Really depends on your client. I'm a sports videographer. I film for first grade to Varsity. There is no one looking and noticing that it's not filmed in raw or S log.. the majority of people that would notice it is us the content creator. I shoot with a Sony a6400 and a Sony A7 Sii. Last year I decided to upgrade my lens to a Sony G Master f2.8 70-200 instead of a new camera. Told myself I would upgrade to a better camera this year but decided to go ahead and buy another a 6400 with the Sony G master f1.4 35mm. I shoot 1080 hundred megabits 120p and 4k. Now I am making more money than my 9-5. It comes down to camera skills lighting and editing if you can Master those you can make money.
My finding on cameras within the past 4yrs is that when they have 10-bit LOG codecs with good bitrate, you really don’t need raw. The sensors are all running their ADC at 12-bit, and the LOG curve does an excellent job of holding that 12-bit sensor data within the 10-bit container without losing any info. Raw can look DIFFERENT, or have different color science, but on a pure technical level of data loss, there’s almost zero difference. If you always find yourself custom white balancing, maybe raw has a small advantage there, but I always treat my cameras like a film stock, and stick to either 3200 or 5600K.
Great video. Idk if my experience is helpful, but as a VFX artist, prores raw and footage without denoising in general is better for what I do. Integrating VFX and CG into footage is much easier when you can properly conform footage into aces simply because of the fact that there's more detail to work with. The in camera denoising is just low quality, there's posterization artifacts and as another commenter mentioned, there is also temporal denoising at higher isos. For most projects shooting raw on the fx3 isn't necessary, especially when there's no or minimal vfx, but on projects where there's time, budget and processing power to work with it, i think it should be used.
The cool thing is though with something like an fx6 you can turn off the internal denoising on the camera while still not using RAW which is nice. I think the FX3 is supposed to allow that too?
Exactly. I've done some light compositing with the internal codecs, and with some 12,800 ISO shots, the blobby noise reduction artifacts don't always look great with chroma keying and working with alpha channels and blending modes. I'd rather have a more uniform, sharper grain in that situation, or bring in a third party plug-in to deal with the grain. Also, that added little bit of sharpness does help with chroma keying when refining edges. Not a huge advantage, but every little bit helps.
I’m a VFX artist myself, and we’ve done all sorts of VFX work with XAVC live action plates, without any issues. I’ve never personally seen any notable posterisation artefacts or IQ issues.
Actually, regarding noise reduction, internal 12,800 iso does some bad temporal noise reduction that has ghosting and trails, at least on the A7S3. You can check this by recording at 12,800 with a high shutter speed to avoid motion blur in a dark scene and panning around. You'll see these artifacts, especially in dark areas. This makes it impossible to stabilize shots at second base iso with catalyst or motionflow. My solution is hooking up the ninja v in raw mode. It disables temporal noise reduction for the internal recording as well, even if the ninja is not recording. :)
Happy someone mentioned this. Really makes things difficult as a VFX artist. Have worked on projects shot with fx3 for B-cam and the temporal noise reduction creates all kinds of problems. Really wish Sony would release a firmware update that gives people the option of turning that off because using an external recorder makes the fx3 useless for helmet pov cams and things like that.
In addition to flexibility and control over NR, one quirky advantage is that ProRes RAW is often faster to work with than the internal Sony codecs, avoiding the need for a proxy workflow for older PCs.
I’m pretty sure the shift in color and contrast can be explained by legal vs full. For some reason I’ve noticed that the Sony cameras output legal and the atomos records full. I bet if you were to use a legal to full conversion lut that they’d match up almost perfectly.
You are correct in assuming it's full vs video But it's not a Sony issue In my testing prores is recorded incorrectly on Canon and Panasonic as well When recording prores the recorder doesn't know whether to save it as full or video, as far as I'm aware there is no metadata that tells the recorder to record it one way or the other, so you would have to actively enable the feature when shooting in log or know to interpret it correctly later Checking the Legalize option when shooting log can help, the issue is when recording on a device that doesn't have this option, or when working with older footage not recorded with the legalize feature Luckily you can just interpret the footage as full in your NLE
I really wish atomos would just come out with a press statement about this Watching all the misinformation fly around or widespread people incorrectly using the devices is insane
I appreciate these tests thanks Josh. Every time I have decided I'm going to shoot raw I end up defaulting to an internal codec on almost every camera I've ever used. Most notably my C70 which obviously can do 12 bit raw internally at a more manageable file size than ProRes raw, but I never use it for paid work, only ever a couple of passion projects. It's kind of ironic but to my knowledge at least, most people who are shooting raw with more affordable camera offerings are doing it for the fun of it, not because it provides any tangible benefits to their professional lives. I bet I still end up messing around with raw again though, ha. That's life.
No need Atomos, no need prores raw HQ. I did many tests with and without the Atomos and I can say, everything is on the grading; The color science of the FX3 is fantastic! It's not the Atomos will bring more in the image, except the sharpest like you said because it's a 12-bits. For the color balance, can be adjust 100% same. Good work here! Thanks
Gerald Undone did some testing and came to the same conclusion that you cannot get any more dynamic range by shooting raw vs 10bit SLOG3 at high bitrate. As for colours, try XAVC HS which has slightly different/punchier colours with a slight shift to green and better skin tones.
Much appreciated and comprehensive. I think i must have missed the point that prores raw locks you out of davinci. Apologies if i did. For social media work no one will notice the extra coloyr but they will notice the lack of sharpening and NR thta they see all the time from so t cameras on youtube posts. There is a reason why they are so popular for youtube. And the turnaround isn’t usually justified for raw if it isnt blackmagic and resolve native. For professional work and especially multi camera or multi source wb and the ability to nail colour to spec is often more important especially for colour sensitive client products. Also pushing the image in terms of grade and look isnt something most clients need but the use of film emulation’s definitely benefit from a higher bit depth and raw. With these sort of straightforward images the differences arent really being tested. When you shoot mixed source lights with multi cameras typically an a cam like an fx9 or fx6 then having a raw image to match and adjust is very useful. That said for most youtuber social media users it’s important not to underestimate how much the internal camera codec is doing to help out. NR sharpening and lens distortion are very good examples that people shooting sony especially underestimate. Again no surprises that the creator shot in raw but thats an 80 million dollar feature and not a 2 minute social media post. Horses for courses obviously. Thx for the continued excellent and thoughtful posts.
You're welcome. Yeah that was interesting for me to see through doing these tests how much the camera is doing when recording internally. Thanks for watching!
The difference in detail and color depth to me is 100% worth it for anything important, not to mention lack of compression artifacts - especially in chroma channels in All-I recordings. To compensate for the noise and lack of ISO control in RAW exposure, I recommend creating a couple additional LUTs that will pull the image 1 to 2 stops. I typically will monitor with such a LUT in either base ISO and feed the sensor a stop more of light, which kills the low end noise (or two stops for the 12,800 base, making it effectively ISO 3200). You just have to watch highlights and when in a contrasty environment (lots of windows for example) then just expose with your standard LUT.
I get exposure control in stops with premier. It's the same as iso and works the same as using the iso in raw tab in resolve. Ironically, resolve exposure doesn't feel like it works the same and it's more like a regular exposure tool. That's after transcoding to DNG. Catalyst Browse will let you change iso too in xavc. When exposure is perfect nothing beats DNG in Resolve in my opinion.
Thanks for doing this Josh! That raw reminds me of the incredible beauty that comes out of the 5Diii with Magic Lantern on it so it can pump out raw. It's amazing, but such a work flow, and no modern advantages like good autofocus and whatnot. Sort of makes me want to get an FX3 at some point...
IDK what you're using to convert ProRes RAW, but for anyone curious in my experience the free converter tool that's on several app stores to convert to cDNG does not set metadata properly for colorspace conversions, as well as failing to include the correct 3x3 matrix metadata for white balancing. This results in slight hue shifts depending on the camera once the cDNG file is opened in Resolve. The tools that convert ProRes RAW correctly are Adobe (such as Media Encoder) and Assimilate (Play Pro Studio). After that, the cDNG or regular ProRes files you create will work and look better in Resolve.
This comparison is what i was looking to see and everybody says you have more color information to work with on the prores raw. But nobody actually shows how far you can push the files till they fall apart in terms of color. That is actually what i want to know. Can you make an episode about heavy color correction comparing both file types? You only showed sooc files without any color adjustment and no film luts applied. I m having an anamorphic rig and because of many lens elements, the colors are way off on the chart. So i have to bring them back and the internal footage falls apart before applying any film lut. Will the prores raw save me ? And how much will help?
You can definitely push further with ProRes RAW. The starting point isn't as good sometimes though. I noticed I could stabilize better in post than using ibis, active mode, lens stab and catalyst Browse combined, then hitting stab. Color detail is on another level but less accurate to start.
Thank you for the NERD info! I hate when people just talk and talk and don't SHOW. Earned another subscriber! PS. I am a Fuji still shooter, do you think the FX3 is still worth it today, at almost the end of 2024...?
You're welcome! I appreciate the support. FX3 is a great camera and will be so for many years. You can see my long term review here: ruclips.net/video/L2EDzbCUeYo/видео.html
My understanding of ProResRAW on Sony is that it doesn't increase dynamic range, you're still limited to whatever the camera offers in SLog3. So in that sense most of the differences are going to be in color, and then the difference in how colors work on a Log curve vs. raw sensor data. One of the advantages to me would be that RAW bypasses all of the cameras lens corrections and noise reduction, in that sense it offers a little more freedom in post processing allowing you to precisely control that. Overall, for me there isn't enough of an advantage to RAW over the internal codecs. It is also an absolute pain to try and work with on Windows, I've given up trying to use Premiere with the RAW plugin, it's too unstable. So I convert to CinemaDNG for Davinci Resolve which is not ideal, however I haven't seen any degredation when doing so, but CDNG has some odd color space and gamut choices that would not be present if working natively on OSX with Final Cut.
Note that saturation and perceived "sharpness" are both factors of contrast. If you adjust the internal recording to match contrast, the saturation and sharpness would also match.
Yep. Adding contrast adds saturation and makes the image look sharper. I find it likely that matching levels would largely mitigate the difference. To me it looks like there's something fishy with the conversion. There should be no color difference whatsoever, if the conversion is done correctly. Or, it could be that the internal recordin's picture profile settings are not at "default" values, at least not the same "default" as the NLE assumes. Also, when people claim that they can get more info on RAW than on the internal log files, sure that can be true, technically speaking. Is it really true in practice, in real world color correction cases is a whole another question. Using 4 x the data rate to save pristine chroma noise is not my idea of smart.
Which reminds me, i think i'm seeing some debayering artifacts with the raw - the black plastic texture in the chart seems to have some discoloration, red and blue static noise, that's different from the noise that changes from frame to frame... quite visible in the 200% zoom.
I always make my videos as ojective and transparent as possible. Yes I know you can change the perceived sharpness of an image with more contrast, but in this case the Prores Raw is slightly sharper than the internal codec. This is true even if I adjust the contrast to match. It might not be perceivable through RUclips and that is challenging but I stare at this on my 32" 4K monitor before making these conclusions. There is nothing fishy going in with the conversions. I literally showed you what I used in this video. I am also editing in Final Cut Pro and this is the native editor for Prores. Please feel free to replicate the tests and post your results for everyone to see. This is what I saw with the gear that I have. I hope this was helpful for people interested in the differences between these two recording methods.
@@Josh_Sattin Hi Josh - i do not meran "fishy" as in you doing something wrong, and i am sure you saw what you saw. I mean "fishy" as in the conversion is the very likely culprit for the differences, rather than the file formats themselves. The software we use is rarely perfect, and the cameras may not be doing the conversions exactly equal to what was specced by the manufacturer for the software vendors in practice. There are quite a lot of additional settings in Sony cameras' color profiles, which may differ from the defaults used by editing software. There definitely are also differences in how various software handle footage - starting from whether you see the actual full range in the file, or just the legal portion or Rec709. One can't really completely trust the editing software to work correctly. I do not own FX3, but I would actually be quite interested in comparing the original files from the camera myself, if you are willing to share a take or to, and share my results with you (and the rest of your audience, if you allow me to).
I know you don't have a FX3 but here is some additional information to help clarify. When shooting in Cine EI in the FX3 you can't make any adjustments to how the camera records the Slog3 information like you could adjust Slog3 with the older system of adjusting picture profiles. It is what it is. There aren't any of those full or legalize settings in the Ninja V when recording ProRes Raw either because you are recording raw. Using FCP and the built in decoding and LUTs is the most objective way I can observe and present the differences. Also FCP is a NLE that can handle ProRes Raw natively and it converts the raw footage into Slog, which I show in the video. So the differences you are seeing are the differences that we get.
I shot my FX3 w Ninja and the external files recorded at 1920 and not 4K. I can't find how to switch resolution on ninja? I just update ninja to OS11. thanks with any help Cheers
Your client will never notice the higher quality, but why would any one wants to deal with the raw workflow? Raw workflow results in more cost (storage ) At the the end, you've getting paid the same per your contract agreement.
Comparing Prores Raw 12bit to All Intra 10bit is like comparing 12bit compressed RAW vs 10bit HEIC Image, you won't able to see much difference if you only grading it slightly
Hello, thank you for video. I have a question. What is faster for export on MBP late 2023 16” M3 pro 36gb. 1 hour video ProRes or XAVC S-I? I like 12bit colors and option. Thank you
Great tests! The luma and color density differences make me wonder if FCP is properly showing the raw and S-Log3 levels. Did you happen to try doing a Full to Video levels LUT to see if that makes the luma and saturation levels of two line up better? I wish FCP had an option to choose to interpret via full and video levels.
Thanks! I used the standard conversion LUT on both clips for the test charts. For the outdoor examples I copied the grade from the internal to the Prores Raw. I explained this in the video. The differences in both sets of shots were consistent.
@@Josh_Sattin I believe FCP assumes the file's levels are Legal, unless they have metadata to assume Data levels. Internal S-Log3 is recorded in Data levels, while ProRes (and raw?) are using Legal levels.
About contrast - it can be issue on atomos. Because on my 7iv I have the same thing in xavs-i and prores hq. Prores hq have more contrast then internal.
Well there is no other way to record Prores Raw so this is what we get in Prores Raw. I did not test the difference between Prores Raw and Prores Raw HQ.
Hey! Nice comparison! I have a question, did you use an ND filter in front of the camera? Probably The magenta cast would produce that render in outdoors
When you record in raw, it only records one way, in raw. You then convert it in your editor. I explain it all in this video: ruclips.net/video/0ZqJ46L7-74/видео.html
@1:15 there is nothing called half a frame. 1 frame is the least . And moving one of the clips by 1 frame should solve the issue. For color grading . If you only adjust the contrast and saturation without using the official Lut you will lose dynamic rang. Gerald undone made a video about it and showed that this method is not good
I am using the camera recording to trigger recording in the Atomos. The Atomos recording does not start at the same time as the internal recording. There is a delay. When I try to line up the clips in post the closest I can get them is a half frame apart. It's just the delay that is built into this configuration lands it in between two frames of the internal recording.
@@Josh_Sattin I see! It seems they have step away from letting them record more bits. I rember seeing this keynote from Jeromy back in 2018. ruclips.net/video/9f3G9mP9Tsk/видео.html. Dont know why they choose to cap it at 12-bits when all Sony cameras clearly are capabel of moore. Any how thanks agin for a fantastic video! Really love the way your cannel has been evolving!
I swear y'all and not you personally gotta just love editing 😅 I get that it's a video to make money off of, but all I can say is, the right tool, for the right job. If you're running a business figure out the camera that does all that you need in-camera. Get your damn time back. I know people love to hate on Nikon, but Nikon can do what prores raw does using Active D-lighting straight out of camera. You can add a custom curve if needed prior to shooting. I get where Raw is needed, but that just means you need better lighting skills which we all do and its a life long process to perfect. They dynamic range is never going to surprise you until you can surpass 13 stops otherwise it's just an exercise in redundancy.
Im gonna do some prores raw shooting this christmas on a new ninja v coupled w an fx3. Just wondering if i change iso in post, does that mean I’m deviating from native iso and therefore changing my dynamic range to the worse? Should i only change the ”exposure adjust” and stick to 800 and 12800 in the iso tab? Thank you
As an Adobe premiere user. I have no idea how to apply noise reduction. I don’t think there is anything directly in Adobe that allows you to do it to my understanding. Can anyone help me figure this out?
I do not know for certain, but i find it likely that the answer is "none". Perhaps in some extremely high ISO situations, where using noise reduction in post may give better results than in-camera.
Thank you. As I stated in the video, the main advantage is more control over the noise reduction and a heavier file (bitrate and bit depth) for more advanced color grading and manipulation.
I was wondering if you have a subscription to storyblock and publish to RUclips but after some time you decide to cancel it, does the previous published material needs to be removed?
“Royalty-Free” is a type of license that allows the unlimited use of content for any media projects without having to pay royalty fees. All footage, templates, and motion backgrounds are royalty-free and they are yours to keep forever - even if you don’t maintain your Storyblocks subscription. help.storyblocks.com/en/articles/3622237-what-does-royalty-free-mean
I just encountered something weird. I recorded internally with XAVC-S 4K and externally to NINJA + in 4K ProRes RAW on Sony FX3. I fed audio to the camera via Rode Wireless Pro. I'm currently trying to sync audio. The internal FX3 recording and the Rode Wireless Pro WAV file match perfectly. However, the Ninja V+ ProRes RAW file's audio does not match perfectly with either the Internal FX3 recording's audio, nor with the Rode Wireless Pro's audio. There is a slight drift. I even converted the MOV files into Cinema DNG and tested it in Resolve and I get the same exact issue I was getting in Premiere. In just once scene that is 3 minutes long, I can sync the audio at the beginning of the scene, and by the end of the scene, the sound is 4 full frames out of sync. I'm wondering if anyone else had had this issue?
And why on that clip of you in the shadows of the trees, is the pro res so much noiser? Like I've noticed on a lot of cameras that can record raw externally that it gets wildly noisy in the shadows
What would you like to know about HS? I recorded in S-I for this comparison because it's the highest quality that can be recorded internally in the FX3.
The way you describe prores raw being more contrasty and more saturated leads me to believe your NLE is not interpreting the data levels correctly Make sure the prores files is being interpreted as a full data level and not a video data level This is a common, industry wide issue where the NLE doesn't know to interpret LOG prores as a full signal, resulting in the whites blowing out, the blacks crushing and the colors being too boosted This has also been in issue for as long as I've had the a7s1, Sony claimed it was an atomos problem, atomos claimed it was a Sony problem Gerald undone got it wrong by saying it was a Sony problem (I 1000% believe that and would be willing to prove it) Atomos added the ability to record in a full data level "Legalize" which fixes the issue, this is what they should have done years ago and just shows it was always a problem on their end (to be fair, there is an inability to flag log footage when recorded externally to be saved as a full level) Note though that Legalize records the image in a full data level right out of camera, so switching a clip shot in full data level to full data level would give unintended results
Really depends on your client. I'm a sports videographer. I film for first grade to Varsity. There is no one looking and noticing that it's not filmed in raw or S log.. the majority of people that would notice it is us the content creator. I shoot with a Sony a6400 and a Sony A7 Sii. Last year I decided to upgrade my lens to a Sony G Master f2.8 70-200 instead of a new camera. Told myself I would upgrade to a better camera this year but decided to go ahead and buy another a 6400 with the Sony G master f1.4 35mm. I shoot 1080 hundred megabits 120p and 4k. Now I am making more money than my 9-5. It comes down to camera skills lighting and editing if you can Master those you can make money.
10000 %
Ping thinks you should make video
Here I am praying for next gen of Sony cameras all comes with X-OCN 16Bit codec.
Having the option for the better RAW quality is important for those that need it. Great comparisons.
My finding on cameras within the past 4yrs is that when they have 10-bit LOG codecs with good bitrate, you really don’t need raw. The sensors are all running their ADC at 12-bit, and the LOG curve does an excellent job of holding that 12-bit sensor data within the 10-bit container without losing any info. Raw can look DIFFERENT, or have different color science, but on a pure technical level of data loss, there’s almost zero difference. If you always find yourself custom white balancing, maybe raw has a small advantage there, but I always treat my cameras like a film stock, and stick to either 3200 or 5600K.
Great video. Idk if my experience is helpful, but as a VFX artist, prores raw and footage without denoising in general is better for what I do. Integrating VFX and CG into footage is much easier when you can properly conform footage into aces simply because of the fact that there's more detail to work with. The in camera denoising is just low quality, there's posterization artifacts and as another commenter mentioned, there is also temporal denoising at higher isos. For most projects shooting raw on the fx3 isn't necessary, especially when there's no or minimal vfx, but on projects where there's time, budget and processing power to work with it, i think it should be used.
Thanks! I appreciate you sharing your perspective on the uses for raw vs internal recording.
The cool thing is though with something like an fx6 you can turn off the internal denoising on the camera while still not using RAW which is nice. I think the FX3 is supposed to allow that too?
@@FinalGrade No, you can't turn off NR while recording internally in the FX3. It is one of the differences between the FX3 and FX6.
Exactly. I've done some light compositing with the internal codecs, and with some 12,800 ISO shots, the blobby noise reduction artifacts don't always look great with chroma keying and working with alpha channels and blending modes. I'd rather have a more uniform, sharper grain in that situation, or bring in a third party plug-in to deal with the grain. Also, that added little bit of sharpness does help with chroma keying when refining edges. Not a huge advantage, but every little bit helps.
I’m a VFX artist myself, and we’ve done all sorts of VFX work with XAVC live action plates, without any issues. I’ve never personally seen any notable posterisation artefacts or IQ issues.
Actually, regarding noise reduction, internal 12,800 iso does some bad temporal noise reduction that has ghosting and trails, at least on the A7S3. You can check this by recording at 12,800 with a high shutter speed to avoid motion blur in a dark scene and panning around. You'll see these artifacts, especially in dark areas. This makes it impossible to stabilize shots at second base iso with catalyst or motionflow. My solution is hooking up the ninja v in raw mode. It disables temporal noise reduction for the internal recording as well, even if the ninja is not recording. :)
Happy someone mentioned this. Really makes things difficult as a VFX artist. Have worked on projects shot with fx3 for B-cam and the temporal noise reduction creates all kinds of problems. Really wish Sony would release a firmware update that gives people the option of turning that off because using an external recorder makes the fx3 useless for helmet pov cams and things like that.
@@nicholaiexe If someone came up with some sort of hdmi dongle that just tricked the camera into raw mode I would buy it in an instant.
@@LDLCGO would be cool but the internal is still denoised if you choose to record internally while recording raw simultaneously
@@nicholaiexe In my tests I've found it's not temporally denoised. The only denoise it goes through is the encoder. Try it.
@@LDLCGO noted, will definitely take a look. Thanks!
In addition to flexibility and control over NR, one quirky advantage is that ProRes RAW is often faster to work with than the internal Sony codecs, avoiding the need for a proxy workflow for older PCs.
I’m pretty sure the shift in color and contrast can be explained by legal vs full. For some reason I’ve noticed that the Sony cameras output legal and the atomos records full. I bet if you were to use a legal to full conversion lut that they’d match up almost perfectly.
You are correct in assuming it's full vs video
But it's not a Sony issue
In my testing prores is recorded incorrectly on Canon and Panasonic as well
When recording prores the recorder doesn't know whether to save it as full or video, as far as I'm aware there is no metadata that tells the recorder to record it one way or the other, so you would have to actively enable the feature when shooting in log or know to interpret it correctly later
Checking the Legalize option when shooting log can help, the issue is when recording on a device that doesn't have this option, or when working with older footage not recorded with the legalize feature
Luckily you can just interpret the footage as full in your NLE
I really wish atomos would just come out with a press statement about this
Watching all the misinformation fly around or widespread people incorrectly using the devices is insane
I appreciate these tests thanks Josh. Every time I have decided I'm going to shoot raw I end up defaulting to an internal codec on almost every camera I've ever used. Most notably my C70 which obviously can do 12 bit raw internally at a more manageable file size than ProRes raw, but I never use it for paid work, only ever a couple of passion projects. It's kind of ironic but to my knowledge at least, most people who are shooting raw with more affordable camera offerings are doing it for the fun of it, not because it provides any tangible benefits to their professional lives. I bet I still end up messing around with raw again though, ha. That's life.
You're welcome. Thanks for watching!
No need Atomos, no need prores raw HQ. I did many tests with and without the Atomos and I can say, everything is on the grading; The color science of the FX3 is fantastic! It's not the Atomos will bring more in the image, except the sharpest like you said because it's a 12-bits. For the color balance, can be adjust 100% same. Good work here! Thanks
Gerald Undone did some testing and came to the same conclusion that you cannot get any more dynamic range by shooting raw vs 10bit SLOG3 at high bitrate. As for colours, try XAVC HS which has slightly different/punchier colours with a slight shift to green and better skin tones.
best video comparing the two I have seen so far. Thanks for the hard work 👍
Fantastic comparison Josh! Thank you for putting this together for us, you are awesome.
You're welcome. Thank you for the kind words and for watching!
Much appreciated and comprehensive. I think i must have missed the point that prores raw locks you out of davinci. Apologies if i did. For social media work no one will notice the extra coloyr but they will notice the lack of sharpening and NR thta they see all the time from so t cameras on youtube posts. There is a reason why they are so popular for youtube. And the turnaround isn’t usually justified for raw if it isnt blackmagic and resolve native. For professional work and especially multi camera or multi source wb and the ability to nail colour to spec is often more important especially for colour sensitive client products. Also pushing the image in terms of grade and look isnt something most clients need but the use of film emulation’s definitely benefit from a higher bit depth and raw. With these sort of straightforward images the differences arent really being tested. When you shoot mixed source lights with multi cameras typically an a cam like an fx9 or fx6 then having a raw image to match and adjust is very useful. That said for most youtuber social media users it’s important not to underestimate how much the internal camera codec is doing to help out. NR sharpening and lens distortion are very good examples that people shooting sony especially underestimate. Again no surprises that the creator shot in raw but thats an 80 million dollar feature and not a 2 minute social media post. Horses for courses obviously. Thx for the continued excellent and thoughtful posts.
You're welcome. Yeah that was interesting for me to see through doing these tests how much the camera is doing when recording internally. Thanks for watching!
i'm gonna tell people The Creator was shot internal
Pahaha
Was it really? 😅
False
Nooooooo ....
really great video! it might be interesting to do the same tests for the different internal compressions (I, S and HS)
Thanks for breaking this down!
You're welcome. Thanks for watching!
first video that convinced me that prores raw is actually better!👍🏾
The difference in detail and color depth to me is 100% worth it for anything important, not to mention lack of compression artifacts - especially in chroma channels in All-I recordings. To compensate for the noise and lack of ISO control in RAW exposure, I recommend creating a couple additional LUTs that will pull the image 1 to 2 stops. I typically will monitor with such a LUT in either base ISO and feed the sensor a stop more of light, which kills the low end noise (or two stops for the 12,800 base, making it effectively ISO 3200). You just have to watch highlights and when in a contrasty environment (lots of windows for example) then just expose with your standard LUT.
I get exposure control in stops with premier. It's the same as iso and works the same as using the iso in raw tab in resolve. Ironically, resolve exposure doesn't feel like it works the same and it's more like a regular exposure tool. That's after transcoding to DNG. Catalyst Browse will let you change iso too in xavc. When exposure is perfect nothing beats DNG in Resolve in my opinion.
Thanks for doing this Josh! That raw reminds me of the incredible beauty that comes out of the 5Diii with Magic Lantern on it so it can pump out raw. It's amazing, but such a work flow, and no modern advantages like good autofocus and whatnot. Sort of makes me want to get an FX3 at some point...
You're welcome. Thanks for watching!
IDK what you're using to convert ProRes RAW, but for anyone curious in my experience the free converter tool that's on several app stores to convert to cDNG does not set metadata properly for colorspace conversions, as well as failing to include the correct 3x3 matrix metadata for white balancing. This results in slight hue shifts depending on the camera once the cDNG file is opened in Resolve. The tools that convert ProRes RAW correctly are Adobe (such as Media Encoder) and Assimilate (Play Pro Studio). After that, the cDNG or regular ProRes files you create will work and look better in Resolve.
I’m using Final Cut Pro to natively edit the ProRes Raw. I show it in this video.
@@Josh_Sattin just saw that, apologies, thanks for the reply.
This comparison is what i was looking to see and everybody says you have more color information to work with on the prores raw. But nobody actually shows how far you can push the files till they fall apart in terms of color. That is actually what i want to know. Can you make an episode about heavy color correction comparing both file types? You only showed sooc files without any color adjustment and no film luts applied. I m having an anamorphic rig and because of many lens elements, the colors are way off on the chart. So i have to bring them back and the internal footage falls apart before applying any film lut. Will the prores raw save me ? And how much will help?
You can definitely push further with ProRes RAW. The starting point isn't as good sometimes though. I noticed I could stabilize better in post than using ibis, active mode, lens stab and catalyst Browse combined, then hitting stab. Color detail is on another level but less accurate to start.
Definitely need a comparison for the fx30
Unfortunately I do not have a FX30. A big difference with the raw on the FX30 is that it crops in on the sensor.
@@Josh_Sattin that's crazy smh
Thanks for doing this, I was interested in finding out how they differ.
You're welcome. Thanks for watching!
Thank u 🙏🏻 for this video i was thinking to buy zve1 or no because there is no RAW VIDEO option but i changed my mind actually 😊🌷
You're welcome. Thanks for watching!
thank you Josh
You're welcome. Thanks for watching!
Thank you for the NERD info! I hate when people just talk and talk and don't SHOW. Earned another subscriber! PS. I am a Fuji still shooter, do you think the FX3 is still worth it today, at almost the end of 2024...?
You're welcome! I appreciate the support. FX3 is a great camera and will be so for many years. You can see my long term review here: ruclips.net/video/L2EDzbCUeYo/видео.html
My understanding of ProResRAW on Sony is that it doesn't increase dynamic range, you're still limited to whatever the camera offers in SLog3. So in that sense most of the differences are going to be in color, and then the difference in how colors work on a Log curve vs. raw sensor data. One of the advantages to me would be that RAW bypasses all of the cameras lens corrections and noise reduction, in that sense it offers a little more freedom in post processing allowing you to precisely control that. Overall, for me there isn't enough of an advantage to RAW over the internal codecs. It is also an absolute pain to try and work with on Windows, I've given up trying to use Premiere with the RAW plugin, it's too unstable. So I convert to CinemaDNG for Davinci Resolve which is not ideal, however I haven't seen any degredation when doing so, but CDNG has some odd color space and gamut choices that would not be present if working natively on OSX with Final Cut.
It’s honestly not that much of a difference but good to have the option.
thank you for this! great video!
You're welcome. Thanks for watching!
Note that saturation and perceived "sharpness" are both factors of contrast. If you adjust the internal recording to match contrast, the saturation and sharpness would also match.
Yep. Adding contrast adds saturation and makes the image look sharper. I find it likely that matching levels would largely mitigate the difference.
To me it looks like there's something fishy with the conversion. There should be no color difference whatsoever, if the conversion is done correctly. Or, it could be that the internal recordin's picture profile settings are not at "default" values, at least not the same "default" as the NLE assumes.
Also, when people claim that they can get more info on RAW than on the internal log files, sure that can be true, technically speaking. Is it really true in practice, in real world color correction cases is a whole another question. Using 4 x the data rate to save pristine chroma noise is not my idea of smart.
Which reminds me, i think i'm seeing some debayering artifacts with the raw - the black plastic texture in the chart seems to have some discoloration, red and blue static noise, that's different from the noise that changes from frame to frame... quite visible in the 200% zoom.
I always make my videos as ojective and transparent as possible.
Yes I know you can change the perceived sharpness of an image with more contrast, but in this case the Prores Raw is slightly sharper than the internal codec. This is true even if I adjust the contrast to match. It might not be perceivable through RUclips and that is challenging but I stare at this on my 32" 4K monitor before making these conclusions.
There is nothing fishy going in with the conversions. I literally showed you what I used in this video. I am also editing in Final Cut Pro and this is the native editor for Prores.
Please feel free to replicate the tests and post your results for everyone to see. This is what I saw with the gear that I have. I hope this was helpful for people interested in the differences between these two recording methods.
@@Josh_Sattin Hi Josh - i do not meran "fishy" as in you doing something wrong, and i am sure you saw what you saw.
I mean "fishy" as in the conversion is the very likely culprit for the differences, rather than the file formats themselves. The software we use is rarely perfect, and the cameras may not be doing the conversions exactly equal to what was specced by the manufacturer for the software vendors in practice. There are quite a lot of additional settings in Sony cameras' color profiles, which may differ from the defaults used by editing software.
There definitely are also differences in how various software handle footage - starting from whether you see the actual full range in the file, or just the legal portion or Rec709. One can't really completely trust the editing software to work correctly.
I do not own FX3, but I would actually be quite interested in comparing the original files from the camera myself, if you are willing to share a take or to, and share my results with you (and the rest of your audience, if you allow me to).
I know you don't have a FX3 but here is some additional information to help clarify. When shooting in Cine EI in the FX3 you can't make any adjustments to how the camera records the Slog3 information like you could adjust Slog3 with the older system of adjusting picture profiles. It is what it is.
There aren't any of those full or legalize settings in the Ninja V when recording ProRes Raw either because you are recording raw.
Using FCP and the built in decoding and LUTs is the most objective way I can observe and present the differences. Also FCP is a NLE that can handle ProRes Raw natively and it converts the raw footage into Slog, which I show in the video. So the differences you are seeing are the differences that we get.
Once you'd mentioned noise reduction I hoped you'd actually compare internal vs mannual
Yes. Raw is worth it. Feels so much better
😂
Thank you!
You're welcome. Thanks for watching!
Good stuff :)
Thanks for watching!
Hi! Thanks for the video. I wonder if it's expected by Sony to release a mirrorless camera that can internally record ProRes anytime in near future?
I shot my FX3 w Ninja and the external files recorded at 1920 and not 4K. I can't find how to switch resolution on ninja? I just update ninja to OS11. thanks with any help Cheers
seeing the individual eyes blinking is soooo uncanny 3:42
Great video, thank you.
You're welcome. Thanks for watching!
Cheers to me internal's better skin tones but raw looks more cinematic except that can be easily matched by white mist filter 😊❤
I love how we have the option to shoot 12bit, id rather save money and just shoot 10 bit its good enough for my work.
on point.. i love it
green has always been more saturated out of the gate when I shoot raw on my FX3. Just a lot more info
What happens with lens correction on the ninja footage? Is this something that you compromise when you use the atomos recorder?
Your client will never notice the higher quality, but why would any one wants to deal with the raw workflow? Raw workflow results in more cost (storage ) At the the end, you've getting paid the same per your contract agreement.
Waste of time and money for most videographers. Not worth it for the average client. Only useful for big budget productions and pixel peepers.
I believe Davinci Resolve Studio now has ability to recognize XAVC S-I from FX series sony line as RAW, and can be plugged in like RED RAW footage.
It has to be in a .mxf wrapper
The RAW 14bit 1080p and 3.5k 14bit lossless I get from a 5D mark III has that same odd green tint, its so odd
How is ML?. Do you feel the need to upgrade to one of the modern mirrorless cameras?
Comparing Prores Raw 12bit to All Intra 10bit is like comparing 12bit compressed RAW vs 10bit HEIC Image, you won't able to see much difference if you only grading it slightly
excellent thanks
You're welcome. Thanks for watching!
Hello, thank you for video. I have a question. What is faster for export on MBP late 2023 16” M3 pro 36gb. 1 hour video ProRes or XAVC S-I? I like 12bit colors and option. Thank you
Been shooting prores raw for a few years cant go back to internal recording period
Great tests! The luma and color density differences make me wonder if FCP is properly showing the raw and S-Log3 levels. Did you happen to try doing a Full to Video levels LUT to see if that makes the luma and saturation levels of two line up better? I wish FCP had an option to choose to interpret via full and video levels.
Thanks! I used the standard conversion LUT on both clips for the test charts. For the outdoor examples I copied the grade from the internal to the Prores Raw. I explained this in the video. The differences in both sets of shots were consistent.
@@Josh_Sattin I believe FCP assumes the file's levels are Legal, unless they have metadata to assume Data levels. Internal S-Log3 is recorded in Data levels, while ProRes (and raw?) are using Legal levels.
About contrast - it can be issue on atomos. Because on my 7iv I have the same thing in xavs-i and prores hq. Prores hq have more contrast then internal.
It could be, but the only way to record ProRes Raw is by using the Atomos.
@@Josh_Sattin yes. I know. I mean that difference in contact can be due to this.
Is there difference in prores hq and prores raw?
Well there is no other way to record Prores Raw so this is what we get in Prores Raw.
I did not test the difference between Prores Raw and Prores Raw HQ.
Hey! Nice comparison! I have a question, did you use an ND filter in front of the camera? Probably The magenta cast would produce that render in outdoors
Thanks! Yes I used an ND filter for the outside shots but both files were recorded simultaneously.
@@Josh_Sattin thanks for answering Bro
You're welcome. Thanks for watching!
Is there an advantage in recording raw in a log profile?
When you record in raw, it only records one way, in raw. You then convert it in your editor. I explain it all in this video: ruclips.net/video/0ZqJ46L7-74/видео.html
@1:15 there is nothing called half a frame. 1 frame is the least . And moving one of the clips by 1 frame should solve the issue. For color grading . If you only adjust the contrast and saturation without using the official Lut you will lose dynamic rang. Gerald undone made a video about it and showed that this method is not good
I am using the camera recording to trigger recording in the Atomos. The Atomos recording does not start at the same time as the internal recording. There is a delay. When I try to line up the clips in post the closest I can get them is a half frame apart. It's just the delay that is built into this configuration lands it in between two frames of the internal recording.
Quick question the Fx3 is 16-bit RAW Out , ProresRAW can handel 16Bit but is it downsampleing to 12-bit?
It can only record in 12-bit. I go over that in this video: ruclips.net/video/0ZqJ46L7-74/видео.html
@@Josh_Sattin I see! It seems they have step away from letting them record more bits. I rember seeing this keynote from Jeromy back in 2018. ruclips.net/video/9f3G9mP9Tsk/видео.html. Dont know why they choose to cap it at 12-bits when all Sony cameras clearly are capabel of moore. Any how thanks agin for a fantastic video! Really love the way your cannel has been evolving!
I swear y'all and not you personally gotta just love editing 😅 I get that it's a video to make money off of, but all I can say is, the right tool, for the right job. If you're running a business figure out the camera that does all that you need in-camera. Get your damn time back. I know people love to hate on Nikon, but Nikon can do what prores raw does using Active D-lighting straight out of camera. You can add a custom curve if needed prior to shooting. I get where Raw is needed, but that just means you need better lighting skills which we all do and its a life long process to perfect. They dynamic range is never going to surprise you until you can surpass 13 stops otherwise it's just an exercise in redundancy.
Im gonna do some prores raw shooting this christmas on a new ninja v coupled w an fx3. Just wondering if i change iso in post, does that mean I’m deviating from native iso and therefore changing my dynamic range to the worse? Should i only change the ”exposure adjust” and stick to 800 and 12800 in the iso tab? Thank you
As an Adobe premiere user. I have no idea how to apply noise reduction. I don’t think there is anything directly in Adobe that allows you to do it to my understanding.
Can anyone help me figure this out?
I use (older versions of) Neat Video noise reduction (plug in). Topaz AI does also a good reduction (stand alone app).
the blinking😂😂😂😂😂
Nice video, very informative. I'd still like to know in which situation Prores Raw is way better than the internal recording of the FX3.
I do not know for certain, but i find it likely that the answer is "none". Perhaps in some extremely high ISO situations, where using noise reduction in post may give better results than in-camera.
Thank you. As I stated in the video, the main advantage is more control over the noise reduction and a heavier file (bitrate and bit depth) for more advanced color grading and manipulation.
I was wondering if you have a subscription to storyblock and publish to RUclips but after some time you decide to cancel it, does the previous published material needs to be removed?
“Royalty-Free” is a type of license that allows the unlimited use of content for any media projects without having to pay royalty fees. All footage, templates, and motion backgrounds are royalty-free and they are yours to keep forever - even if you don’t maintain your Storyblocks subscription.
help.storyblocks.com/en/articles/3622237-what-does-royalty-free-mean
I just encountered something weird. I recorded internally with XAVC-S 4K and externally to NINJA + in 4K ProRes RAW on Sony FX3. I fed audio to the camera via Rode Wireless Pro.
I'm currently trying to sync audio. The internal FX3 recording and the Rode Wireless Pro WAV file match perfectly.
However, the Ninja V+ ProRes RAW file's audio does not match perfectly with either the Internal FX3 recording's audio, nor with the Rode Wireless Pro's audio. There is a slight drift.
I even converted the MOV files into Cinema DNG and tested it in Resolve and I get the same exact issue I was getting in Premiere. In just once scene that is 3 minutes long, I can sync the audio at the beginning of the scene, and by the end of the scene, the sound is 4 full frames out of sync. I'm wondering if anyone else had had this issue?
In theory you don't need RAW capture, unless your ISP does things that you are out of control... (like in my world smartphones)
Can the FX30 also do this?
Yes but it crops in on the sensor.
And why on that clip of you in the shadows of the trees, is the pro res so much noiser? Like I've noticed on a lot of cameras that can record raw externally that it gets wildly noisy in the shadows
As I explained in the video, the internal recording has noise reduction applied and the raw does not.
@@Josh_Sattin I made this comment before I got to the end my bad lol
what about XAVC HS ?
What would you like to know about HS? I recorded in S-I for this comparison because it's the highest quality that can be recorded internally in the FX3.
WHICH ONE IS BETTER THE FX3 OR THE C70?
FX3
Depends
this is the correct answer
Was yelling necessary?
yes!!!@@POVwithRC
What is ur studio camera here?
Blackmagic 6K G2 - geni.us/JD5fRDA
Sigma 18-35 1.8 - geni.us/0JZdi
RAW really never is about better quality as such, just more flexibility
Do we really need to drink water?
Sponsorship gang. 😁
Looks like a legal vs full by the looks of it.
Don’t see enough of a difference for the raw. That said where’s your iPhone log videos lol
Haha. I have an iPhone 12 Pro.
we want the 15!!!!!! ;)@@Josh_Sattin
@@ReachFilms Haha!
it's 16 bit vs 10 bit
The external raw is recorded as 12 bit ProRes Raw. I explain it in this video: ruclips.net/video/0ZqJ46L7-74/видео.html
Thank you for the nerd stuff. It matters.
You are welcome. I love this stuff!
Ping!
CGI, AI special effects will work better with raw files.
The way you describe prores raw being more contrasty and more saturated leads me to believe your NLE is not interpreting the data levels correctly
Make sure the prores files is being interpreted as a full data level and not a video data level
This is a common, industry wide issue where the NLE doesn't know to interpret LOG prores as a full signal, resulting in the whites blowing out, the blacks crushing and the colors being too boosted
This has also been in issue for as long as I've had the a7s1, Sony claimed it was an atomos problem, atomos claimed it was a Sony problem
Gerald undone got it wrong by saying it was a Sony problem (I 1000% believe that and would be willing to prove it)
Atomos added the ability to record in a full data level "Legalize" which fixes the issue, this is what they should have done years ago and just shows it was always a problem on their end (to be fair, there is an inability to flag log footage when recorded externally to be saved as a full level)
Note though that Legalize records the image in a full data level right out of camera, so switching a clip shot in full data level to full data level would give unintended results
16 bit and the abitity to set ISO and White balance in post (Mic drop) ProRes RAW WINS!
All these influencers speaking like they are professional colorist and will be fixing stuff in post.
pro res raw is no real raw
Real enough that The Creator DP was happy with it
both sucks i guess
yawn, lets see some of your client work now brother
There is some in my FX3 review: ruclips.net/video/L2EDzbCUeYo/видео.html
thank you good sir @@Josh_Sattin