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

nyvz said:
The idea that there is only marginal difference between 50/100Mbps XAVC and 200/800Mbps Prores seems like a matter of different perspectives.
Click to expand...

It's knowing that the FS5 and FS7 RAW output is 12bit Linear which nobbles you from the very start and is no way equivalent to the output of higher end cameras. At worst you get horrible noisy banding in the darks tones at best you're limited in dynamic range so the better codec doesn't necessarily help.
 
Nirv said:
It's knowing that the FS5 and FS7 RAW output is 12bit Linear which nobbles you from the very start and is no way equivalent to the output of higher end cameras. At worst you get horrible noisy banding in the darks tones at best you're limited in dynamic range so the better codec doesn't necessarily help.
Click to expand...

My understanding is that the implementation of RAW on the FS5 and FS7 are NOT the same. But I'll let someone else explain exactly what those differences are.
 
Nirv said:
It's knowing that the FS5 and FS7 RAW output is 12bit Linear which nobbles you from the very start and is no way equivalent to the output of higher end cameras. At worst you get horrible noisy banding in the darks tones at best you're limited in dynamic range so the better codec doesn't necessarily help.
Click to expand...

Do you know that? My understanding is that all these cameras encode the RAW with a log curve to fit it over the 12bit SDI signal. Do you mean the specs of CDNG in particular?

I know the FS7 had some issues that may or may not have been full fixed but from my experience with FS700 RAW is that none of what you are talking about is an issue at all. Seems likely the FS5 may have more in common with the FS700 in this regard.
 
nyvz said:
Do you know that? My understanding is that all these cameras encode the RAW with a log curve to fit it over the 12bit SDI signal. Do you mean the specs of CDNG in particular?
Click to expand...

According to the brains trust on this forum the FS7 outputs 12 bit linear RAW. There is a long thread about FS7 RAW with examples of banding and crapness. I assumed the FS5 was the same if it isn't I'd like to know what the difference is.
 
Funky cool test video Doug! Raw video might be a holy grail and when you finally hold it in your hand it doesn't look as shiny as you would have though.
 
As some people mentioned directly or indirectly, all RAW isn't created equally.

Companies/engineers may also not be providing the best possible output from different cameras for various reasons. It's also probably complicated stuff.
 
Dear Dough,
if someone prepares some tests then there will be some limitations always. Simply because the design of the test is an answer to the question one had in mind. And if the question was wrong then the test will be of limited value only.

The right question should be what is the final result of shooting and editing the footage. Regardless your unfriendly comments it is no valid approach to compare footage that was refined in the signal processing of a camera with footage that was not refined. You know that this is the same with the Cine EI slog3 footage of the FS7. It is noiser simply because it was not denoised! So the comparison is hard.

And I also do not take the point with your 1000$. While the denoiser
in Resolve is great (and I do not share Mitch point that good denoising is simple) there are other packages available that are cheaper (e.g Hitfilm what has a denoiser or BCC). So sure it can be done (and is done in real live).

But more important you miss the point that it is never ever the target to get the final footage from the shoot when recording to raw. You know quite well that you have to grade that. While it should be the target to use the 8bit footage from the FS5 at the best as it comes from the camera. So to exclude the impact of grading is not the best idea.

And it was said earlier that it would make sense to have more critical movement in the shoots. Sure the H264 XAVC (not I) version in the FS5 is a powerfull codec - but it is neither XAVX I nor ProRes nore CDNG.

Overall the test requirements for all of that are not so simple. That is the reason why we do not have good tests for such an end-to-end comparison up to now. And it is no shame to state that someone do not want to test that because it is not simple but hard work. But simply to ignore relevant factors is also no solution. That is my approach and I would appreciate a friendly language since we are no boys but grown up mens and are talking about technical aspects only.
 
@wachmid

If you find Doug's tests so inadequate an option would be to conduct your own then you would know they've been carried out to your exacting requirements.
 
Nirv said:
@wachmid

If you find Doug's tests so inadequate an option would be to conduct your own then you would know they've been carried out to your exacting requirements.
Click to expand...


Well, he wouldn't even have to do that if he would have bothered to accept the offer I made in post #17 to tell me exactly what/how he wanted me to shoot something for him. I was willing to invest my time on his behalf, but he ignored my offer and didn't give me any specific instructions. Apparently, there was something unfriendly in my tone and I have offended him in some manner so he just wants to tell me why I'm wrong. I have no idea what exactly he is referring to in my post, but since I'm not accustomed to having to walk on egg shells in a professional environment, and just to avoid any further offensive comments, I'll rescind my offer and move on.

BTW, just for the record, my name is not Dough no matter how many times some wants to misspell it in this thread.
 
Last edited:
morgan_moore said:
Doug

It seems (to me) that to some extent you miss the purpose of raw whose strengths (if there are any) would only show when the footage is given an extreme manipulation or some other codec stress. It is not only operator error that stresses a codec.

-frizzy hair on green screen
-matching two shots either shot in different locations or during seriously changing light
(eg. inteview of two CEOs, one in LA, on in NY)
(long interview as sun drifts in and out changing ambient colour temp)
-tracking shot (eg flouro kitchen steadicam to tungsten eating area)
-Heavy 'brand colour' look
eg all LLoyds bank images are green - but a rebrand might require them all to go pink!
-huge DR - room with sea view
-fashion grey backdrops with subtle tonal ranges
-complex architectural scenes, that may need tombstone correction.
-complex architectural scenes
-fibre, fabric

Obviously you would not be able to recreate these shots during a no budget test.
Click to expand...

Some good ideas there, but as you say, not within my resources to test. Besides, if the power of RAW is so impressive, we shouldn't have to setup such specific tests and squint at the monitor with a magnifying glass to see the difference.
 
Doug Jensen said:
.. specific tests and squint at the monitor with a magnifying glass to see the difference.
Click to expand...

Not at all. One needs to do 'stress' tests because raw should only start to shine (if it does shine) at a point where a regular codec would break.

A simple test being shooting a scene under tungsten and then correcting it to daylight in post. (camera set on 5500k)

This is a scene simple I built for my F5 vs FS100 test, which has tones, shadows, nasty bit of metal to have sketchy highlights and of course some narrow resolution lines at many angles. It happens to show XAVC going very miserable in the shadows.

https://dl.dropboxusercontent.com/u/33167160/ouch.jpg
 
Last edited:
I did a short Internal/RawExternal comparision of the FS5 on an Atomos Shogun and have some images:

http://alexandermejia.net/shooting-raw-on-sony-fs5/

The biggest things to know about shooting externally is that a lot of internal processing the camera does DOES NOT go over the wire:
  • No Temporal noise reduction
  • No Lens corrections
I'm also seeing a difference between how Atomos bakes in Slog-3 vs how Sony bakes it in. I wasn't able to test on multiple NLE's just to see if it was some QuickTime container wonkyness.
 
Last edited:
morgan_moore said:
A simple test being shooting a scene under tungsten and then correcting it to daylight in post. (camera set on 5500k)
Click to expand...

This is not a testing scenario that I would personally care about. I know you are just using it as an example, but nevertheless, I'm not going to go around shooting RAW just in case I might muck up my in-camera settings or exposure. I'm not saying I'm perfect, but it's not just something I feel I need insurance against at that price. What matters more to me is if RAW makes a difference when I hit the mark nearly perfectly the other 99% of the time.

Also, I never make any image judgements from still frames, especially when noise is of particular interest.
 
Demistate said:
I did a short Internal/RawExternal comparision of the FS5 on an Atomos Shogun and have some images:
Click to expand...

Well, done. I like it.
I would also add that about 1/3 of the shots I did for my testing ended up with mis-matched timecode on the Odyssey. Not just a few frames off, but hours off. This was a major pain in the ass to get the split screen shots lineed up. And if I had been hoping to use the XAVC clips as proxies in a real edit, it would have failed miserably. I didn't mention this before because I'm not 100% it wasn't my fault somehow, but it did happen about 30% of the time so I don't think so. And even if I did mess it up myself, something that touchy is dangerous in the field.
 
Doug Jensen said:
This is not a testing scenario that I would personally care about. I know you are just using it as an example, but nevertheless, I'm not going to go around shooting RAW just in case I might muck up my in-camera settings or exposure.
Click to expand...

It is not about mucking up - I trust you not to do that :) - It is about making ambitious shots that would not be possible with a thinner codec.. for example the tracking shot that leaves the flouro kitchen and arrives at the tungsten restaurant. Shots that could take a huge amount of light/gel to make work in a 'traditional' manner.

To me (when doing my own DIT/Post!) I would only use raw very selectively.. only on shots that I knew would break a traditional codec.

S
 
Hey Doug,
Thanks for the test. It's amazing how great the image is straight from the camera!
I don't think I'll stop using RAW to CDNG anytime soon. Even with the compressed delivery of test, I see meaningful improvements in the RAW image.
Is it worth the extra time, hard drive space, and hassle of running an external recorder? For me it is. But if I had to knock out a long project, in minimal time, for web delivery, internal recording could make it a lot faster and easier.
YMMV. Thanks again!
 
morgan_moore said:
To me (when doing my own DIT/Post!) I would only use raw very selectively.. only on shots that I knew would break a traditional codec.

S
Click to expand...

If you're using RAW very selectively wouldn't a days hire of a RED be the order of the day instead of crippled expensive RAW from the FS7?

I'm hoping Doug's test can put a lot of the arguments to bed.

@Doug, Below is an cropped example of XAVC-I keeping the detail in the trees blowing the the wind while heavily compressing the sky. Would it be possible for you to replicate a similar scene while recording ProRes externally? This might be one area that an external ProRes recorder is better? The compression artefacts are not normally visible unless a contrast adjustment is made. These artefacts do limit how far a grade can go before the sky becomes noisy with mechanical looking noise. In the example below there is an extreme amount of contrast pulled to clearly show the compression artefacts.

The crop comes from a clip that was 75% trees moving in the wind with 25% skies and clouds. XAVC-I is impressive at keeping detail where you need it.

Cheers and much appreciated you taking the time to do the side by sides, very informative.

2yuaosm.png
 
Nirv said:
If you're using RAW very selectively wouldn't a days hire of a RED be the order of the day instead of crippled expensive RAW from the FS7
Click to expand...

Im just commenting on why Doug should do tests that are of challenging scenes because to me raw should ​shine under pressure.

Im not making any comment on actual shooting practice or camera choice just test methodology.

Say I owned a raw and proRes camera (I do!) then I might swap on a scene by scene basis.

Similarly when shooting HD delivery project on the FS7 I might choose to shoot internal 4KforHD in certain scenes, like greenscreen, or when I will need to correct or stablize even if the majority of the project is HDforHD.

I need to buy hard drive pair 35 today, I dont enjoy burning data unless there is a need.

S
 
Last edited:
morgan_moore said:
Im just commenting on why Doug should do tests that are of challenging scenes because to me raw should ​shine under pressure.
Click to expand...

I was just making the point that there are other options than ploughing a considerable amount of money into a substandard RAW experience on the FS7 especially if one needs better quality on an infrequent basis.

If I had a 4k green screen shoot I wouldn't hesitate to hire in a RED or F5 and RAW etc. An HD green screen shoot is a different matter as the downsize cover a multitude of sins quite nicely.

I'm more interested in what Doug can determine with ProRes and if it is better in challenging situations and can overcome the 'skies' problem above I might buy an external recorder for those slightly more important shoots and maybe even the 4K green screen scenario. I might just have to hire some kit myself and give it a run through one can't expect someone else's tests to cover every situation.

I despise data and only want to capture the bare minimum, shooting CDNG is just bonkers IMHO, RED is the only RAW I'd consider as a regular option.
 
You must log in or register to reply here.
Back
Top