Sony FS100 white balance at depth

Here are some samples from a recent dive showing the halo/banding effect that has been mentioned. This was shooting without a filter. I manually white balanced on a grey slate as I was descending at around 25 feet. I don't recall the temperature that popped up, but it looked good at that depth. I then descended to about 50 feet without touching the white balance. I've used this method with other sony cameras (EX1R and Z1) and while it's not ideal (images are fairly blue), the results are usually acceptable and can easily be corrected in post. This was shot with a 1080/60p PS setting, 0 db Gain, and everything on manual. The effect seems worst in darker areas at F stops of 16 to 13. You can see them come and go in the video clip as the image becomes more exposed and then darker again. You'll notice me adjust the iris manually from F16 to F13 aat around 15 seconds int0 the video clip.

I originally thought the banding was some sort of reflection inside the dome port, as I have had issues with that as well. Silver is a bad color choice for a lens. It's interesting to hear that the effect has been seen topside, as mentioned on this thread about the FS700.

http://wetpixel.com/forums/index.php?showtopic=49724&hl=%2Bfs100+%2Bwhite+%2Bbalance&do=findComment&comment=330239

Under what conditions are you recreating the halo/banding effect topside Simon? Seems like it would be easier to experiment on the dry side if that's possible.

Mike

Hi Mike

I've managed to reproduce the effect topside in dark conditions with the iris was closed up more than it should be for correct exposure and only when using the same MWB setting from a dive. From memory the Iris was wider than F13 that you report, but it was very dark so maybe a similar level of light entering the camera.

I'm actually probably going to get it back in the water tomorrow as I'll be diving at 6m all week in murky conditions and thought I'd give it one last chance. Sadly I tried all my manual lenses but none would fit.

Awesome Simon. The best way to work this out is to get it replicable and work out the causes. Please do try it with/without filter, MWB/AWB/Daylight, and underexposing the image by 1 stop or more.

It's worth seeing if it's the sensor reacting (or not reacting) to the lack of colors under certain conditions, like the LED spikey spectrum effect.

Ok back at home for a day and can report some findings. Unfortunately I can't really post any footage or screenshots for copyright reasons, but hopefully I can give a decent explanation.

The FS100 was used for 2 days diving this week, once by myself and once by another cameraman at an average of 6m.

The colour reproduction at this depth was poor using both MWB and preset K settings when compared to the other cameras used (PMW200 & 7D).

Filters were not used due to the shallow depth.

There were a couple of incidents where the halos appeared and one incident in particular that may shed some light. I wasn't using the camera that day but it was set to MWB, 1/50 shutter, 0db Gain and Auto IRIS. The camera in open water was well exposed, but then the cameraman entered a building where it was very dark. Although the iris opened up it was still underexposed and the halo effect immediately appeared. When the camera was then turned towards a well lit exit the iris closed up and the halos immediately disappeared.

I know this is only one example, but the halos definitely appeared to be visible in dark underexposed conditions. You could claim that most conditions when diving below 10m and when not using lights or when out of the range of lights are underexposed conditions to some extent. This tends to tie in with my other findings, but I'd never seen such a clearly definitive situation where the halos appeared and then disappeared before.

After the second day the FS100 was put away and not used again due to image quality reasons. That really was its last chance as far as I'm concerned.

Cheers, Simon

Hi Simon

Do you know if Sony will look into the halo problem? If such phenomenon could be replicated then they have no excuse but to look into it.

I have no more answers from Sony here other than that the camera should be okay within working parameters. I am talking about the FS700 with the latest firmware and not the FS100. I am not sure if they are hinting at the higher resolution sensor or a different processing in camera as the camera has the ability to pump out RAW stream. One thing is for sure for top side this camera is flying off the shelves here. I think it may be due to a cheaper RAW recorder is soon available.

I am going to have to hire a camera and really check out the footage for myself. I was told FS 700 has better managed colour processing and also that with RAW recording would provide more latitude in low light and underexposed conditions.

regards

David

Dave,

Will be really interested to see your results as am not being 'sold' on this camera at the moment !!! Please let me know when you test the 700 . . .

Cheers

Paul

Hi Paul

Desperation really. I only see footage from other people but to be really sure I will have to grab the bull by the horns so to speak and try it out for myself.

I know from RAW Cinema cameras the S35 larger sensor is not that much easier to get correct white balance UW than smaller chip (APS-c or 3CCD cameras. I have to constantly be aware of exposure and the RGB colour channels. Since it is RAW it is all meta data and all goes into the media recorder pot.

For AVCHD the camera footage is baked in. The Sony cameras from consumer CX to NEX has always been tricky for MWB. Need to trick the camera into not extrapolating too much red channel or green at depth. The Cyan or Magenta cast is always problematic with presets. I once got terrible reds dancing around even at two metres depth. At least with RAW with 16bit bandwidth (even 10 would make a big difference) there are lots more information in the footage to correct in post. Compared with 8 bit compression.

Well I am planning to go for a long weekend and try and shoot some wide angle at various depth and see if all the issues discussed so far manifest themselves with the FS700. I would also try shooting with video lights. Balance with that and switch off and then see what happens. I won't say more until then.

cheers

David

Very interested to see your testing results David.

I should add that the halos appear to be on most images even when they are not immediately visible. They tend to become visbile when you colour correct and move the image towards orange/red/magenta even when you thought it was a well exposed shot.

I will email Sony again with my latest findings. Hopefully they will come back to me at some point!

Cheers, Simon

Here are some clips that were shot with the Sony FS100 in a Genesis housing. I will admit there is a learning curve towards these new super 35mm sensor video cameras and once mastered they can and will achieve excellent underwater imaging.

https://vimeo.com/49757188

https://vimeo.com/46491638

https://vimeo.com/46487991

https://vimeo.com/46487990

https://vimeo.com/61721292

https://vimeo.com/39102598

https://vimeo.com/58836639

Have you not read this thread Joe??

Thanks for the links Joe. It does show that the issue is when the MWB is pushed toward R, some sort of ring banding occurs.

The banding is clear on the right side of the pic.

Again it is clearer on the right side but it is faintly on the left side.

What's stranger is the interlacing artifacts on the coral.

Initially, I thought this issue was depth related. Looking at the videos you posted, it seemed they were shot in 60ft (18m) or less. If so, the conjecture may now have to be shifted toward the MWB and S35 Exmor sensor's sensitivity to RG.

Have you not read this thread Joe??

Eventhough Joe does have a vetted interest, I think it was ok to remind us what is possible in certain lighting conditions. Each and everyone may estimate is his/her diving range (shallow water) and video shooting needs similar to the samples. For example macro shooting under well-lit artificial lighting might be totally ok. And whether the picture quality, tonal range and slight red rings are acceptable or not.

These samples also confirm that even with multiple different shooters the red rings might and do appear. I was initially concerned that the problems might have been related to individual cameras and/or personal settings on the camera.

Hopefully Sony will fix this.

This at the end of the day is still crap for a camera set up that costs this much money

Each and everyone may estimate is his/her diving range (shallow water) and video shooting needs similar to the samples. For example macro shooting under well-lit artificial lighting might be totally ok. And whether the picture quality, tonal range and slight red rings are acceptable or not.

 

I'm not sure if shallow depth will alleviate the issue. It may just hide it enough that most don't see it. But of course you are right, in the end, if people don't mind or see it, it works then. If it's a MWB/S35 EXMOR sensor issue, Sony may not be able to fix it even if they wanted to.

If it's a MWB/S35 EXMOR sensor issue, Sony may not be able to fix it even if they wanted to.

Do Sony ever fix issues in the existing range? Especially if it only affects a small subset of their customers. Or, due to planned obsolescence, do they usually wait to fix it in the next product?

This problem, or something very similar, has been around since the CX550. Maybe something in the firmware of their colour engine that only impacts under certain lighting (very low red) conditions, does not impact topside shooters and keeps getting copied across as they develop new models. Looks like the color engine is increasing the gain in the red channel.

Can gain be locked in the FS100?

Can gain be locked in the FS100?

Hi Peter

All of my recent examples have been with the Gain locked at 0db and the halos still show up.

Cheers, Simon

Eventhough Joe does have a vetted interest, I think it was ok to remind us what is possible in certain lighting conditions.

As Drew pointed out even the examples Joe posted from shallow water show the halos and the really awful colour reproduction / WB issues that we have been talking about which is why I asked if he had bothered to actually read the thread. I've been so busy lately and I have a huge backlog of work but when I find the time in addition to the Socorro footage that I have already posted I can post a FS100 showreel of what it is capable of in certain circumstances. Sadly even the good examples are inferior to cheaper and similar priced cameras when used under the same conditions.

I've possibly used this camera underwater more than anyone else in water conditions ranging from 30oC tropical to 8oC temperate and I can tell you that this is not a steep learning curve issue to do with a S35 sensor - it is down to fundamental technical flaws in the cameras image quality when used underwater. We have talked in depth now about the halos and the poor colour reproduction / MWB issues but there are others. The DR underwater seems to shrink beyond belief and you get horrendous digital noise in white highligts - perhaps I'll dig out an example of that and throw another can or worms into the FS100 pot.

I too have a vested interest in 'promoting' the FS100 as I am trying to sell my Genesis housing, but on balance I feel that I need to talk about these issues to let others know rather than trying to paper over the cracks or ignoring them as others appear willing to do. Hand on heart I couldn't do that to people when I know differently and they are about to spend a large amount of money getting these cameras underwater. A casual user may be ok with all of the flaws, but this is a pro/semi pro camera and should without a doubt be achieving a whole lot more than it appears to be capable of. It is probably costing me quite a few thousand dollars by saying all of this so believe me I do not say it lightly.

Cheers, Simon

Do Sony ever fix issues in the existing range? Especially if it only affects a small subset of their customers. Or, due to planned obsolescence, do they usually wait to fix it in the next product?

 

This problem, or something very similar, has been around since the CX550. Maybe something in the firmware of their colour engine that only impacts under certain lighting (very low red) conditions, does not impact topside shooters and keeps getting copied across as they develop new models.

Well Sony has stepped up in the past with fixes for issues, but just can't think of one off the top of my head.

The more important issue for Sony is that IF this is a deficiency with the sensor line, then it's big problems for them and the housing manufacturers who support them. Those color rings don't look like gain noise but some sort of channel clipping that's not on the entire sensor, which is very weird! It would do well for the housing manufacturers to bring this up to the Sony reps themselves.

I wonder if the red rings would be the combined interaction of the following:

a) internal sensor reflections (see extreme samples below shot against bright windows). Similar problems have been reported in low light conditions

b) the round dome

c) unbalanced for, reddish light from surface spilling from almost directly above and somehow reflecting via dome and sensor reflections

d) the (exagerrated) WB red gain of the camera.

If it is, you could try mocking up something like a temporary mattebox top and side hood on the outside of the dome. Or use another diver to cast his/her shadow on your dome. Eventhough a permanent solution might require more creativity, it would be interesting to see could the red ring problem be eliminated this way. The other remaining problems are color space/profile/WB related.

http://vimeo.com/36858333

http://vimeo.com/37389803

Ok I've put a couple of examples together and compared them to footage shot on a 7D during the same dive. I've listed a detailed breakdown below.

https://www.youtube.com/watch?v=xFScY98itrM

This is a short video to quickly demonstrate a couple of the image quality issues discovered when using a Sony FS100 underwater. I have included the 7D footage as a comparison and to give an indication of the results that the FS100 should have been able to produce. In summary these are the issues currently found when using the FS100 underwater.

- An Orange/Red/Magenta 'halo' effect on most images when the camera is manual white balanced underwater. This becomes more pronounced the deeper you go or the darker the conditions.

- Very poor results from MWB with some very, very strange results when used below 10m.

- Very poor results when using the preset K settings for white balance with the colours shifted towards Green/Cyan.

- The images at depth appear to have had their saturation boosted even if you reduce your saturation with picture profiles.

- A lack of Dynamic Range

- Digital Noise on highlights/blown out areas (which is not helped at all by the lack of DR).

- While the FS100 shines topside in low light underwater it tends to fall apart. It is quite acceptable to shoot up to 18db Gain topside, but anything other than 0db Gain underwater shows lots of noise. What looks organic, even filmic topside looks downright nasty underwater.

- Impossible to colour correct / colour grade in post due to technical issues with the footage. Basically as soon as you try to colour correct and move the images towards Orange/Red/Magenta the halos start to appear even if they were not visible before.

All the footage used in this video was shot at Yolanda Reef, Ras Mohammed Marine Park, Egypt on the same dive. Light penetration in the Red Sea is excellent and it generally makes cameras look better than they actually are.

No filters were used on either camera.

Example 1:
Canon 7D, 160 ISO, Tokina 10-17mm.
Apart from the scratch on the dome port I was very happy with this footage. The colours look great which is to be expected from the 7D and although the resolution is lower than ideal it is a trade off I'd happily make. Remember this is straight out of the camera with no correction in post.

Example 2
Sony FS100, 0db Gain, Sony E 16mm
What can I say. The same depth on the same dive and what different results. The reason I posted this example specifically was to show the lack of DR that the FS100 shows when used underwater and the awful digital noise that can be seen on the white toilet. The Iris was closed down to assist the highlights, but the toilet still appears to be blown out. This clip also shows the unsatisfactory colour reproduction with a clear bias towards cyan/green. This shot was manual white balanced off of the sand around the Yolanda Wreck, but I have attempted many other methods all with similar or far worse results.

Example 3
Sony FS100, 0db Gain, Sony E 16mm
This is the same clip as the previous example, but an attempt has been made to colour correct. The results are an improvement, but quite frankly the image still looks awful. The digital noise is still shown on the toilet and although the toilet now looks white the cyan colour cast in the rest of the image has been replaced with Magenta. Due to the colour correction the halos have become visible.

I have shot nicer footage than this on the FS100, but this gives a good example of the challenges faced when using this camera underwater. I genuinely believe that 90% of the footage I've shot is a waste of hard disk space purely due to technical flaws.

Cheers, Simon