Comparison of Internal Recording on FS5 vs. RAW output to Odyssey 7Q+

[Doug Jensen]

Thanks. Edited in 1080 so the 4K is down-rezzed in post. And I have tried to be as fair and neutral as possible to both sides of the screen.

[Cliff Totten]

Doug, I'd love to get your candid opinion of your own results. I have talked to you at the Sony booth at NAB and I know your relationship with Sony, so I don't mean to put you on the spot but,.....

We can talk about the wonderful factors about the FS5, form factor, ergonomics, professional features,...the list of great things is endless. However, based on your results in your video, (and other videos you have shot) I'm curious to know your opinion on just TWO categories: "Noise" and "4k resolution".

I LOVE the FS5 for dozens of reasons but I'm disappointed in it's noise level and what I think is a very "soft" 4k image. Seeing your video just confirms my feelings with the footage I have shot myself.

It seems to me that Sony took the FS7 and tweaked it down on the FS5 to give it market separation but was very keen to keep it competitive with the C100, BlackMagic models and JVC LS300. So, for it's price point, the FS5 easily beats those guys with no trouble. The problems come when you compare the FS5 against Sony's own state of the art image sensor technology in their Alpha division. The FS5's 6 year old sensor starts to show it's age. Ouch.....

Here is small part of a larger FS5 test that I did against the A6300:

SONY FS5 vs. A6300 - "IMAGE SENSOR" Test. - YouTube

Resolution and noise...what do you think?

CT

[Doug Jensen]

Cliff, that is an interesting test and I'm sure that the FS5 is on the left in the daytime shot and on the right on the nighttime shot or you wouldn't be posting the test at all. Yes, there is a big difference and in those tests the A6300 is the clear winner by a wide margin.

But, here's what I think. You either have a sick FS5 or you're using settings that aren't right. If my FS5 looked like that I'd send it in for service. In fact, if my FS5 had looked like that when I first bought it I would have put it back in the box, returned it to B&H, got a refund, and would never have wasted time producing a training video for that camera. My number one rule when choosing cameras to produce training videos for is that I have to actually like the camera.

Take a look at the first 60 seconds of the opening montage from my training video. I think the FS5 looks pretty good in all those shots -- and I have done nothing special to doctor them up.

Doug Jensen’s Sony PXW-FS5 Master Class - CHAPTER ONE FREE! on Vimeo

Look at the right-side of the screen in this test footage. I think the FS5 looks great and way better than yours. I think there is something wrong with your camera.

Sony PXW-FS5: Comparison of Internal Recording vs. RAW output to Odyssey 7Q+ (PART 1) on Vimeo

Sony PXW-FS5: Comparison of Internal Recording vs. RAW, SDI and HDMI output to Odyssey 7Q+ (PART 2) on Vimeo

I also do not buy into your theory that Sony purposely dumbs down some cameras so they don't look as good as more expensive models. I have never found that to be the case and I think you are dead wrong there. Is there a visual difference between some of the cameras? Yes, just as there is a performance difference in different models of cars from the same manufacturer. Is a Ford Focus going to perform as well as a Mustang? Nope, but it doesn't cost as much either and nobody expects it too. Of course the A6300 looks better than the FS5 in your tests, so the analogy doesn't work that well. But I have nothing but good things to say about my FS5's image quality.

One of these days I'm going to find the time to compare the FS7 and FS5 on the exact same shots using the same settings and lenses and then we can see what the differences truly are. But my opinion right now is the differences are negligible. I prefer the FS7 over the FS5 for a lot of reasons and I think it is the much better camera, but that is for other reasons other than image quality.

I'm sorry to say it, but i think it is YOUR camera that is at fault and tainting your whole opinion of the FS5.

[Cliff Totten]

I must admit, comparing the FS5's 1:1 pixel readout to an A6300's 6K oversample, a 2:1 pixel readout is not a "fair" fight.

The A6300 takes in 16 megapixel RGGB video image to start with. That's literally 8 million green and 4 million red and blue. This is the very latest generation Sony readout processing technology. Sony has specifically designed this for MAXIMUM 4k resolution. It literally resolves frequencies right up to the very highest Nyquist frequency sampling limit of UHD. It's new and it's brilliant. This is not "normal" and is the only camera that Sony markets today that does this. It's possibly the only camera in the world today that does this for 4k recording. (The A7r-II does a 1.8x 4k oversample and the new A99-II also does a 1.8x 4k oversample....two other extremely detailed 4k Sony cameras)

All the FS5 footage I have seen to date is exactly like mine. The settings I used for this were picture profiles OFF for both cameras. I didn't want Picture Profiles to be a big point of debate. Even in your footage, there is very little difference between the 4k side and 1080 side of your tests. Granted, this is Vimeo and you reduced the entire project to 1080 anyway. Generally, I have found that 4k downscaled to 1080 WILL produce more detail than native 1080.

Forget my video. Looking at your video. What do you see? Can you change your timeline to a 4k timeline and compare that 4k side to the 1080 side?

Better yet, if you still have your Z150, do a side by side with that camera. That sensor takes in about....13 or 14 megapixels in 16x9?...using Sony's full pixel readout technology.

So far, ALL the FS5's that I have seen from ANY FS5 camera has those two things in common. "Soft" (compared to Alphas...not FS7 or FS700) and "noisy" in higher ISO's.

Dont get me wrong....the FS5 all around is a GREAT camera that is a joy to shoot with. What I call "soft" might be perfectly fine to many other people. I'm just a pixel peeping junkie,...it's true. ;-)

I'm talking about this NOT as a FS5 vs A6300 "camera" comparison. This is strictly an "image science" discussion. These two camera's are not ones that are in the same league to be compared. They are so far different, it' s not funny. Agreed! We are just talking "science" here.

I do disagree with you about Sony tweaking image quality for each price point. I believe the FS100 shared the same sensor as the F3. The FS100 was softer on charts than an F3. Sony gave the FS100 a "thicker" low pass filter that softened the image and reduced moire. The F3 had a tiny bit more moire but was way sharper because it seemed to have a higher frequency low pas filter. It would have been a terrible marketing mistake to give the much cheaper FS100 the same detail and resolution as the F3. (even though they shared the same sensor) A simple production way to fix that is simply use a lower OLPF on the FS100. Bingo, problem solved....softer image and you kill moire too at the same time. The Canon 5D DSLR crowd will love that in particular. Right? (that was who the FS100 was marketed to anyway)

The original FS5 firmware had terrible CODEC banding and noise reduction flaws. It's my guess that they were too aggressive on tweaking it down early in development, released it out like that, got caught with their pants down and raised it back up.....fast. Real fast. No, I dont know this as a fact and yes, I could be overly synical or paranoid about it. I admit that. ;-)

Sony does tweak each camera's image for it's price point....and just that!...no more is allowed. All the companies do and this to an extent is a normal part of the camera marketing process.

I dont own an FS7. However, it's been said that the 8 bit, long GOP recording on the FS7 is better then the same 8bit Long GOP on the FS5. Especially in higher ISO's If this is true, with both sensor's being the same and the CODECs being the same....what is the FS7 doing that the FS5 is NOT doing? The raw sensor data in the front end of both cameras is the same, yet the 8bit FS7 supposedly looks better? (I have seen online shootouts...nothing I have done myself) Again,..this is 8bit for 8bit UHD. Not a 10bit vs 8bit shootout.

You have both,..you could try that test.

CT

[Cliff Totten]

Doug, here is another video of the A6300 vs the FS5 that shows the exact same results as mine. These are not my two cameras and it wasn't made by me.

A6300 VS FS5 - Image Detail, Lowlight, and Rolling shutter comparison - YouTube

Again, the A6300's 6K readout science is really the star of the show here. It's ability to read 16 megapixels of video, 24 frames a second, DeBayer it, noise reduce it and scale that down to 4k really shows. It even makes "splotchy" noise "smaller" and "finer" too.

CT

[Doug Jensen]

I woke up this morning to find that my shoot was cancelled due to weather, so I decided I could spend an hour shooting some 240 fps test footage with the FS5 and 7Q+. Nothing fancy, just some stuff around the kitchen and garage. I think the quality looks very good and almost on par with slow-mo footage from my F55. Grading was done in Resolve and included a healthy dose of noise reduction to clean it up.

Sony PXW-FS5: 240 fps Slow Motion test footage recorded with Odyssey 7Q+ on Vimeo

[Cliff Totten]

@Doug,

Here is more independent evidence that shows the advantages of SONY's new amazing image sensor 6k oversampling techniques. This is the FS5's five year old 1:1 readout against SONY's A7R-II that uses a 1.7x 4k over sample data collection and scales down to 4k. It's important to remember that the A6300 samples a FULL 2x 4k readout (6k) which is even higher than the A7R-II.

If there is one thing that SONY's is teaching the industry in 2016 it's this: Having allot of pixels on a sensor is one thing....but how you collect those pixel voltage readings, de-Bayer them, process and scale them generates a very different result. Many companies have simply used line skipping, pixel binning or taken the other "easy" approach of simply using the bare minimum amount of RGGB pixels to make 4k. (about 8 million full color pixels made from 4 million green, 2 million blue and 2 million red)

1:1 pixel readout is simply the rock-bottom bare minimum pixels required to call your video a certain resolution. What SONY is now developing is so much more brilliant. Doubling the sampling amount allows them to START with a green channel that is already 8 megapixels...full 4k faster...and build the red and blue channels around that maximum resolution "foundation" green channel. Your classic 1:1 readout only starts with a green channel that's 4 million pixels...only HALF 4k raster supported by two 1/4 raster red and blue channels.

The fact that Sony now has the processing power to do this is staggering. You really got to tip your hat to their engineers on this. Killing noise in the 6k world and down-sizing it to 4k is again, insane. Imagine the possibilities!

The FS5 sensor was designed in 2012 for the original FS700. Look how far Sony sensor and processing architecture has come in 4-5 years! Wow!

This is not my video or my FS5:

Comparing S-Log2s: Sony FS5, A7R2, RX1004 - YouTube

CT.

[Doug Jensen]

Well, it sounds like we have some great sensor performance and technology to look forward to down the road when/if it reaches the next generation of camcorders. But in the mean time, I have zero interest in shooting professional video with a DSLR or mirroless camera just for the sake of some slight improvements in picture quality in someone else's footage. it doesn't matter to me. I am not willing to ditch all the professional features, functions, and ergonomics of a true camcorder for slight improvement in image quality. And you know what, I am quite certain that I could post some video where the FS5 blows the doors off those mirrorless cameras in other ways that would be even more important to me. Care to try tracking a flying bird with manual focus and a 300mm f/2.8 lens? I am quite happy with image quality from my FS5, FS7, and F55 using my custom camera settings, high quality lenses, good lighting, proven shooting technqiues, and grading workflows in post.

The intent with the testing I did with the FS5 was to see if jumping through a bunch of hoops and spending a few thousand dollars for hardware and software upgrades was worth it or not. And I answered that question for myself already through these tests. But when it comes to comparing the sharpness or noise levels of the FS5 to other types of cameras -- I have no interest whatsoever.

[Cliff Totten]

Doug -"Care to try tracking a flying bird with manual focus and a 300mm f/2.8 lens?"

Well, not sure how much you have followed SONY's recent engineering acomplishments with the new A6500. It has the fastest, most accurate phase detection and object tracking auto focus in the world. Even more amazing is that they have given this incredable sensor a 5 axis stabilization base. It "floats" in the body like a gimbal.

On your new A6500, I'm sure that you will track your bird on your 300mm lens with amazing pinpoint, razor sharp acuraccy and with stabilization smoothness that no XDCAM body has ever done for that lens today. That bird will be cleaner in high ISO and you will see even the thinest detail in its feathers!

Doug....I love XDCAMs...always have, always will. Tried and true, trusted hardware with ergonomics that you can always count on in battle. They are a true warriors tool.

I will agree with you for eternity on this!!! I completely stand with you.

However, SONYs engineering department is moving at an absolute lightning speed right now. Sony R&D is leaping over the entire industry like a kangaroo on steriods and Viagra. They have nowhere to install their new technology!! XDCAM life cycles are 3 to 5 years turn arround. XDCAM moves at a far slower pace. So the ony place Sony can implement this incredable, industry changing work...is Alpha!!

Alpa gets the most high tech sensors. They get aluminum wiring, BSI, and fully stacked sensors. They get the latest micro lens archiceture and they get RAM now mounted onboard the sensors for additional speed and cashe....and 5 axis stabilized sensors! Alphas are now being updated evey 11 to 18 months!!!

its a very wierd marketing problem for SONY. The argonomics are all in XDCAM. The latest tech, low light performance and 6k processing is all in Alpha!!

I hope XDCAM in the future can find some way to move to a faster refresh schedual and start adoping more of SONYs newest technology, and processing. The FS5 body is spectacular...its 5 year old sensor?...I dont know.

Lets see what this new FS7-II announcement is going to bring next week.

[Doug Jensen]

Obviously there is always going to be something better just around the corner. So what??? I don't waste my time or attention thinking about that kind of stuff until it can actually be purchased in a tangible product. And even then I still might decide that I don't care enough about it to buy it.

And if the idea of a great auto-focus system appeals to you, then I'll bet you also look forward to self-driving cars and other automation that takes the fun out of living. Just as a I love to drive, I also enjoy operating my cameras manually. So, I have no interest in either technology no matter how good it performs. I absolutely hate being a passenger in a car, and I'm sure not going to become a passenger behind my camera and watch it do the work that I enjoy.

You know, if I didn't already own an FS5 or FS7, I guarantee you that I could buy one tomorrow and have it completely paid off and earning excellent money day in and day out for many months before your dream camera is ever officially announced, let alone shipping. And would my clients even give a crap about the new camera? Not a chance.

[Cliff Totten]

I hear ya Doug,

It's true that everybody has their different interests and different things that makes them say "WOW!"

I was talking to an ARRI guy a few months back and we were discussing that his camera only has 2.7k sensor and we were talking all about 4k. All he could do was tell me how much he "hates" 4k today. He ended it by saying that he will be happy to shoot 1080 until the day he dies.

Although I'm completely different form him, I can respect his opinion. Allot of guys are just happy to do what they do everyday, are comfortable with it and have no love or desire to want anything different. That's cool. I get it. We all do this stuff for different reasons.

As far as auto focus? The A6300 is the FIRST camera that I have ever used in my life,....where I actually LIKED it's auto focus. When I first got it, I was in my back yard opened up with ND and a fairly shallow depth of field. I was chasing my two dogs that were running at and away from me at full speed. I was BLOWN AWAY at how fast that A6300 locked onto those pups. I tried to manually pull for a few minutes and I did pretty good but I'll be damned if the AF shots weren't faster and considerably more accurate than my best manual labor. (not always but mostly)

That night I fell out of my chair in front of my monitor watching allot of what I shot frame by frame. All I could think was: "OMG,...this the beginning of a serious shift in what can be possible from here on in."

Doug, before last year, I never would have expected that kind of AF would have been possible....and I never expected that I would have every "liked" using it..

For me, I like to be "wow'd" and I dig tech. I guess I share a similar sense of wonder that many Sony engineers have. I'm sure they love pushing the limits of imaging science.....and I LOVE, LOVE,...L O V E... seeing them do it.

I'm guilty as charged, Doug! ;-)

CT

P.S. No,...AF will never completely "replace" MF.....ever.

[Doug Jensen]

Cliff, I appreciate your enthusiasm for the new technology and I do want to keep an open mind on the things that are coming down the pike. I certainly don't want to be the old guy that says "I've always done it this way and I don't need anything new." But with that said, I don't care what is on an engineer's drawing board until I can buy it, and I look at anything new with some skepticism until I see what it can do for me. I'd gladly write a check right now for several thousand dollards if I could retroactively add Variable ND to my F55 and FS7. That is an example of new technology I eagerly embrace.

One thing is sure, we certainly live in interesting times. Long gone are the days when a Betacam, lighting, and post, went basically unchanged for years and years.

[Ole Salomonsen]

Hey, great thread.

Any chance we can see a lowlight shootout between the FS700R and the FS5 ?

I used to have the FS700R, but had to sell it. Now I shoot a lot in low light conditions using the A7S Mark II / A7S, often using ISOs up to 32000 or more. Even if the A7S has the edge over the FS700R, I remember the FS700R to be REALLY good in low light using correct CINE profile (but stay away from SLOG if noise is a concern). Particularly when shooting RAW on the O7Q recorder.

Now I am considering gettin back a used FS700R, or possibly a new FS5.

But looking at the low light shots of the FS5 I have to say I am really disappointed!
I could get REALLY clean shots from the FS700R up to ISO25600. The FS5 looks smudgy, soft and has more noise. WHY? I know it has the same sensor? Why not tune it at ISO2000 like the FS700 and not ISO3200?

It's been a while since I had the FS700R, but still...
The best would be to see a shootout between the FS700R. I will definitely not be buying the FS5 only for the convenience factor if the FS700R has better IQ.

[Cliff Totten]

Ole,

You ask some great questions here. Yes, the FS5 uses the same sensor and raw readout technology as the FS700. They essentially took an FS700, gave it and awesome new ND filter, new XAVC-L CODEC and put it all into a spectacular, well designed, ergonomic body.

However, so many people were stunned by Sony's decision to raise the minimum ISO in SLOG to 3200 over the FS700's 2000 ISO. That really upset a lot of people.

You have to remember though,..when the FS700r was developed, it was a "stand alone" Sony camera in it's price range. The FS700 enjoyed that fact that it had no FS7 above it to cannibalize in it's day. (The FS7 had not been invented yet) So Sony seemed to give the FS700 everything it could in those days...it was fully optimized for image quality.

Fast forward a few years and we now have the FS5 and the FS7. These two cameras can clearly overlap markets if Sony is not careful to separate them. The FS7 is a GREAT selling camera for Sony and they certainly want to keep that model protected form the FS5 as best they can.

So, in my opinion (and many others too), Sony carefully made sure that there were strong image quality reasons to favor the FS7 over the cheaper FS5. I believe noise processing was one of them. Sony seemed to allowe noise in to the FS5 that it does NOT allow on the FS7. I also suspect that the FS5's low pass filter is lowered on the FS5 which slightly "softens" it's image over the FS7 and even FS700's image. (low pass theory is only a guess) Yes, the FS7 (and I believe the FS700) is clearly cleaner and sharper than the FS5.

Good news though.....Sony has officially decided to return it's minimum SLOG ISO back to FS700 levels!!! The next FS5 firmware WILL allow it to have the original FS700's 2000 ISO!!!

I'm really glad that Sony had has reversed it's original "crippling" decision and restore it to where it always belonged in the first place. I think they were just "over zealous" about protecting the FS7 and got carried away. Also, that original FS5 1.0 firmware was absolutely noisy as Hell. It's CODEC was shearing edges like crazy. So many people said "yuck!" that Sony publically aknowledged it and decided to clean it up on the next release.

Interesting though....it's internal XAVC does not do the FS5 justice. In my opinion that CODEC is badly tuned down. Plain HDMI ProRes recordings look noticeably better that's internal CODEC and raw recordings look very FS7-like at at the same ISO.

So,...two cameras take the same sensor's raw data....pass it through different processing and record to XAVC with very different noise results. Take them both raw out to a Shogun and they look about the same. The only question one could ask is....what in the Hell is the FS5 doing to it's image processing between raw to H.264 that the FS7 is NOT doing? An Atomos Shogun doesn't "tweak down" either cameras raw image processing and you get similar image.....right?

BTW,...we are hearing strong rumors that Sony is about to update the FS5 with a new model already. Dont do anything until NAB this month.

One other thing. Its completely and thoroughly unfair to compare the low light ability of the FS5 against a complete low light monster that is the A7S-II. The A7S-II uses a much larger FF sensor than the super35 chip on the FS5. The A7S-II sensor is also a much newer generation architecture than the FS5's 5 year old chip. Its unfair to compare the A7S-IIs signal to noise ratio to ANY Sony camera...FS7, F5, F55 or even F65! Nothing will touch the A7S-II in that catagory.

The FS5 is designed to compete against BlackMagic, JVC and Canon models...which it does very very well.
CT

[Ole Salomonsen]

Cliff, thanks for your thoughts and great news!

I will definitely wait and see how the FS5 performs with the new firmware, and wait for the new FS5m2 (possibly 8K?) before I make a decision.

What is the base/minimum ISO on CINE2 / CINE4 profiles on the FS5 ?

Yes I know you cannot compare apples and oranges, but we still do ;) The file quality of my FS700R I remember to be fantastic on my Odyssesy 7Q. Even if I did not get to use it as much in low light as I planned before I had to sell.

The XAVC is an excellent intermediate codec. But with the FS700R (AVCHD) I shot Prores externally on the O7Q. I have done extensively comparisons between ProRes and XAVC, and IMO the XAVS has just as good IQ, if not slightly better. So even if I will be using the O7Q+ for slomo on a possible FS5, I am hoping to get good quality 4K XAVC files from it.

I use the A7S/A7Sm2 extensively in my work, and also did a few projects with the FS700R.
Now I miss the quality of the RAW files, and the slomo capabilites like 4K@120p of the FS700R.

Some of my A7S low-light work:

Aurora Living 4K - Life under the Aurora on Vimeo

Lights for friends - 4K on Vimeo

FS700R:

FLOW on Vimeo

To be honest I really hope the FS5 image quality will be upgraded with the new firmware to where it belongs. The FS5 will fit so much easier in my Gimbal setup. But it will just feel wrong paying so much more for an FS5 if the old FS700R has superior IQ. I will wait and hope.

And perhaps someone can show a comparison between the FS5 and the FS700R, will be really interesting.