View Full Version : FS5 - Macro blocking codec issue in 4k??
Cliff Totten December 30th, 2015, 07:42 PM Can you upload some short samples to a Google Drive or Dropbox?
So your camera DOES show these blocky eges in other standard Rec709 settings?
SLOG and Rec709 800% are great for grading. However, there are times when you want to give out a "baked in" finished look.
So it seems the FS5 is stuck in UHD SLOG and Rec709 800% ??
We gotta get more samples and proof to Sony ASAP.
Thanks!
Josh Swan December 30th, 2015, 08:02 PM Yeah it shows the nasty edges at 3200 and above on other gammas. Below 3200 and it completely goes away. I think you could get a pretty good look with 709 800% and adjusting the profile if you wanted the baked in look. I actually prefer 800% over the standard 709 gamma
Cliff Totten December 30th, 2015, 08:12 PM Wow, uggg,...well? I guess maybe we can add your FS5 to the list of problem cameras.
Have you contacted Sony about it yet?
I am really hoping that this issue gets reported and makes its way up the Sony tree and up to Japan. If it doesn't exist in SLOG or Rec709 800% than maybe there is hope that they can correct it in standard Rec709 gamma?
FS5 owners need to also be very aware of this problem too. It sucks to see everything fine on your small camera monitor only to get shocked when you see it in your NLE at home.
Not good!
Josh Swan December 30th, 2015, 08:44 PM As long as you understand the camera and its limitations it's an awesome camera.
Cliff Totten December 30th, 2015, 10:07 PM True....
But having a limitation of not being able to record standard Rec709 without horizontal edge artifacts, something that even my cell phone can do properly,...is a very embarrassing problem for a $5,500+ pro camera.
On the FS5, does Rec709 800% force a minimum ISO? My AS7-II forces it at 1600 ISO. (Same as SLOG)
Ouch!....
Jack Zhang December 30th, 2015, 11:58 PM Edit: The A7S-II got a firmware update this month. One of the Sony stated fixes was "image quality improvements". I did a few short (original firmware) recordings with the lens cap on to record all black and different ISO settings. I then upgraded to the new firmware and did the same recordings with the lens cap on. I DID notice some slight noise changes between the firmware versions. When I heavily stretched the contrast in both blacks upward, I saw that the original firmware had very faint "horizontal" noise "lines" while the pattern with the new one was more "dot and spot" looking.
Could they have modified the noise reduction process between the two firmware?
It's possible they tuned some of the random noise reduction to a lower level. Yes, you can turn down the amount/intensity of the different noise reduction types with firmware, but the method/algorithm is baked into the chip. I suspect Sony turned down some noise reduction features that were too strong to begin with on the A7s II. The intensity of the noise reduction is way lower in S-Log mode even on the FS5. It's not off, cause some things like fixed pattern noise reduction need it to remain on. You still have to keep in mind the different imagers and I suspect they cranked up some of the noise reduction at standard gammas to hide problems with the mismatching algorithms, hiding stuff better at the cost of introducing other artifacts.
Cliff Totten December 31st, 2015, 12:49 AM Wow Jack.....if what you are saying about the FS5 is true, Sony has one Hell of a defect on it's hands here with a camera that they really want's to sell allot of in 2016.
If Sony cant repair this problem what are they going to do? Can they try to sweep this under the rug and pretend it's not happening? More and more problem samples will come out and complaints are sure to grow and grow.
I'm definitely looking for more FS5 owners out there to shoot more straight line tests with motion.
Better yet, if somebody can shoot a resolution chart and "tap" or "wiggle" the camera and post it in UHD, that would be awesome.
This is depressing......
Wacharapong Chiowanich December 31st, 2015, 01:22 AM Sorry I am new to the FS5's REC.709 800% terminology here. What does it mean that makes it different from the usual REC.709? I'm planning on getting an FS5 myself. Thanks.
Noa Put December 31st, 2015, 03:43 AM If Sony cant repair this problem what are they going to do?
Newsshooter has reported the problem to Sony and they claim Sony is looking into it and I"m sure newsshooter is not the only source that has reported the problem, so until Sony decides to fix it with a firmware upgrade it is what it is, you either buy the camera and work around it's limitations and wait until it maybe will get fixed, or it might not get fixed and if that would bother you then get a fs7.
Josh Swan December 31st, 2015, 07:14 AM Wouldnt proving that the camera can record beautiful images on three gammas prove its not a hardware issue. Gammas are a software change, so it's clearly not the chip.
Josh Swan December 31st, 2015, 07:19 AM Sorry I am new to the FS5's REC.709 800% terminology here. What does it mean that makes it different from the usual REC.709? I'm planning on getting an FS5 myself. Thanks.
They are identical until it starts dealing with highlights. https://www.google.com/search?q=rec709+800%25+gamma+curve&client=safari&hl=en-us&prmd=isvn&source=lnms&tbm=isch&sa=X&ved=0ahUKEwihqcWzm4bKAhXC5yYKHRl-BOYQ_AUIBygB#imgrc=M9EFxMHd8XQHvM%3A
Cliff Totten December 31st, 2015, 10:31 AM Yeah it shows the nasty edges at 3200 and above on other gammas. Below 3200 and it completely goes away. I think you could get a pretty good look with 709 800% and adjusting the profile if you wanted the baked in look. I actually prefer 800% over the standard 709 gamma
Rec709 800% is great....IF....you dont mind being stuck at 3200 as your lowest ISO. It sucks to be forced to run in a WAY higher ISO than you need only to have to dial all that light back with ND.
It does make you wonder though, if Sony can give you a clear image at rec709 800%, why cant they do that at standard Rec709 at ISO 100-3200?
How is Sony's implementation of this possibly faulty noise reduction system different between these two modes?
Calling all FS5 owners! More test samples please, please, please. ;-)
Jack Zhang December 31st, 2015, 12:06 PM Wouldnt proving that the camera can record beautiful images on three gammas prove its not a hardware issue. Gammas are a software change, so it's clearly not the chip.
Yes, changing gammas can also change in the intensity of the noise reduction algorithm who's FUNCTIONAL ALGORITHM, not INTENSITY can't be changed for a new sensor. Gammas are easily tweakable, so is what you tell the chip how MUCH something should be noise reduced, but the function and algorithm is baked into the chip once it's an ASIC.
Imagine this, you try to make a Bitcoin mining ASIC change into a general purpose CPU... You just can't. The function is VERY SPECIFIC and baked inside the chip, and to change instruction sets so that it turns into a general CPU is neigh impossible. (and if it is possible, incredibly inefficent, nullifying the point of changing the specific purpose of a ASIC after it's been made.)
Cliff Totten December 31st, 2015, 01:03 PM So, if you are right about these ASIC chips providing "broken" or "improper" noise reduction algorithms to the FS5's image processing, the only way for Sony to fix this is to dial back on its usage amount.
Dong so, might reduce the problem the NR algorithm is causing but will of course, allow more noise back into the image.
I'll take it! Noise I can handle! I can always clean that in post if needed. What we cant do is repair the artifacts this is causing on the image today. Once resolution and detail is destroyed at the image processing level, it cant be recovered in post afterwards.
Maybe this is Sony's only hope for a FS5 firmware fix?
Question for Jack: Could Sony have baked in MULTIPLE noise reduction patterns/algorithms into this ASIC? (Designed for multiple Sony sensor types)
Is it possible the the FS5's firmware is implementing or activating the wrong one? We know the FS5's S35 sensor is not a new one, it's used on the FS5, F5/55 and FS700 too. So we know that Sony has the right noise reduction algorithm for that sensor in it's technology somewhere.
Could this noise reduction ASIC have those S35 sensor noise models baked in too and not being accessed properly? (The FS5 firmware is calling on the wrong NR model in that ASIC?)
CT
Olof Ekbergh December 31st, 2015, 01:53 PM OK I can now reproduce the problem reliable, after extensive pixel peeping and I found the best way to spot the problem. I use my Odyssey O7Q+ in UHD and go pixel to pixel or even 2 to 1. This way I can see the effect w/o going back and forth downloading to my edit suite.
It took a long time for me to even spot this problem as I was working on developing PPs and test shooting outdoors and studio setups. I never saw it in any of my footage. I really had to work at it but now I can see it clearly and reproduce it. I will contact Sony and send them the video below that shows what is happening very clearly.
So my conclusions are:
It is a very special circumstance that causes the problem. A high contrast level edge slightly out of focus moving up and down slowly, fast is fine. In focus seems better but sometimes shows up.
It is interesting that it is not every frame, just some frames, others are fine, and rarely are there 2 bad ones in a row.
It happens UHD in all the CINE gammas 1-4 at 3200 iso and up. It can be seen but rarely at 1600 iso.
All other Gammas are fine in my tests.
It does not happen in HD.
Noise Reduction settings make no difference at all, I tried all the variations.
All my tests were in 29.97.
So here is a short video it is ΒΌ of an UHD, so it is pixel to pixel viewed in HD. I can see this in reference monitors and on my Mac screens, so it is not the same noise you sometimes see in a computer screen when the GPU is overwhelmed.
My bottom line is the 5 is a great camera, but it has a definite problem with CineGammas in UHD at iso 3200 and up, but this will only show up in fairly rare circumstances, at least the way I shoot. Hopefully Sony will look into this and correct it. It can be fixed in post if you get it in a shot and it needs to be used, just do a tracking blur over the affected area with soft edges. It is not stopping me from using Cinegammas at 3200, I will just be aware of what may happen in some rare circumstances.
Anyway Happy New Year to everyone. These are interesting times...
FS5 Ripped Edges in some shots on Vimeo
Andy Wilkinson December 31st, 2015, 02:00 PM Just to say thank you Olof for your efforts in establishing these details. Happy New Year!
It will be interesting to see how Sony respond.
Peter Newsom December 31st, 2015, 02:56 PM I think that perspective has been lost here, and some of it starting to sound a lot like Sony bashing.
Sony has developed a very nice little HD camera in the fs5, which includes UHD capabilities. It is pleasing to the hand and to the eye and is offered it at an attractive price. It's a bargain...if it meets your requirements. If not move on.
There have been many beautiful examples of the quality images that fs5 can produce posted by a number of professionals on this site and elsewhere.
The 'macro blocking' in the dark could be resolved by throwing a bit of light on the scene and holding the camera steady.
Any camera can be made to look bad if you try hard enough.
Cliff Totten December 31st, 2015, 03:40 PM I see no Sony bashing here. In fact, I see people trying to identify a problem so that Sony can fix it and sell more cameras.
Nobody wins if we all try to ignore this problem, pretend it's not there and try to sweep it under the carpet.
This artifact happens in good light and bad light. It ruins edges in UHD and ultimately the hurts the overall resolution potential of the camera.
No other Sony UHD camera has this problem that I know of and it's a shame that a $1300 RX10-II can out resolve a $5600 XDCAM FS5 because of this edging problem.
Let's not "bash" Sony on this. Let's help Sony identify the issue so they can fix it and make the FS5 all that it can really be.
We had a fit about the A7S-II over the "black hole sun" problem. We didn't just sit back and pretend it didn't exist. Sony took notice and fixed in in 3 weeks for their customers.
I have owned 18 Sony cameras in 20 years. I'm certainly in the Sony "fanboy" catagory. I'm about to buy the FS5 in 3 weeks...so no, I'm not a Sony basher...I'm quite the opposite, for sure.
Jack Zhang December 31st, 2015, 03:47 PM Question for Jack: Could Sony have baked in MULTIPLE noise reduction patterns/algorithms into this ASIC? (Designed for multiple Sony sensor types)
DSLR ones, maybe. But not the Super35 sensor.
I didn't know the A7s II had a black sun issue. That must meant they REALLY rushed the firmware and the camera out to market. Usually black sun issues only apply to industrial CMOS sensors or unprocessed raw sensor data. They must have rushed these cameras to market, like the PXW-X200 with it's infamous lens focus issue.
Ray Lee December 31st, 2015, 04:49 PM Rec709 800% is great....IF....you dont mind being stuck at 3200 as your lowest ISO. It sucks to be forced to run in a WAY higher ISO than you need only to have to dial all that light back with ND.
It does make you wonder though, if Sony can give you a clear image at rec709 800%, why cant they do that at standard Rec709 at ISO 100-3200?
How is Sony's implementation of this possibly faulty noise reduction system different between these two modes?
Calling all FS5 owners! More test samples please, please, please. ;-)
I have been unable to get the issue to show in standard rec 709... and seems like people who have the camera available for testing have not either. (its getting hard to sift through the testing, what people read and re post, and technical "theory" I worry Sony is going to get tons of "samples" and emails that really are not even related to the real issue.
Not sure how its "sucks" to have use ND in this camera and we will have to work at 3200 to use SLOG3 as well, it is not like the ND is causing any color shift or any IQ loss... I can't se any difference with or without ND engaged... sharpness and contrast all look the same to me, but I admit I didn't zoom to an absurd level just 100%-200% UHD on a 1080 time line and I have been watching some on a 60 inch UHD tv.
I have been able to recreate the dancing line issue in CINE gamma at 3200/6400iso at 1 to 1 , hope it can be fixed, I wouldn't want to see busy scenes like tree branches in the background of an image go crazy with artifacts
but as far as the original large blocking issue (the issue News Shooter re-posted) I have not been able to get anywhere near that... with a couple of exceptions grossly underexposing and bringing up in post and/or a combination of under exposure and digital zoom at high ISO values.
Cliff Totten December 31st, 2015, 06:31 PM When I sat it "sucks"....I simply mean that I have never...oh let's say: Crank up to 12db of gain only to throw on the ND3 filter on an EX1.
Gain and ND are counter intuitive when used like that.
It's not absurd to want to shoot in a standard Rec.709 gamma and run at 200 ISO or low gain like that.
You don't think it's odd that the FS5 can't do that without getting hit with edge blocks?
I dunno....maybe I'm expecting way way to much from a $5500 camera?
CT
Ray Lee December 31st, 2015, 07:35 PM "When I sat it "sucks"....I simply mean that I have never...oh let's say: Crank up to 12db of gain only to throw on the ND3 filter on an EX1."
I never owned an EX1, but with the FS5 3200 in log or REC709 800% 3200 isn't 12db of gain its the native gain
"Gain and ND are counter intuitive when used like that."
Its not really gain for that gamma just like LOG the native is 3200 so your only options are close iris, add ND or speed up the shutter....in this case ND wont even cause color shifts or any loss of image quality so thats the way I would go
"It's not absurd to want to shoot in a standard Rec.709 gamma and run at 200 ISO or low gain like that."
Nope not at all, but this camera simply wont do it in 709 800% or SLOG (wont do 200 in any setting, 800 is the lowest in Cine1)
"You don't think it's odd that the FS5 can't do that without getting hit with edge blocks?"
Not sure what you mean by "that" I think the line/blinky/artifacts are really F'd up to be 100% honest.... luckily for me even as it sits the pros outweigh the cons.. I desperately needed a smaller camera, the vari ND is going to be amazing for what I shoot (big lenses that are very hard and expensive to filter) and the better HD/SLOG option will be absolutely wonderful.
"I dunno....maybe I'm expecting way way to much from a $5500 camera?"
I have no idea what you expect or how this applies to what I wrote. I personally expect it will get improved but do worry Sony is going to get blasted with strange requests and "issues" that might obscure the real ones.
Seems like maybe I was unclear in my post but I just don't have much time to go on about it so I will just repeat the main point of my post. I am able to replicate the blinking line artifact issues in the CINE modes and at 3200 and up..... but after trying extensively I am unable to replicate the big blocked background look of the footage News Shooter reposted without using digital crop or underexposing
I hope you don't feel I was being rude or anything, just posting what I found with my fs5 and if you disagree with it... or if yours is doing something different than mine so be it... post away I am out :)
Cliff Totten December 31st, 2015, 07:46 PM No, you are cool and I dig what you are saying. I don't own an FS5 so I can't say anything conclusive at all about it.
It's funny, on the A7s, the "native" ISO was 3200 with SLOG. On the A7S-II, with the same sensor, that magically changed to 1600 ISO.
Is it correct to say the the FS5's native 3200 ISO is equal to 0db of gain? Is any ISO below that running negative gain?
My brain operates better in "gain" than it does in "ISO"...ha!
Can't wait to test my FS5 for myself soon anyway.
CT
Wacharapong Chiowanich January 1st, 2016, 02:46 AM They are identical until it starts dealing with highlights. https://www.google.com/search?q=rec709+800%25+gamma+curve&client=safari&hl=en-us&prmd=isvn&source=lnms&tbm=isch&sa=X&ved=0ahUKEwihqcWzm4bKAhXC5yYKHRl-BOYQ_AUIBygB#imgrc=M9EFxMHd8XQHvM%3A
Thanks Josh for making it very clear on this and thanks to Olof too for taking the time and effort to pinpoint exactly where the problem lies. Still a great 4k camera for the price despite all these downsides IMHO. Hopefully I'll keep those in mind and make sure to get around them. Happy New Year to all.
Andrew Reid January 1st, 2016, 08:30 PM Have been shooting with my FS5 for only a few days but it's clear the 4K image pipeline on this camera needs a firmware fix.
For now I recommend avoiding the ripped edges by shooting S-LOG 2 and no higher than ISO 6400. Also ITU 709 800% at the base ISO of that profile seems fine too.
What I am most disappointed about is the 10bit 1080p performance. 50Mbit/s is way to low for the smooth gradation 10bit is capable of, it just gets compressed to hell! Macro blocking and banding all over the place and nothing like 10bit on a Blackmagic.
Keith Moreau January 2nd, 2016, 09:17 AM @Olof: Thanks so much for narrowing down the issues. It's very helpful to know when these artifacts happen and how to avoid them. I did a shoot in Cine the other day and was wondering why the edges of interviewees seemed a bit weird, and this explains it. Cine gammas are my go-to settings at this point as they offer higher dynamic range but avoid the log workflow. But for now I may choose log or rec709 800. Hopefully Sony will address these issues soon, I was pleasantly surprised when they issued firmware updates so quickly for the A7sII and FS5, which means they either responded very quickly or they knew of the issues and had the fixes in the pipeline and were just QA'ing them further.
I'll be discussing this in an upcoming issue of my cinematography podcast, TechMove.
-Keith
Listen To TechMove Podcast (http://www.techmovepodcast.com)
Cliff Totten January 2nd, 2016, 10:06 AM I'd love for Sony to issue a statement that acknowledges the problem and states that they are working on a solution.
They did this recently for the A7s-II "black hole sun" problem and fix.
At this point, we don't even know if Sony is even really aware of this today. However, If they do know about it, they might not want to say anything about because they are afraid it could hurt FS5 sales. Who knows? I certainly believe that being upfront with product problems improves your brand reputation and is a very noble way to handle these things. In the end, I think this strategy improves customer relations and increases consumer confidence too. How many people where impressed when Sony talked about that A&S-II firmware problem and fixed that thing in 3 weeks? I know I was very impressed.
If this thread topic hasn't arrived on Sony firmware programmers desks yet, at least it's getting the warning flag up for the FS5 user community.
There is nothing worse than shooting in a mode that you think is OK in the field only to get home and find out that you recorded a nasty problem with edges in your UHD Rec709 shots.
CT
Cliff Totten January 4th, 2016, 11:05 AM More information is starting to come out:
Sony FS5 user reports artifact issues triggered by changing exposure, ND. Visible on HDMI output (http://www.newsshooter.com/2016/01/04/sony-fs5-user-reports-artifact-issues-visible-on-hdmi-output-and-also-triggered-by-changing-exposure-nd/)
And more samples and here along with other issues discussed:
Sony FS5 codec problems and 4K ripped edges bug - EOSHD (http://www.eoshd.com/2016/01/sony-fs5-codec-problems-and-4k-ripped-edges-bug/)
I REALLY hope that Sony in Tokyo is aware of this and is working on it.
CT
Ray Lee January 4th, 2016, 12:01 PM No, you are cool and I dig what you are saying. I don't own an FS5 so I can't say anything conclusive at all about it.
It's funny, on the A7s, the "native" ISO was 3200 with SLOG. On the A7S-II, with the same sensor, that magically changed to 1600 ISO.
Is it correct to say the the FS5's native 3200 ISO is equal to 0db of gain? Is any ISO below that running negative gain?
My brain operates better in "gain" than it does in "ISO"...ha!
Can't wait to test my FS5 for myself soon anyway.
CT
That EOSHD post is just what they (he) does, look no further than the NX1 review (better than the RED epic and a few other $10k-$20k cameras) I worry crap like this is going to hurt our chances of getting any improvements via firmware.... just to many people confusing normal 8bit compressed UHD issues with the actually problems that probably could be improved on.
NewsShooter should be ashamed, they could easily get a camera and test for themselves but they just repost someones video and write an article... they will both get thousands of hits and I guess thats what counts but starting to look like the "sky is falling" pack has taken over and I can't see any reason for Sony to fix anything now... its over, the damage is done, even if they could fix all the issues why would they? The camera is dead, those posts will get thousands and thousands of hits... and keep getting hits even after an firmware fix came out.
Bummed, I just wanted a better image in the LCD when I used focus mag and to be able assign AF on to the grip along with some other easy firmware tweaks to make the camera even more usable.. bet its not going to happen now
Cliff Totten January 4th, 2016, 01:04 PM "its over, the damage is done, even if they could fix all the issues why would they? The camera is dead,"
Not sure what you mean. The camera is not "dead", it's a brand new model and I'm guessing they CAN fix it in firmware.
Sony plans these XDCAM models to have at least a 3 year shelf life. So, I'm sure they have allot planned for this in 2016.
Let's hope they are aware of these problems in Tokyo
Cliff Totten January 4th, 2016, 09:33 PM More developments from two important industry groups:
Alister Chapman:
My response to all the FS5 artefact issue complainers. | XDCAM-USER.COM (http://www.xdcam-user.com/2016/01/my-response-to-all-the-fs5-artefact-issue-complainers/)
EOSHD:
My response to Alister Chapman's response to all the Sony FS5 complainers! - EOSHD (http://www.eoshd.com/2016/01/response-alister-chapmans-response-sony-fs5-complainers/)
I must say that I LOVE Alister's work and his teachings. I have been following him for years and I have learned allot from his work. However, I must disagree with him on this one issue.
I don't feel at all that this FS5 and it's "ragged edges" problem is "normal" or OK in any way. I'm not sure why he is trying so hard to defend this. Maybe he has not seen many problem samples yet?
Anyway, I suspect that Sony is now aware of this problem and I have faith that they will fix this problem soon.
CT
Noa Put January 5th, 2016, 02:24 AM It looks to me that the "heavyweights" in the industry are having a totally different opinion on the matter, I"m sure Sony will take Alistair's comment more seriously, he says small improvements might occur trough a firmware update in the future but to not expect miracles and especially no fs7 performance since its a totally different camera. How fast this will be implemented by Sony will probably depend if all this negative response will have a impact on sales.
Also Alistar owns the fs5 and fs7 so he knows what he is talking about. I also wouldn't call the eoshd blog a "important industry group", it's just a blog owned by and for video enthusiasts, the reason why he shouts out now is that his first fs5 post was to tell everyone why he thinks this camera was the best money could buy but he made the mistake basing his conclusions on specs only (like he always does) and when he finally got the camera he saw it looked different then expected so now he wants Sony to fix it. I think Alistar accurately described it on his own blog that the article on eoshd was full of incorrect or sensationalist statements that really do show how little the blogowner understands video cameras, this is going to be an interesting discussion to follow :)
I personally feel people are overeacting, from what I have seen so far the fs5 can produce some lovely images and if you still think it's not good enough, invest 2k more on a fs7 and be done with it. The fs7 is actually a bargain if you compare it to a c100 user looking to upgrade to a c300 for better image quality and/or features, the c300 is 3 times the price of a c100 so it's unlikely for most c100 users to upgrade as the extra investment is too high, I also read about fs5 users thinking about getting an external recorder to improve on IQ but don't understand why as you will loose all benefits of that small formfactor + the extra cost will bring it close or at the same level as a fs7 body and that would be a no-brainer for me if I had to choose.
Jack Zhang January 5th, 2016, 04:13 PM Everyone's complaining about it being a codec problem when it is not a codec problem. As Alister said, the FS5 uses half the power of the FS7, so the image processing is therefore crippled into fixed function ASICs that work well with DSLR sensors, but not the Super35 sensor. The FS7 uses more power cause it has more high logic FPGAs that are reprogrammable and can handle the image load.
I agree with the first point in the EOSHD article. An off switch is needed but on specific types of noise reduction. Fixed pattern noise reduction needs to remain on, because without that kind of noise reduction, your footage will be full of nasty lines. Random noise reduction needs to be off. That is what is causing the issues observed. And overload correction (black sun issue) needs to remain on.
Codec is moot point in this discussion since this is clearly an image processing issue.
David Heath January 5th, 2016, 06:24 PM I agree with the first point in the EOSHD article. An off switch is needed but on specific types of noise reduction.
Yes, assuming it is (as Alister thinks) due to inter-frame noise reduction, it's the obvious answer, and should be pretty easily implemented(?)
Perhaps the most unfortunate thing (if the above is true) is the way it was initially incorrectly reported as a codec issue. I believe I've read that a few people have tried an external recorder and it not making any difference? Which certainly seems to back up what Alister Chapman is saying about noise reduction.
Cliff Totten January 5th, 2016, 07:52 PM Any way to change the name of this thread and get the "codec" part of the title out?
We have come a long way with this issue in the last few days!
Christopher Young January 6th, 2016, 04:51 AM Perhaps the most unfortunate thing (if the above is true) is the way it was initially incorrectly reported as a codec issue.
Agreed. I haven't seen this type of edge artefact noise even on X70 XAVC-L 4K footage at 60Mbps. Which is intrinsically the same as the FS5 being that its XAVC-L 26/30p. The only difference being on the FS5 you do have the option of 100Mbps.
Interestingly on a 4K upgraded FS700 the 4K S-LOG2 ISO3200 appears to have less noise than the FS5. even though the S-LOG2 default is ISO2000 on the FS700. Added to which the FS700 doesn't exhibit these torn flecking horizontal edges at any ISO even in AVCHD. It has to be something in the way this four year old S35 sensor has been married to the encoding processor. Be it in the temporal noise reduction circuit or what.
Tonight via a colleague I heard of an FS5 being returned tomorrow because of this issue and the fact that the 18-105 kit lens will not hold focus when you re-frame. Being a parfocal lens it should. As I mentioned in another post elsewhere even changing to another kit lens doesn't solve the problem. You cannot quickly reframe using the camera zoom servo while rolling without having to re-focus. Oddly the problem is not there with the FF 28-135 or the original 18-200 servo lens that became a later option on the FS700.
All of the above was all physically demonstrated to Sony on one of their own demo FS5s yesterday so we are doing our bit down here in Oz to get some traction on these issues.
Hope Sony are listening.
Chris Young
CYV Productions
Sydney
Andy Wilkinson January 6th, 2016, 07:24 AM "All of the above was all physically demonstrated to Sony on one of their own demo FS5s yesterday so we are doing our bit down here in Oz to get some traction on these issues".
Thanks for doing this Christopher. There have been some very ugly discussions about this whole FS5 debacle flying around on some blogs and other web forums (not on here thankfully - which is why I often turn to this forum for information).
The "heat" on the interwebs will no doubt get Sony's attention - but the method you've chosen is by far the better way. I imagine FS5s being returned is also going to raise the profile with Sony too.
I'm patient enough to wait for Sony's official response to all this.
Keith Moreau January 6th, 2016, 01:00 PM I got my FS5 at the end of November, I think I would be called an 'early adopter'. After reading these posts, I'm going to try to avoid Cine and normal Rec709 (PP3) until further notice. I did a recording yesterday in Rec709, and then switched it today to Rec709 800% (PP4). I can't scientifically prove it, but in viewing in through a SmallHD 501 monitor (excellent monitor BTW for general checking of focus and exposure) I 'feel' that the image is a lot better and is also a bit brighter. I'm at ISO 4000 because I have to be. I'm doing this with a FS5 with a Metabones Speedbooster Ultra and a Canon 70-200 f4.0 L IS lens and I'm effectively shooting at f2.8. I'm also using the 'clear image zoom' to get even closer or reframing without having to zoom the Canon lens which isn't parfocal. I started with the 18-105 kit lens, and while this lens is 'ok', it's not as sharp and the image is not great, and manual focusing is pretty terrible and unusable because of the 'focus by wire' feature of the lens. I feel that all the 'focus by wire' lenses by Sony are at a disadvantage to manual or EF lenses, the only thing they do better is autofocus, because even with the 'cadillac' of electronic EF-Emount adapters, Metabones, the autofocusing feature seems to be non-functional with the FS5.
I'm using just one FS5 rather than what I did before which was to use a Sony EX1 for closeups, and a GH4 for a wider view. I'm using this gig to see if I can replace my EX1 + GH4 with just one camcorder and then crop in post for the closeups. The image quality being superlative isn't a requirement for this gig so I'm using it as a "test" for the FS5, profiles, and workflow. The 'bug' of the FS5's internal LCD/EVF being disabled is tolerable since the 1.10 firmware update. Having info on the FS5 and image on the external monitor works, though when the SmallHD suddenly shuts off due to batteries running out is inconvenient because you are basically blind! You can't switch of the FS5 setting that turns on the internal LCD in lieu of the 4K HDMI output while recording! That should be a button assignable function and should be switchable while recording but I doubt it's technically possible.
In doing this, I feel I can replace my trusty old EX1. I feel the image quality with this combo I have now is way better than the EX1, more light sensitive, and almost as versatile and certainly more versatile in post. This gig extends to tomorrow and I plan to try out some SLog profiles along with trying to get a LUT into my SmallHD 501 for previewing the final image.
All all, I'm kind of OK with the FS5. I'd like it to be perfect, I feel a little bit fooled by the advanced marketing, including the filmmakers / paid promoters reports on the FS5. I feel Alistar Chapman is an awesome and amazing resource for all of us, I've been following him since the EX1 introduction. I'm slightly perplexed by his stance on the FS5 as I feel that it should be at least as good as a A7rII or A7sII, Canon C100 or C300 in low light and I don't feel it is at this point. For example, for this gig, I have a very 'fast' lens setup with a 2.8 and I still need to be at 4000 iso because of the environment. Seminars are often like this. I feel the my new A7sII (and even the old A7s) is better, out of the box, image wise, than the FS5, which is a bit disappointing to me. But using the A7sII as a camcorder (I've tried) is a hassle compared to the FS5. Having dual SDXC slots with mirror or relay recording (I'm using relay recording for this current gig) unlimited recording time, electronic ND, Clear Image Zoom, large battery capability, and ergonomics of the FS5 is awesome! But we are early in the FS5. Within a month of the release, we've already had a significant HDMI output update that makes the FS5 much more usable in the real world.
I knew going in that the FS5 might not be a 'Cinema' camcorder like the FS7 or the C300MkII or the upcoming Ursa Mini 4.6K. I was hoping for a bit more initially, and we may get closer, and if Sony comes out with a Raw output update it may get there, but then it loses it's portability advantage.
I'm not sure if I've ever seen as many posts, blogs and reports about a new camcorder's problems as this new FS5, I feel Sony has to be hearing these reports and will take them seriously. But we shouldn't hold back! If you see problems let Sony (via this and other forums) know!
I'll be discussing all this and more on my podcast about Cinematography, Post Production and Workflow, TechMove, TechMove is the Podcast all about Digital Filmmaking, Gear, Mac Computers, and iOS Devices from a Unique Perspective! (http://techmovepodcast.com).
-Keith Moreau
Cliff Totten January 6th, 2016, 02:50 PM It wouldn't be fair to expect the FS5 to match the A7s-II or A7R-II. The A7x cameras have much larger sensors that completely out class the FS5's Super35 sensor.
The FS5's noise reduction issue is reported to happen at any gamma setting. For now, try to run at 0db only or as close as you can to 0db. Avoid high gain as best you can, even if it means slowing down to 1/30 shutter...in my opinion.
I'm sure Sony will try to do something in firmware soon. If they do nothing, the community will continue to harass and scold this model to its early death.
Some want to say the problem is perfectly fine and normal to see. It's just par for the course to live with that atrifact. However, I don't believe the rest of the market agrees with that concept.
Sony traditionally has done very well with +6dB or +9dB or +12db gain with other camera models in the past 4 years. The FS5 cannot be the only exception to Sony's great noise reduction technology.
I think Sony will fix it. I just hope they don't try to "wish it away" and they will put their engineering teams on it as a top priority.
Noa Put January 6th, 2016, 03:44 PM It wouldn't be fair to expect the FS5 to match the A7s-II or A7R-II. The A7x cameras have much larger sensors that completely out class the FS5's Super35 sensor.
Sorry but that makes no sense, a full frame sensor alone does not automatically mean a superior image quality, if I would follow your logic a canon 5d mark 2 would outclass a Canon c300 mark 2.
The FS5's noise reduction issue is reported to happen at any gamma setting.
Like it does on any camera out there that uses gain, like Alister said, "use gain and you get artifacts", saying to people now to not shoot higher then 0db gain is silly, ofcourse you can and will have to go higher then 0db gain for most uncontrolled shoots, like a wedding, but don't expect squeaky clean footage, my gh4 is noisy at 3200 iso and even noisier beyond that but I don't go complain this is not acceptable, it's just the best this camera can do in these circumstances and I and my clients can live with that. The fs5, even with it's artifacts still looks better at 3200 iso then what my gh4 can do. Most people did not even see the noise issue on the fs5 when the camera came out but that might be because they didn't wave their camera constantly up and down at 3200 iso to prove a point.
If they do nothing, the community will continue to harass and scold this model to its early death.
Sony won't loose much sleep over it, Sony does do small firmware improvements but Sony prefers to do improved mark 2 versions on their bestsellers so you can go out and buy a new camera again and if it's not a bestseller they just forget about it, just like they did with the nex-ea50, or vg900.
Sony traditionally has done very well with +6dB or +9dB or +12db gain with other camera models in the past 4 years.
Not all sony's are great at high gain, my ax100 is very noisy at high gains, same applied for my nex-ea50 and my rx10's image turns to mush at high gains as well.
I think Sony will fix it. I just hope they don't try to "wish it away" and they will put their engineering teams on it as a top priority.
I think in about every response you gave about the fs5 you hope Sony will fix this but look at it from the bright side, you at least have not wasted any money yet on the fs5 with it's dreadful image so you can still choose any camera you want :)
Just kidding but I think by now Sony knows what the problem is and maybe it's time people just start to use the camera, after all, it's not as bad as it looks, it might get fixed or it might not but in the meantime I"m selling yet another wedding to a happy client that's full of noisy 6400 iso gh4 footage, I just put black bars on top and bottom and say it's cinematic, that helps a lot too. :)
David Heath January 6th, 2016, 07:02 PM ......... my gh4 is noisy at 3200 iso and even noisier beyond that but I don't go complain this is not acceptable, it's just the best this camera can do in these circumstances and I and my clients can live with that. The fs5, even with it's artifacts still looks better at 3200 iso then what my gh4 can do.
I'm somewhat on the fence about this, as I feel that sometimes a slight imperfection can be more annoying than another which is quantifiably greater. And there's a lot of psychology in it. By all accounts the "flaw" in the FS5 is slight. It seems the camera fundamentally performs extremely well (certainly for the price!) with high sensitivity. But I do wonder if here - and assuming it is a noise reduction issue - it may be overall more satisfactory to have a higher level of noise but lose the artefacting.....?
Well, easily done (as said before) - allow the noise reduction to be user switchable.
So whilst on the one hand I do feel there may be an issue for Sony to acknowledge and hopefully introduce such a "switch" via firmware, on the other hand there does seem to have been a lot of over-the-top hysteria from some. The issue is slight, and let's just bear in mind only came to light just before Christmas. There seem to be some voices screaming "why aren't Sony doing something about this!?!" - I can only say that a lot of Sony staff may only now be returning to work after the holidays......
And to Christopher Young - you mention "via a colleague I heard of an FS5 being returned tomorrow because of this issue ...." Which begs the question of what they will get to replace it....? (I'm not expecting you to answer that! :-) ) But just everybody think about it. In sheer quality terms, I've no doubt the FS7 may be better, but it's heavier, bigger and more expensive - if anybody is so unhappy with the FS5 issue, then what is on the market as a direct alternative?
Cliff Totten January 6th, 2016, 07:24 PM NOA PUT - "Sorry but that makes no sense, a full frame sensor alone does not automatically mean a superior image quality, if I would follow your logic a canon 5d mark 2 would outclass a Canon c300 mark 2."
The 5D II has a full frame sensor but it line skips heavily. It scans one line, throws away 3, 4 or 5 lines and scans the next. So, it's really only keeps a very small fraction of it's sensor to make a 1080 image. Even with that, it's still not bad in low light. The C300 does a 1:1 readout and uses it full Super35 surface area. The C300 has the surface area advantage.
On the gain topic, yeah I guess we all have wildly different expectations about what makes an image look noisy or sharp. It's all in what you think looks good to your eyes and brain.
NOA PUT - "Sony won't loose much sleep over it, Sony does do small firmware improvements but Sony prefers to do improved mark 2 versions on their bestsellers so you can go out and buy a new camera again and if it's not a bestseller they just forget about it, just like they did with the nex-ea50, or vg900. "
I think this is very true with the Handycam and maybe NXCAM groups. These cameras are a bit less profitable. We know that Sony does significant work on XDCAM updates and we know they are working on FS5 raw firmware. There is certainly significant programming hours involved with that. This is also a perfect time to revisit the FS5 noise reduction process.
NOA PUT - "Not all sony's are great at high gain, my ax100 is very noisy at high gains, same applied for my nex-ea50 and my rx10's image turns to mush at high gains as well."
Again it's all about what we call "high" or "low". What is "high gain" to you? For some that's just +6db. In 2016, I call "high gain" +9db or +12db. For my taste, my AX100 looked fine at +9db and looked good even at +12db. I thought +15db was absolute emergency max.
My FS100 looked good up to +18db, The FS700 looks good up to +15db. The X70 at +15db is "OK". VG20 is OK at 15db. (another big line skipper) My A7s-II? Holy crap, that camera LAUGHS at +18db like it's nothing. Yes, I know this camera is a GIANT exception to the rule.
So for me, I expect the FS5 to at least hold +12db before falling apart. That's not asking much by today's standard.
The FS5? like a ton of other people in the world, I'm super concerned about the samples I see out there. However I WILL ABSOLUTELY BUY ONE at the end of January. I'm going to get that thing onto my charts on my wall, compare it side by side to my other cameras and of course, the most important thing, go out and shoot with it for 2 weeks.
Luckily B&H has an excellent, customer friendly return policy. It's a "no risk" $5,500 purchase. If all this testing reveals that it ranks at the bottom of my camera collection and doesn't even outclass my little, tiny, baby 1 inch-type sensor RX10-II or PXW-X70 in tests....It's going back!
I'm not paying $5k+ for a Super35 camera that doesn't even outclass a small sensor Sony point and shoot with SLOG-2. (I think it should though)
CT ;-)
Noa Put January 7th, 2016, 02:27 AM Cliff, you can click on the quote button or make your own quotes like [ quote ] [ / quote ] (but without the empty space) which makes it much easier to read.
About sensor size, you don't get what I mean, you said the fs5 was completely outclassed by the a7 series because of their sensor size difference, so that would mean a c300 II would be outclassed as well and we all know that is nonsense. A big sensor alone doesn't automatically mean superior image quality.
About gain performances, sony has done well in some camera categories but to put things into perspective, my oldest camera in my gearbag, my little trusty cx730 outperforms my ax100, rx10 and the nex-ea50 I had when it comes to noise reduction so no, I don't think that Sony has done well with all their cameras the past 4 year. The ax100 is my only camera where I have to apply neatvideo to on occasion and I can tell you that is fun on a 4K source. :)
You know, if that one guy had not been staring at his screen waving his camera up and down and discovering a artifact at 3200 iso everyone would be praising the fs5, over at eoshd I see yet another pointless test where for the half part I don't even see what he is seeing but apparently people do start to see ghosts if you tell them a house is haunted.
Maybe I was tired yesterday but I just get annoyed about all this complaining over what not only I but also some seasoned professionals not even find worth mentioning, but that might be because I work with camera equipment that performs worse at higher gains and I am totally ok with that. I guess the fs5 falls in a category that is just reachable for people that have to stretch their budget to get it and expect nothing less then perfection while someone who is used to working with c300 alike camera's and get's the fs5 as a b or c-roll camera understands that it can't be perfect for it's price. I think the problem lies in the segmentation of different camera models and the odd choices camera manufacturers have to make to protect higher end models. One would expect that in a fs5 price range it would automatically get the best codecs and have the best image processing behind the sensor but no, Sony wants you to get the fs7 for a reason so they deprive you from getting that, same as Canon has been doing on the c range but only with a more ridiculous price difference. In that sense Sony makes it easier for you by offering a "better" camera for not that much more. I would be very happy with the fs5 just the way it is now but that comes from a user that had been shooting on a xh-a1 not so long ago where any added gain made the image almost unusable to today's standards. That helps to put things into perspective.
If you have clients that have such high demands, well then get a fs7 or a c300 II, otherwise I"m pretty sure no-one will ever notice that problem people have been screaming about the past days. I"d say, just get the camera and start shooting. :)
Olof Ekbergh January 7th, 2016, 11:12 AM This issue is just greatly overblown, I feel the need to post my experience with the FS5. It really is a great camera for what it is.
What do I mean by that. Well I use it as a replacement for my C100. And compared to the C100 it is far superior in HD, 422 10bit. I never had any problem emulating the "Canon Colors" with any Sony cam, just a bit of CC. Slomo, VariND, and it actually has 4K so...
Is it a baby FS7, NO. And it was never meant to be a 4K Cine cam. The 4K is a bonus. If I need Cine quality 4K I use my FS7.
In all my tests and shooting, I have had the FS5 since early December the images amaze me. I never spotted the ripping problem until I read about it by pixel peepers. And yes then it took me quite a few tries to replicate it. But the only way I can replicate it is very specific and not my way of shooting anyway, so it is almost a non issue in my mind. However I have contacted Sony and got a response from ( I won't name names here) some senior engineers and they are now looking into what can be done to improve this possible NR problem. And I am sure they will have a statement about it in a while and improvements in FW soon, just like they have on many other Sony Cams.
In the mean time I am having a great time using this very compact cam for winter outdoor shooting here in NH, mostly HD but some UHD and I love the Slomo. I can say I actually prefer the cache recording to the FS7 continuous. I am so used to the FS700 way and it uses much less media, you only record what you want, you don't burn through cards like crazy waiting for something to happen. And the media is a lot cheaper than the FS7 media.
I am sure the dust will settle here soon and shooters will just appreciate this great cam for what it is. Another tool in the quiver.
For low light I use the A7s, I only have V1 but it is great at that. For serious 4K I use the FS7 as A-cam and the FS5 is fine as B-cam in UHD as well. I rate the FS5 at 1200-1600iso for Log and use 800 or 1000 iso in cinemas. Upping the gain a bit is OK but if it gets real dark I use the A7s.
Using the FS5 the way I do you will never see the "horrible ripping" you can see blowing pixels up when looking at shaky cam footage shot underexposed with horizontal high contrast lines.
So I for one am just out there using this innovative cam for what it excels at. And I can't think of another camera that has all the great features in such a small and ergonomically great package.
I must say that if you can't get great images from any of the current available cams it is not the cameras fault. I don't just mean Sony cams, even iPhones can make great images, and I have used iPhone video in commercials intercut with FS7 footage and the client never noticed. You need to understand your camera, light well and you need to tell a great story. And you need to buy the right gear for what you are doing.
Keith Moreau January 7th, 2016, 11:51 AM Olof, thank you for your excellent comment! I think that you are pretty much right on in many ways. Do you use S-log or Cine mostly? Or Rec709 ever? I was using Rec709 by default at first for lower light shooting, really based on Alistar's excellent Vocas seminar on the FS5 (recommended to all FS5 users) where he suggested that in low light that S-log isn't a good idea because there isn't usually much dynamic range in lower light conditions and you want to use more bit depth allocated to darker areas. But by nature lower light conditions will require more gain.
However, it seems that the noise and artifacts are the worst in Rec709 and Cine gammas with any gain, even as low as +6db. (I'm going to use db instead of ISO because ISO on the FS5 is relative to the picture profile one is using, which I think is confusing everybody, for example 0db gain at Rec709 = ISO 1000, 0 gain at S-Log2 or 3 is 3200 which implies that the native ISO changes and is lower for non-slog gammas.
But what I'm finding is that for the FS5, using Cine or Rec709 with enough gain to get decent exposure seem to result (at least subjectively) in more artifacts at with +6 to +9 db gain.
For example, I've been shooting a technical seminar for the past 3 days. The first day I used standard Rec709 at ISO 4000 which seems to be equivalent to about +6 to +9. I found the image to be a bit grainy, particularly on the edges, at this gain / ISO.
The 2nd day after studying this thread I switched to Rec709 800% and I found the image to be brighter and cleaner with the same gain. At the end of the 3rd day I switched to Slog3 and used my SmallHD 501 along with a 3D LUT for Slog supplied by them to see what that might look like. I thought it looked very pleasing and relatively clean.
So today, on the 3rd day, I'm using Slog3 and using the LUT along with false color to get proper exposure. I'm at +3 to +9 db gain and I find I'm really liking the image quality at this point when viewed through the LUT.
Anyway, just wanted to report my subjective findings on the matter.
Regards,
-Keith
TechMove is the Podcast all about Digital Filmmaking, Gear, Mac Computers, and iOS Devices from a Unique Perspective! (http://www.techmovepodcast.com)
Ilya Spektor January 7th, 2016, 12:15 PM Olof, I absolutely agree with you. I don't have this camera yet but I am planning to buy it within a couple months and just doing my research now...
...In all my tests and shooting, I have had the FS7 since early December, the images amaze me...
Do you mean, FS5 here, not FS7?..
Olof Ekbergh January 7th, 2016, 12:29 PM Olof, I absolutely agree with you. I don't have this camera yet but I am planning to buy it within a couple months and just doing my research now...
Do you mean, FS5 here, not FS7?..
Ops fixed it thanks.
Olof Ekbergh January 7th, 2016, 12:48 PM @Kieth
I use ISO because a lot of times I actually use a light meter. And I rate all my cams and profiles. So in a multi cam shoot I just take a light reading with the LumaPro and then I can quickly set all the cams the same. A light meter is also a great way to figure out how many stops of DR you need. And it is a lot smaller and faster than setting up scopes. Although I must say that I do often use my Odyssey just for the scopes, it has a great multi cam in function so you can quickly match cams, more cables though.
You are right it is confusing that you can rate either way. Back in the film days it was always ISO, video was always gain. I was happy to go back to the ISO a few years ago as it quickly let me know what amount of light a cam needs to look good. It helps me make lens and lighting choices, before even getting the cam out.
On the FS5 I shoot mostly Cinegamma2 right now, because of an issue with FCPX clipping at 100ire. But if I need more DR I shoot Slog 2 or 3.
Even on my FS7 I use custom most of the time so I can set white point. But CineEI is great when you need it.
Doug Jensen January 7th, 2016, 03:01 PM Olof, if you ever have the chance to rate the various gamma modes I'd love to compare notes with you. I'm 100% certain the displayed ISO on the camera is not correct for several of them but I'd like a second opinion. Send me a private email with your numbers if you ever check them. If you don't that's cool too. Don't waste your time if you don't want to bother.
BTW, with the right settings the camera performs amazingly well in low light. Much better than a lot of people are giving it credit for.
Scott Hepler January 11th, 2016, 03:48 PM I was doing a bit of slow motion testing with the FS5 at 240fps, and was thrilled with the look. I then burned a DVD and viewed it on a samsung 55" led tv and the blocking around each moving image was awful. I was shooting geese landing on a local lake and around each bird was this horrible pixilation so I went back to the NLE and I had to really look but I could sort of see it on my apple 30" monitor. So I decided to do a bit more "testing" and I burned a blu-ray at 1920x1080 and it is as clean as a whistle. Absolutely perfect both on the LED tv and on my plasma. But when I output it to a mov file for youtube there it is, just not as bad as that DVD on a television.
Has anyone who is seeing the pixelation in 4K tried to burn a Blu-ray and view it large? I think I will try to see if I can replicate what people are complaining about and out put some test.
And is there a remedy for a setting to get a clean DVD output? and or a Youtube setting to minimize any pixelation.
It seems like the exporting is having a bit of trouble with these files.
Thanks
|
|