Comparing ProRes RAW and Internal Recording on the Sony FX30
ฝัง
- เผยแพร่เมื่อ 2 พ.ย. 2023
- In this video I take the ProRes RAW from the Sony FX30 and compare it to the internal 10-bit footage in camera. I do all of this in Davinci as with Sony FX cameras if you feed Davinci an MXF file, it gives you RAW like features for 10-bit footage. I wanted to see what the difference is using both codecs, since I am in DaVinci Resolve, I had to transcode the ProRes RAW into CDNG using the program ASSIMILATE. This was in order to keep the 12 bit files as true raw and not a compressed format.
Tech discussed in today's video ( links are affiliated and help with the content creation for this channel):
Atomos Ninja V:
amzn.to/3MrwUHW
Tamron 35-150mm Sony E:
amzn.to/46vJSMY
SmallRig Full Camera Cage for Sony Alpha 7R V/Alpha 7 IV / A7R IV/Alpha 7 S III/Alpha 1-3667B:
amzn.to/3s0e2ZA
SmallRig Mounting Plate for Attaching FX30 / FX3 XLR Handle to Camera Cage 2999/3241 for Sony A7 SIII, Camera Cage 3667/3639 / 3594 for Sony A7 IV (A7M4) - MD4019:
amzn.to/3QpWhwi
18-50mm F2.8 DC DN Contemporary for Sony E:
amzn.to/46LI9m6
Sony Cinema Line FX30 Super 35 Camera:
amzn.to/3ClKl7t
K&F 82mm polarizer & VND filter:
www.amazon.com/dp/B0BHSNR69R?...
Sony Alpha 6700:
amzn.to/3Qphqaf
Sigma 18-50 2.8:
amzn.to/3ND6eUp
Sony 11mm 1.8:
amzn.to/3ODYzHh
Sigma 16mm f/1.4 DC DN Contemporary Lens for Sony E:
amzn.to/3qdrFnj
Sigma 18-35 1.8:
amzn.to/43fYb5N
Sony A6700 Video Facebook Group:
/ 948121776270462
Sony 15mm 1.4 E Mount:
amzn.to/3WI34m5
Zhiyun Crane M2s:
amzn.to/3NQ7PWB
Zhiyun m2s/m3 quick release plate:
amzn.to/3Wmv4Ls
Sigma MC-11 Mount Converter/Lens Adapter:
amzn.to/3gX7vsG
PD Cable I use:
amzn.to/3Xx7sDt
Gear used for this channel:
Canon EOS R7:
amzn.to/3bsmWGM
RF Cameras for Video Facebook Group: groups/15510...
Canon RF35mm F1.8 IS Macro STM Lens:
amzn.to/3Gt25xq
Canon 16mm 2.8 RF:
amzn.to/3tPNpmU
SmallRig RED Komodo Cage:
amzn.to/3thFayR
Core SWX RED Komodo V-mount battery plate:
amzn.to/3qaGggW
SMALLRIG Right Side Rosette Wooden Handle:
amzn.to/35RmImW
SmallRig Super Mini Top Handle:
amzn.to/3inwfXP
SMALLRIG DSLR Camera Side Handle Bracket with NATO Clamp:
amzn.to/38SVYnG
SMALLRIG Field Monitor Holder Mount with Quick Release NATO Clamp:
amzn.to/39GOUd3
Blackmagic Design Pocket Cinema Camera 4K:
amzn.to/3nda3Uj
TOKINA ATX-i 11-16mm F2.8 Canon EF:
amzn.to/30y2grv
BLACKMAGIC DESIGN Blackmagic Pocket Cinema Camera 6K:
amzn.to/3iJO0Sl
Sony ZV1:
amzn.to/3qtTaTu
Vaxis Atom 500 SDI:
amzn.to/3ngcOBF
SMALLRIG BMPCC 4K & 6K Cage:
amzn.to/352rDSc
Neewer 60W LED Video Light:
amzn.to/3dyFaTZ
#sonyfx30 #fx30 #a6700
I think most average folks need to stop thinking about 8, 10, and 12 bit and focus in on the basics. I caught myself almost going down this route only to realize it starts with the basics. You can get phenomenal imagery with proper lighting and composition. This is especially important if you're getting paid as your job. Clients could care less of what you used to get it, it's the final product at the end. If you're rushing about, that's not being a pro.
100%. There are many "bells and whistles" the average person obsesses about and gets confused about. It's like 32 bit audio recording... you need to understand when and why it might be useful, and what the limitations are in relation to the gear you're using. It all becomes just marketing terms. You really need to understand the fundamentals and make good decisions based on what you're doing, who it's for, where it will end up etc. Bigger and better isn't always the way to go, and often it creates more headache than it's worth.
For me, it's not about the visual differences between internal recordings and external raw recordings that matter. It's the flexibility in post that I look to when considering recording in a raw format. I think a better comparison would be how much flexibility in POST do you have between the two formats. Show us how you can work with each format in Resolve.
The obvious differences between raw and internal, is that with raw you can change the white balance and ISO in post without a loss in image quality. Usually raw brings more noise to the footage, but that can be relatively easily resolved in post as well. Resolve's noise reduction tools are quite powerful and rather easy to use.
With raw you often have to deal with much larger file sizes. So, having a powerful and fast computer would go a long way in working with raw in post. I think it would be useful to show us how you would work with raw and internal recordings in post. Let us see the challenges we would need to consider when deciding if we will shoot internal 10-bit or external 12-bit raw recordings. I would love to see that.
You are totally right, but if you haven’t caught my Sony RAW video, you can change the iso and white balance in the raw tab to FX30 footage if you transcode the files to MXF like the fx6 and fx9 output. That’s why I reference the video, because aside from latitude which the raw footage obviously has more of, a lot of people just want the ability to change white balance and iso in post which you can. The key is I wouldn’t transcode the fx30 footage unless something is messed up and you NEED to Change it as the transcoding process is yet another step.
All good points. But I will add that shooting in Pro Res will actually be easier for your editing system to work with since it doesn't have to deal with all the extra compression. In fact Sony Cameras even show you a warning if you attempt to shoot in the H.265 codec which reminds you that you'll essentially need a newer gen Mac (M1,2,3) to playback the footage.
Nice comparison! Really great info :) 🙌
Thank you!!
Thanks a lot, quite important info!
Such a great video, Ping loves it!!!!
Thanks Ping!
Thank you Geoff. Can't wait for the Speedbooster test.
You and me both! I finally got auto focus to work on the speedbooster which may me another video in itself haha, but plan to do something this week
Dude, thanks so much, total game changer.
You got it!!
Thank you for doing such an in-depth comparison. Please let us know if you create a video regarding the FX 30 with the speed booster 🙏🏽🙏🏽🤙🏽🤙🏽
Thanks excellent video
RAW sensor data much less noise reduction added to it's file. The camera's internal recording has TWO noise reduction processes added. 1.) Noise reduction on the image processor and sent to HDMI out. 2.) Secondary noise reduction built into H.264 & H.265 encoders. (Yes, HDMI is only hit with one noise reduction process. XAVC has two and is hit the hardest)
Yes this is all true, but everyone commenting this forgets one huge part. The internal footage uses the entire sensor of super 35, while the raw footage is closer to micro four thirds (a bit smaller) that is an even bigger factor to the noise than all the other factors you and others have mentioned. I am not discounting that no denoising is done but if you are going to post this info you MUST take the actual sensor area and iso used into account. Especially because this info you present is true for all cameras that film in raw, but most are not giving you an additional crop than the compressed footage gives you
@@ReachFilms - Yeah, the cropped read window was a big gimp trick on this camera. Sony didnt want to give the full sensor width in raw for this price point. That sucks. And for that reason, I'll only shoot raw on my A7S-III and FX6. Pulling 4k video into ProRes might be a bit of a boost in image quality over the low bit rate XAVC-I (300mbps) internal codec. Plus, you wont be hit with secondary h.264\h.265 noise processing.
@@CLIFFLIX yeah when these recorders are clearly able to (at their cheapest price point) give 6k 30. They should have just given us 6k30 lol
thank you so much, this is a very useful video, maybe I will buy an atomos, do you know if the blackmagic recording is compatible with sony FX30? Braw is muuuuuuuch better than prores raw
Not that I know of, if it was I would have gotten a blackmagic recorder instead. I don’t know why Sony doesn’t make their cameras compatible
Isn't the color difference more a result of turning the linear RAW into BMD Film then Red, as opposed to the internal footage which went from S-Log to RED? It would have been nice to see a like for like comparison. But otherwise, great video! The BMD to RED transform looks nice. Also, it was interesting to see the result of the ISO changes in post. The internal codec has the benefit of internal NR as well. Could you get the RAW image to clean up using NR in Resolve?
One last nerdy comment. The crop is pretty significant, but I think it could have a very Super 16mm film look if you wanted to use it for a project that would benefit from that look. This combo--FX30 + Atomos--is kind of like the new BMD Pocket Cinema Camera OG
Hey Phillip, some of the change totally could have been from the bmd transform, but what was interesting is I tried to do the bmd transform to the 10 bit too and was getting the same difference in the image so I’m thinking it’s either due to the difference in color info.. or the transform in the actual raw tab it more accurate than the color transform plugins
I'm creating a cinematic travel video and currently using the Sony a7C. I'm considering purchasing either the Sony FX30 or the Sony ZV-E1. Could you please recommend which one would be the better choice for my needs?
It really depends, the biggest factors are lens selection, low light and if you need external raw.
What codec & bit rate do you have to shoot in to be able to convert the footage to raw?
You can be in HS or s-I in the highest bit rate up to 4k60
Thanks for the video. I'm looking into this camera and I'm not interested in RAW due to the crop, but I'm wondering if I can use the Ninja to get ProRes 422HQ files without the RAW crop. I understand this would give me smaller files than the internal S-I. Is that possible?
yeah of course you can, as long as your not switching to prores raw you will be able to capture the entire sensor data
On the a7siii the external is a larger image circle that internal doesn't offer. More res too at 4.2K. I thought it was similar with the fx30 but 4.7K vs 4 2K.
@@errol3184 right
Did the Raw have better highlights? I’m a noob. Or was it in the grade? Not sure, but initially I liked the dynamic range of those highlights on his forehead being dropped down a little.
Yes the raw did have a bit more dynamic range on both highlights and shadows
What does 10 bit with RAW 4.7 crop?
How do you get 4k 30p on the ninja V? Mine won’t go past 4k 24p for ProRes Raw!
You have to change your raw output to 30p
I noticed that the highlights are better in the 12bit footage. Seems like 1/4 fstop darker too
It’s quite possible!!
I believe the majority of color difference you're seeing is a contrast difference between transforms, not any actual benefit of 12-bit RAW vs 10-bit YUV. If you were to normalize the contrast, I think you'd see a basically indistringuishable color response.
It’s possible, but I had to grade the 10 bit footage a bit harder to match, the saturation and contrast were pretty off compared to the raw using the same basic grades, which I corrected to make the images a bit more fair as yeah 10 bit footage is still really powerful
I totally agree, all this RAW bs… it’s getting a bit too much.
I think that filming in RAW you get the RAW colors caught by the sensor in 12-bit. I use HLG3 which has way more color with Rec.2020 color space and Rec.2100 color gamut before pushing it to Rec.709. I think you can get more colors if you use DaVinci Resolve then transfer the very wide Sgamut3.Cine or SGamut3 color gamut to DV Wide Gamut, color grade there, then transfer it to Rec.709. HLG3 seems to recover shadows pretty cleanly to me maybe due the lower base ISO's or the HDR Metadata inside the files. SLog3 is great but if you ruin exposure then you really have a very small wiggle room in comparison.
Very interesting insights about HLG. I'm only about to get into Sony systems so I don't know it deeper. Am I right that you advise to go deeper into HLG format and that it brings often more value than Slog3?
@@harry54ck For run and gun, wildlife, easy edit, less risk of noise, very good dynamic range (slightly less than SLog3), and HDR-ready. I can re-expose the scene perfectly in DaVinci resolve using the metadata, but I usually overexpose by at least 0.7+ up to 2.0 depending on the situation. I also tend to use 2nd base ISO of 2,000 (for HLG3) with VND if needed and at least a circular polarizer as it renders pleasing colors.
If you are making a film with 100% full control of lighting and crew then SLog3 is great with plenty of LUT's, but I would rather make LUT's/looks using Dehancer Pro instead if that's my goal or manually.
Play around with them and see what suits you best. I see most people recommend either SLog3 or S-Cinetone, which are the absolute extremes of each sides. HLG3 is very close to SLog3 and very easy to edit with much pleasing colors.
Why is there no p30 in XAVC HS/H.265? Do I have to use external recorder? I'm not happy about this.
Do you have a BRaw comparison? I was thinking about used bmpcc4k... But it's bulky... Although very nice price.
Unfortunately not, I don’t have any blackmagic cameras currently to test. When my red comes back from service I may do that comparison though with R3D
@@ReachFilms Thank You. Seems like i'm going to have cinema time, watching so many info videos. Thanks!
Qué buena imagen tiene en ambas situaciones.
thank you!
I edit with Davinci Resolve. If I record raw with the atomos am I able to transcode the files into another raw codec that is compatible with resolve using sony catalyst browse?
There’s a program called assimilate raw that you would need to transcode the proresyraw to CDNG which you could edit in davinci
@@ReachFilms thanks!
does FX3 have dual gain when shooting external raw at 12800?
they way dual iso works, yes it should, however i haven't personally tested fx3 raw so am not quite sure how its affected
its cleaner on 10bit because of the internal noise reduction. you HAVE to use NR on raw footage.
Yep that’s totally true, but 2 things to note, one, the sensor area for raw is around an MFT sensor size vs the full super 35 sensor, so that being said there will be more noise in the raw just due to the difference in size. The second is the more NR you use the less sharp the image becomes, Sony does some NR magic in camera, so it was unfair to clean up the noise in this comparison as it would make the raw image look much worse to get similar results.
I wonder what im doing wrong here because I bought the studio version and my transcoded files from my FX30 still come in as regular clips with the .MXF files and the RAW tab is still completely greyed out
Are you filming at the highest bitrate in 4k HS or S-I ?
@@ReachFilms indeed I am. XAVC S-I 4k in Cine EI then transcoding to Intra in catalyst browse to the .MXF thats outputed, drag it into Davinci and nothing. Some have claimed apparently Davinci may have gotten rid of this option because people in 18.6 are claiming the same exact thing
@@Kaotix_music I’ll double check but my guess is there is something missing in the chain along the way. The only frame rate I couldn’t get to work is stuff in the S&Q modes
@@Kaotix_music also just to confirm you have the paid version of davinci right?
@@ReachFilmsindeed I do hahaha I actually finally pulled the trigger on buying it for this very reason
Thanks for the comparison!
I agree about a difference in noise because of the crop to a significantly smaller sensor size.
But you cannot compare RAW footage without _any_ noise reduction with footage that has been processed in the camera (which always includes noise reduction). RAW noise reduction is always done in post, the advantage being that other grading steps happen _before_ noise reduction destroys image information and introduces artefacts in steps that follow.
To be fair you should compare the cropped RAW take _with_ noise reduction in Davinci Resolve vs. an internally processed take where the lens is more zoomed in (to compensate for the different crop factors).
Problem is with the fx30, unlike the fx3 the crop is so bad most people arent even going to use raw on the fx30 because you have to jump through too many hoops to not even use the full sensor or get 6k
6:55 Sony has noise reduction that you cant turn off when recording internally. That's why 10 bit has less noise. RAW is RAW, without any NR or sharpening.
You should have also included ProRes 422HQ in the comparison (with the ninja V), I believe it comes out cleaner than internal XVAC S...
If you film in ProRes, you lose the metadata that the camera embeds in the file to use the raw tab in Davinci resolve. Please see my 10-bit raw video for an explanation
I understand that, I meant ProRes VS internal camera codec. You don't make any use of the raw tab for both. I've seen comparisons where ProRes had an edge over internal 10 bit in grading.@@ReachFilms
@@furiousangus6467 the whole iso test is making use of the RAW tab in davinci, it’s not a straight up grain comparison, it’s how do the images look when boosting the ISO in the raw tab of Davinci resolve, as you would think the 10-bit file would get really messed up but somehow there is some magic where it looks great if within a certain amount of stops
Cool test! 👍 I’m assuming you edit on Mac...
Thanks! Yes I do, but this was all in Davinci resolve which is available for windows too
Yes but DR doesn’t support ProRez on Windows...
@@dpixvid so I’m actually converting the ProRes raw into CDNG files with assimilate, because you can’t use ProRes raw with davinci anyways
@@ReachFilms gotcha - didn’t catch you’re converting to CDG thought PR on Mac... My bad 👍
Sony 6700 can film RAW format?
Unfortunately not currently
More noise and crop... slightly different colors that can be adjusted... for all the people arguing you can remove the noise,... its an extra step and you should compare them equally like internal + noise reduction form Davinci for exsample instead only comparing the raw with the add noise reduction from davinci lol... gotta make it fair and squre
exactly, i actually could care less about the noise from raw as most of the people here pointed out raw footage tends to have, but i think a lot of people posting that stuff are unaware how severe the crop is in the fx30 raw and how that makes the noise much worse than just filming in raw itself
Does fx30 has soft skin effect ? Please let me know
In the manual modes I use it does not, not sure about the auto modes but those features are normally on cameras aimed at content creators like the zv series
The Nikon that Z6 3 can do ProRes RAW internally now. (It's shipping shortly). Can anyone match this and do a full frame with IBIS?
Josh Sattin N-raw vs h.265 I would check out his channel
That's more like a MFT or 1" sensor size image after the 4.7 crop though.
Yeah for sure it’s around MFT, don’t shame the poor fx30 with the one inch size haha 😂
4.7K from 6K S35 ends up being much smaller than S35, somewhere in-between 1" to Micro 4/3. @@ReachFilms
Ummm, the internal recording has in camera Sony noise reduction.. the external output does not. Been a while since I saw that youtube video, or else I would point you to it.
absolutely, there's no arguing there, but the noise is even more than no noise reduction, there is an additional crop when using raw vs using the internal sensor using the full frame. When you use noise reduction on the raw you start to lose sharpness so it wasn't fair as sony works magic with their internal noise reduction
Timestamps please
RAW will always have more noise because it’s not doing internal noise reduction - doing your own NR in post on RAW footage will produce much better results than h.264 codecs with internal NR. RAW is preferred for low light for this reason
look at the other comments, this video isnt just about noise from raw video specifically. Literally every person who is making this comment keeps forgetting the massive difference in the sensor size using the RAW video mode vs the normal 10-bit mode on the FX30. That crop using less of the sensor is an even bigger reason for noise over it being raw video, especially the higher you go in ISO.
I think the internal 10-bit looks way better.
in some situations the full sensor 10 bit is definitely better
Useless, I shot an entire series in 10bit. Looks amazing.
I don't think any conclusion said the 10-bit footage was bad. But the fact we CAN get 12-bit RAW from this camera is pretty impressive and useful for higher end production
horrible music choice
Thank you!!!!