is in motion in either manner mentioned above.

I'm really perplexed, because I've seen this on most HD cameras I've used and virtually all DSLRs I've used for shooting HD video. I'm not sure why people think this is a 4K or frame rate driven issue. In my experience it is not. It is simply how most cameras render a very finely detailed item with higher contrast (I agree with you about high contrast being an instigator). The threshold for where this might occur will probably vary depending on the resolution capabilities of the camera. For HD cameras, the fine detail that is seen by 4K cameras won't even be there. But for HD cameras, their 'definition' of fine detail will be at a lower frequency, but it will still reach a threshold where this issue will manifest itself. It's just that it will occur at a different point. But it will and does occur. This is simply not a new phenomena.

With that said, I've seen very few instances of this whether watching my 4K footage on a large screen UHD TV, which I've done a number of times at Best Buy, watching the same footage on my own 4K 28" monitor, or watching the down-rez'd footage on my 64" plasma.

I've certainly seen no higher instances of this issue and if anything, far less of it. I would say it's very similar to what I saw on the RX10. If one pans slowly (a good practice anyway), when there is a need to pan, you can minimize the issue.

I linked a 4K video above that I shot and invited anyone to point out the artifacts in that video. Thus far I have no takers.

To be perfectly fair, I can see one instance of this issue in 4K with my nose stuck up against my 28" monitor. But I was focusing on just looking for this issue. If I was watching this the way normal people would watch a video, actually looking at content (wow, another novel idea, instead of searching for artifacts, watch the content) I'd probably never have noticed it. If the artifacts are so significant, they should hit you in the face whether you're looking for them or not. To me that's the test of serious artifacting...not pixel peeping or endlessly searching for them.

I really think some folks just can't 'watch' a video without over-analyzing it. That's really a pity. I certainly don't like artifacts and like I said before, it was why I gave up shooting video with DSLRs. But I never had to 'look' for those artifacts, they reared their ugly head far too many times. Such is not the case with the AX100.

[Dave Blackhurst]

@ Ken -

This is what has me baffled about Steve's contentions... and it's not the first time I've seen Sony cameras criticized when there was actual visible DETAIL over and above a comparison camera, but somehow there was something "wrong" with the Sony images... and the other camera was of course "superior". I've learned that there is not a lot of love for Sony out there, even when they get it right... or mostly right (I'm the first one to scratch my head and gripe about some "weird" design decisions!)

Pause a video from the AX100, where are the "artifacts"... generally it looks like a good STILL, better than you'll typically get from video, but it's got more DETAIL. I've looked at some "4K" from the BMCC online, and it looks PAINFULLY SOFT by comparison... no thanks, not interested in a "4K" camera that doesn't look as good as the RX10 and costs lots more (and is more hassle to shoot with to boot)... just saying it's "4K" does not evidently equate to resolving that resolution, IMO.

There simply is not nor has ever been a "perfect" camera under ALL conditions - they only capture a representation or impression of what we "see", and when processed, may "improve" upon that, or make it worse. Some are better than others. Again, post processing and display methods/equipment can factor in in many ways. Some good, some not. I've seen LOTS of badly edited stuff "proving" a camera is flawed, when it only shows "user malfunction". It obviously must be the equipment, not the operator... or not.


I for one will admit that the "low" specs of the video/graphic subsystem at 4K, the limits of a cheap 4K TV, and even the limits of the AX100 (OK so all are effectively 30p...) in the "4K" system I've just assembled on the cheap... just looking at all those "30's" gives me pause. I would be surprised if it all worked perfectly and gives "perfect" image results (whatever perfect means....). SO FAR, I'm happy with the sharp image, the color is pretty close, and video playback seems "good enough". My previously shot 1080 stuff looks plenty good too. My text "stutters" when I scroll (yep, 30p "artifact"!), but I can live with it. When I experimented with 24p the mouse "jumps" all over, proving we don't "see" at that rate terribly well... 30 is "working OK", I'm happy with it in the computer/display, and can't wait to do more with all the screen real estate!

I expected some "issues", we just don't "see" 30 frames per second, any more than we "see" 24 FPS... BUT, I have yet to find footage from the AX100 to be "offensive" or fatally defective (then again I need to start shooting/editing, I'm sure I can manage to muck it up a bit!). I do see a few things that didn't show up on my 1080 screens, but again, nothing THAT objectionable.

Do I see some temporal issues under some circumstances? Yes, exactly as expected, and IF I have to shoot under that scenario, I'll switch to 1080/60p/60Mbps, thank you very much, as this is currently the only "HD" camera with that spec at this price... other than the CX900. Of course the upcoming RX100M3 will do it for around $800, with other compromises that come with that small form factor - no free lunch!


Honestly, I want to know where these "truly great HD cameras at half the price" are, I must have missed it somewhere - seriously, what other cameras (not BODIES, complete cameras with lenses) shoot 1080/60p @ 60Mbps? I think that strawman doesn't hunt...


Even more puzzling is this entire paragraph....

"But, there are those of us who will not buy a camera that can create artifacts when others view their video. Hell, I won't watch my video with aliasing. We won't because there is no need to! There are other 4K cameras that have an industry wide professional recognition and there will be many more lower-cost consumer cameras! What's the rush to 4K when there are truly great HD cameras at half the price?"

I'm not sure which artifact free camera we're talking about here, or which artifact free media and display technology the rest of us have missed? And all low cost consumer stuff no less?? Seriously, where are you getting this "good stuff", please pass it along...


In short, yes cameras WILL have artifacts or variance from "reality" to one degree or another, but the AX100 is a good first foray into "cheap" 4K, with high quality 1080 to boot, from a fairly big sensor in a good traditional video camera design... where is this so flawed??? I'll take Cindy, you keep the mole...

It's not a "rush", it's a toe in the water, at a decent price, and overall this thing seems to be a very well built camera that will replace at least a couple of others for my use, meaning the cost after selling some things will be nominal. I'm sure I'll be looking at upgrading to something with MORE capability, but I'll have some good stuff "in the can" in the meantime. I still have absolutely no idea how one "delivers" 4K, but that will come... at least I won't be jumping all the hurdles at one time... and if pan and scan delivers good 1080, I'll be thrilled!

[Wacharapong Chiowanich]

Ken,

I'm perplexed myself. I didn't look for it but it just manifested itself when there's movement under those conditions. Maybe I have to wait some time to have a proper 4K monitor to see what it is about.

[Dave Blackhurst]

I don't think it's that tricky - the high contrast naturally sticks out to the eye, and I do think frame rate has something to do with it being more noticeable, as the series of "still" images exhibit movement between frames, particularly very sharp crisp frames. I suspect the eye detects the sudden moves in the high contrast areas as flashes or pulses of light areas flickering against dark areas. I suspect that the lower the frame rate, the more striking the effect.

If the contrasting edges were less distinct, the frame rate higher so that the movements were smaller, or the image a tad blurry so that the differences between frames were not so prominent, the "problem" would be reduced or eliminated. In that sense the sharpness produced by the larger sensor being fully read and a good lens resolving high detail may be somewhat counterproductive in some high contrast/motion situations.

If that theory is correct, the "artifacts" are just the product of how our eyes "see" a series of high detail/high contrast "moving" images, not a "flaw" in the images.