|
|
|
|
|
July 4th, 2014, 04:56 PM
|
#1531 |
|
Inner Circle
Join Date: Feb 2007
Location: Apple Valley CA
Posts: 4,874
|
Re: Sony FDR-AX100
As Alister (who does have access to, and reviews MANY cameras, with a pretty good accuracy record as reviewers go) explained it, there are several different things going on here...
FIRST, skew when panning is a function of trying to read out roughly 4 TIMES the pixels from the SENSOR as if you were dealing with a 1080 sensor - the physics dictate this will take LONGER. It takes longer to count to 16 than it does to count to 4...even for a computer... Due to horsepower and physics, they are scanned (progressive) line by line, not all at once. This happens "fast", but (as I'll illustrate shortly), MOTION can happen FASTER... so what you see is that by the time the lower portion of the sensor is READ, if there's been fast motion/panning, the position of a vertical object is not the same as it was when the upper portion was READ, so it now appears to lean (or curve with variable pan rate). THUS, why you want to avoid whip pans, PERIOD, and why slower pan technique is required at the current level of technology. I see equivalent levels of RS (at 1080) if I whip my PJ7xx series cameras back and forth - it's easier to see using the VF, but it's ALWAYS been a factor with CMOS sensors! The idea would be a "global" read, but that's not yet an available "option"! Nothing you, or I or Panasonic (or Sony) will do will change the laws of physics, don't expect the FZ1000 to be magically "better"... In fact, some of the macroblocking and artifacting in shadow areas I've looked at in the sample videos indicates that the FZ1000 is tossing out far more data than the AX100... I guess that's ONE way to deal with data overload! The PROCESSOR is another factor, it takes the data read, and munches and crunches and rearranges it into the final file format, and that takes TIME. Todays processing is faster than "yesterdays", that's how it works in electronics... a 4770K i7 is faster and more capable than the older 920... Even with "better" compression, you're still dealing with roughly 4x the data at points in the process... As I've already explained, the time required to read the ENTIRE sensor for 4K is by definition LONGER than reading a sensor for 1080p. Even longer if you consider reading every other line for "i" (anyone remember "mouse teeth", AKA temporal motion artifacts). We are talking "miliseconds" (tiny fractions of a second), but the possible motion during that time can be SIGNIFICANT. We don't perceive it with our eyes, because our brain processes it in the way we "think" is normal. Reproducing vision with a mechanical or electronic means (AKA a camera) can result in "odd" results... Sony no doubt is working on ways to speed up the capture/processing, just like Intel is always working on faster chips... there's no evil conspiracy involved, just engineers continuing to push the envelope, and getting farther and farther over time. That's why "the perfect camera" is and will ALWAYS be "under development", because with each "newer, better", we will find SOMETHING which could be still "better". NOW, back to how to practically deal with the cameras we have available... OR, "how the physics define use parameters". If you want to avoid skew, PAN SLOWLY - the more motion, the more the required time for sensor readout will be apparent. ALSO, the more pixels that change from frame to frame, the more load on the processor, and the more "artifacting" and blurring you'll get, which becomes more apparent with 4K. I've mentioned high shutter speeds freezing the action - this results in the eye seeing "stutter" from 30 sharp frames, which our brain begins to process as a "flip book" of stills. This is worsened by the greater sharpness of 4k over 1080 over SD... AS mentioned, shutter speed then comes into play - how far can a given object move in 1/30 of a second... 1/60th... 1/125th... YES, it DOES matter, and fairly significantly!! Think about it, and in the first, the moving object will move FOUR TIMES more distance than the last... yes, SIGNIFICANT!!! The corollary is "how far can the camera move" in each of those time frames... do you see where understanding the physics of movement in fractions of a second now becomes PARAMOUNT in camera operation?? SO, the goal of the camera operator is to find a balance between shutter speed, camera motion, and "talent" motion blur that "looks" acceptable. With 4K, you must deal with the added number of pixels and resulting sharpness that makes "older" cameras look fuzzy and soft. As Ken and I have both posted, taking shutter speed down gives a degree of motion blur as "talent" or background go through the frame (or the frame pans across the scene) that reduces the jarring effect. BUT you lose some of that "super sharpness" (which is perhaps not a terrible thing for fast action scenes!). You MUST pan like a motion picture camera would (slow and deliberate). I suggest framing wider than "normal", and you can crop/pan in post and STILL have better than typical 1080 resolution. Steadiness matters FAR more, as any movement becomes more noticeable with 4K. OK, perhaps those "limitations" are better suited to a cooking show than a soccer match or auto race, but you can STILL adjust the camera to get decent results for the latter, OR switch to the 1080/60p XAVCS mode for better quality HD. As you begin to take into consideration how much movement you'll be dealing with, the advantages of 2x the frame "samples" becomes apparent/desirable... I felt the 30p would be a "major" issue, and in some ways it contributes a few extra hurdles, but it CAN be dealt with and still get EXCELLENT resulting footage. If you can keep the majority of the frame stable, and get a nice touch of motion blur to things moving within the frame, it looks insanely great. Wiggle the camera about and let the shutter speed run amok... not quite so much... Bottom line, 4K is "new", it's "developing", it's going to have teething pains.... There's a reason that the AX100, the Seiki 4K TV/monitor, and the Intel integrated graphics chipset are ALL 30p "solutions"... #1, I'm cheap, and these were all "affordable", and #2, they will all likely be mostly OBSOLETE in 3-5 years, they will look "dated" as 4K/60p becomes feasible and the norm. BUT, if it opens the door to 4K to mere mortals on budgets... it ain't so bad! And I can't argue with what I can shoot and process RIGHT NOW. THIS is the state of technology, learn to embrace, use, and enjoy it... or stick with "tried and true" HD "solutions"... or be prepared to spend A LOT more for "professional" solutions. I'll take the "bang for the buck" of my current system over what it replaces any day of the week, even with a few limitations. The "old stuff" had limitations TOO... Once again, hopefully you will take the practical experience of others (some of us who DO own way too many cameras and can speak to how they perform...). Railing about technology that simply is not practical or economical or perhaps not even available because of some evil conspiracy is a waste of time. I can almost guarantee that if you're disappointed with the AX100, the FZ1000 might well make you completely apoplectic, even though the price is a lot less... it's all there in the sample videos... and plenty of "gotchas" in the 4k implementation "on paper". You already see the potential in 4K, just accept that the technology has limits, and you'll be far happier enjoying what IS! |
|
|
|
July 4th, 2014, 05:14 PM
|
#1532 |
|
Major Player
Join Date: Apr 2012
Location: New York City
Posts: 329
|
Re: Sony FDR-AX100
The problem is that the shutter rolling or rolling shutter (depending on how the internet enthusiasts like to call it) is more visible on certain cameras and less in others. The GH4 for example should be better because of a faster readout (not because must be bad by default and there in nuthing they can do about it). We know that the AX100 is a consumer camera for consumers and I must admit that Sony is pretty clear about it. no special cards needed, limited mbps and so on. So be it. It doesn't get the best technology. What I'm here to state is that (like I said several times already) it doesn't have much to do with the framerate. So let's take the myth out of this plate.
Ordered the bracket by the way, but I'm not going to get the XLR thing. Took a look at the AX1 and looks great if it wasn't for the cards (and no face detection that means a lot to me) : I need to produce 2hrs up to 4 times a week so I'd need 5 64GB cards and changing them like a DJ , plus 5 cards will cost me a thousand (just the cards). No can't do. In any case B&H have several returns at a good price, I was sertiously thinking about it. But then the cards and other stuff. Sure I'll probably able to use the varizoom (that has a dream touch for zooming) but the cards... not now at least. The AX100 must do the work, and it will. My tests on the field yesterday (with the stabilization OFF) went very well. I'm ok with the AX100. And thanks for taking the time to write that book over there :) |
|
|
|
|
July 4th, 2014, 05:20 PM
|
#1533 |
|
Inner Circle
Join Date: Feb 2007
Location: Apple Valley CA
Posts: 4,874
|
Re: Sony FDR-AX100
I suppose you "could" say it's the inability of the processor to "digest" the output from a progressively scanned sensor... but it's more accurately in how a sensor is read... only the engineers know exactly where the bottleneck is... is it the chicken or the egg, metaphorically speaking... can the sensor be read faster? Can the processor ingest that data faster?
We don't KNOW, we can only deal with the reality. Still photogs shoot FAST shutter for a reason. Video dudes have different considerations. Frame rate is NOT an "internet myth" when dealing with MOTION, the sample rate of moving objects is REALITY, period. As is the reality of how much data a camera can practically deal with (4x as much with 4K). The math isn't exactly 4x as has been pointed out, but from a practical standpoint, it's easier to "keep it simple" to make it easier to grasp where the challenges lie. 4K is more data points than HD, 60p is more data points than 30p. Each increase requires at least SOME increase in sensor read speed, processor speed, memory speed and capacity... you can't get something for nothing! More data points over time give you a more "accurate" representation or "capture" of reality. At a price. Calling it a myth simply denies the physics and reality... The AX100, at $2k broke ground, the FZ1000 at $900 breaks more... but these cameras WILL have limitations with the 4K implementations. At least I was able to stick an SDHC card in my AX and have it work perfectly fine even though the specs are adamant about SDXC! I was expecting a good degree of fun trying to find "fast enough" memory! |
|
|
|
|
July 4th, 2014, 05:30 PM
|
#1534 |
|
Inner Circle
Join Date: Feb 2007
Location: Apple Valley CA
Posts: 4,874
|
Re: Sony FDR-AX100
Well, some have reported some possible RS related issues with the GH4, which I'd expect... again, got to learn to work with the camera.
Panasonic doesn't get to violate the laws of physics... all the things in my "long post" apply to one degree or another to them, same as Sony... You have discovered the "price" to be paid for higher framerates, higher bitrates, and "better" 4K, and yep, it ain't pretty, not one bit! Those fast memory cards that fly away like candy at the circus are needed to deal with the same larger numbers of bits that originate at the sensor, if not reduced via compression along the way... I think the FZ1000 memory specs are also higher (and likely more expensive, offsetting the initial "savings")? No free lunch. |
|
|
|
|
July 4th, 2014, 05:35 PM
|
#1535 |
|
Major Player
Join Date: Apr 2010
Location: Atlantic Coast Canada
Posts: 599
|
Re: Sony FDR-AX100
For the incoming Lumix fz1000, I Just bought 2 x Transcend SDXC 64GB U3 95MB/sec Read 60MB/sec Write Ultimate UHS-1 Memory Card cost 79.00 for the pair shipped
Last edited by Bruce Dempsey; July 4th, 2014 at 06:13 PM. |
|
|
|
|
July 4th, 2014, 06:03 PM
|
#1536 | |
|
Major Player
Join Date: Apr 2012
Location: New York City
Posts: 329
|
Re: Sony FDR-AX100
Quote:
I use the cowboy studio 20bucks shoulder mount and it's fine. It was fine with the EX1 so can definitely hold the AX100 ... it's a necessary thing, in my opinion. The part that pushes on the chest mu be replaced with something softer and thicker or you'll get the camera moving on each and every breath you take. With more layers of something soft it will be fine. Do you know of any eyecup for the AX100? |
|
|
|
|
|
July 4th, 2014, 06:06 PM
|
#1537 | |
|
Major Player
Join Date: Apr 2012
Location: New York City
Posts: 329
|
Re: Sony FDR-AX100
Quote:
if you are referring to the cards for the AX1 then they are not the same: the AX1 needs XQD , that cost 200 bucks each and can take 25 minutes of 4K on 64 GB |
|
|
|
|
|
July 4th, 2014, 06:09 PM
|
#1538 |
|
Major Player
Join Date: Apr 2010
Location: Atlantic Coast Canada
Posts: 599
|
Re: Sony FDR-AX100
Hey Anthony
no no those ones I mentioned are for the fz1000 which is now 3 wks away. Got the 4k seiki tv today all set up ready for my cam regards Bruce |
|
|
|
|
July 4th, 2014, 06:20 PM
|
#1539 | |||
|
Major Player
Join Date: Apr 2012
Location: New York City
Posts: 329
|
Re: Sony FDR-AX100
Quote:
Quote:
Quote:
|
|||
|
|
|
|
July 4th, 2014, 06:24 PM
|
#1540 |
|
Major Player
Join Date: Apr 2010
Location: Atlantic Coast Canada
Posts: 599
|
Re: Sony FDR-AX100
What's the big deal about 30 min? Dont you stop tape at all? Play interrupted for whatever reason, you just keep rolling?
|
|
|
|
|
July 4th, 2014, 06:25 PM
|
#1541 | |
|
Major Player
Join Date: Apr 2012
Location: New York City
Posts: 329
|
Re: Sony FDR-AX100
Quote:
about the 30 minutes thing.. Bruce, it will be a disaster. When you shoot a game you have literally no time to think about anything else than the game. You can't lose the ball because if you do then it will take 3-4 seconds to get it back and everybody can see it. It's a tough business, believe me, no time for anything (even to go to the bathroom LOL, no, seriously the day of the game I don't drink at all, even water, just because of that) . 50 minutes straight per period. I can't possibly deal with that kind of stress of having the camera stop. Next season we'll go live, imagine that ... no, the FZ1000 can't do it (for me I mean) |
|
|
|
|
|
July 4th, 2014, 06:29 PM
|
#1542 |
|
Major Player
Join Date: Apr 2010
Location: Atlantic Coast Canada
Posts: 599
|
Re: Sony FDR-AX100
really fine once I figure out how to get my nvidia 440 card set up manually for 4k.
All I've got to play back so far are some of the promotional 4k videos but even those are much more relaxing to watch Certainly easier on the eyeballs to watch How's the storm going must be over your head bout now It will be here up the coast tomorrow mornign and I'm right on the water so its always a bit of a worrry |
|
|
|
|
July 4th, 2014, 06:56 PM
|
#1543 |
|
Inner Circle
Join Date: Mar 2003
Location: Ottawa, Ontario, Canada
Posts: 4,222
|
Re: Sony FDR-AX100
Sony's professional solution proposed for 4K is two fixed cameras with electronic panning and zooming. The advantage of 4K is this pan and scan capability if one is going to follow the action then a straight HD camera would be better. Anthony I shoot theatre and stand up for each act usually about 60 mins or so at a time the last dance show ran 2 hours and 15 mins straight. No bathroom breaks !!!
XQD cards for the AX1 are $200 for 64G and run about 55mins with auto switch over to second card slot then the first can be changed etc . The PMW-Z100 of course uses these cards up in 12mins for XAVC 422 10bit recording. Ron Evans |
|
|
|
|
July 4th, 2014, 07:09 PM
|
#1544 | |
|
Major Player
Join Date: Apr 2012
Location: New York City
Posts: 329
|
Re: Sony FDR-AX100
Quote:
LOL when I was joking (and sort of complaining but still joking) about the bathroom thing with a still photographer he said "wait few more years and THEN you'll really understand what's like ..." Yeah, the AX1 hold about 26 minutes per card, that with the auto switch to the next would give me 52minutes . Still it would be risky because it depends on the referee , I mean how many extra minutes he'll set. Changing the first card and put another one would do, but still it's something that I don't want to deal with now. I'd rather wait for larger cards, it shouldn't take long, Sony and Lexar are in the market already. But since I'm stubborn in nature I wanna make the AX100 work (LOL , and I will, in 4K). Well, if possible, if not then .. oh well. But I think that I can make it work (frm what I saw yesterday on a field test at least) |
|
|
|
|
|
July 4th, 2014, 07:14 PM
|
#1545 |
|
Major Player
Join Date: Apr 2010
Location: Atlantic Coast Canada
Posts: 599
|
Re: Sony FDR-AX100
Anthony I could help you with the Live streaming .Been doing it now for 2 1/2 yrs
|
|
|
|
 |
| ||||||
|
||||||
|
|
