Jump to content

stephen

Members
  • Posts

    135
  • Joined

  • Last visited

About stephen

Profile Information

  • Gender
    Not Telling

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

stephen's Achievements

Member

Member (2/5)

104

Reputation

  1. Would rather put a new camera in series of practical tests shooting scenes that typically shoot, rather then set expose based on some assumptions. Those assumptions may not be true. We have no idea how exactly a particular manufacture is interpreting the RAW data from a given sensor. If we have to believe this test: Sigma FP shooting raw video handles underexposure much better than overexposure. Some people interpret the Sigma RAW video as BM Log and then do CST (Color space transform) and claim this is the way to extract more dynamic range to the full potential of the sensor (12.7 stops). CDNG files from Sigma RAW video according to them give lower dynamic range. Have similar experience with Canon ML RAW video on hacked Canon 5D Mark III and Canon EOS M. There is ETTR function in the hack and multiple people advice to do ETTR in order to reduce noise when shooting at HIGH ISOs. In practice at lower ISOs (and high as well) risk of burn in highlights is quite high. Canon RAW CDNG files are handling underexpose better than overexposure. I was able to recover at one occasion 4 or 5 stops of underexposure due to forgotten VND for daylight when shooting at night. As result when shooting Canon ML RAW tend to underexpose half a stop, and almost never go for ETTR. And like the results and footage better.
  2. Sigma 24-70mm f2.8 DG DN has almost 0 breathing and manual focus is smooth for a focus by wire lens. With latest firmware that gives extra control over focusing similar to Panasonic lenses it is almost perfect for video. Was pleasantly surprised. But it is relatively big and heavy 🙂
  3. P.S.2 - Forgot to say that when using the hard stops ND filters will use them to adjust exposure in steps of 3,6,9 stops. And then use aperture (if possible) or ISO to fine tune in between the 3 stops steps. I use Dual ISO or second (high) ISO mostly at night. For daylight shoot use the first base (lower) ISO but will not shy away to increase it with 1 -2 stops in order to tune the exposure. Or just overexpose / underexpose a stop or two depending on camera that use. Most cameras tolerate better overexposure.
  4. P.S. If am using fixed focal lenses and changing them during the shoot 2 hard stops (3 and 6) magnetic ND filters is the way to go. Gives both good quality and convenience quite rare when using ND filters and on limited budget. Bought K&F just to try the concept and it worked. Overall am pleased with quality as well. There are better but more expensive magnetic ND filters.
  5. My use of ND filters is mostly for outdoor / daylight video shooting. Almost never need them or use them for indoor or outdoor at night. Here is my experience so far: When shooting RAW almost any cheap VariND will do the job. Like K&F VariND Hoya ProND hard stops ND filters are good enough for 10bit 4:2:2 log footage. They may have slight color cast but it is easy to correct in post. Basically for non RAW footage you can correct WB up to 700-800K, even 1000K but no more. Most good quality hard stop NDs will give you only 100-300K deviation (colour cast) from correct value (without ND filter). Last generation of VariND filters split the 1-10 stops range in 2 different filters to improve colour fidelity and avoid the dreaded X pattern at the extreme end. Based on several reviews decided to buy/try Freewill VariND 6-9 stops (82mm). Curiously this one has no colour cast or let say no visible colour cast and keeps the sharpness of the image but the 1-5 stop one does have quite significant colour cast according to one reviewer. For this reason and based on my calculation about ND stops that need bought only the 6-9 stops one. Also bought 2 magnetic K&F 62mm hard stop filters - 3 and 6 ND stops. Freewill 82mm VariND goes on my 24-70mm f2.8 Sigma or 24-100mm f4 IS Canon. K&F 3 and 6 stops on all fixed focal lenses and of course I never mix them. Both setups are giving me convenience / ease of use and also good quality. They work so well that decided to add another 62mm Freewill VariND and 2 82mm K&F. Same ND range and stops. Most modern cameras when shooting LOG or RAW start at 400 to 800 ISO range. For example my BMPCC 4K base ISO is 400 and my Panasonic S1 base ISO is 640. Having in mind the sunny 16 rule, those are 2 to 3 stops more than base 100 ISO, another stop more comes from the fact that instead of 100 shutter have to use 50. Usually shoot at 24 FPS. And another 2-3 stops come from apperture where typically f5.6 and f8 are my preffered ones when shooting outside at Super35 or FF. When shooting people apperture has to go down to F2.8 on Super35 and f4 on FF. Those are typical and most used appertures in cinema and movies. So 6-9 is just perfect and can stay 100% of the time on my 24-70 f2.8 or 24-100 f4 lenses. Apertures like f1.4 or less or even f2 at FF are almost never used in cinema as it is very difficult to manual focus at closer distances and focal lenght around 40-50mm or more. The strongest ND filter you can get is 10 stops. Two K&F 3 and 6 stops filters are used the same way. By stcking them can achieve 9 stops of ND and have the convinience of 3 stops of ND as well if needed. But as said this is rarely the case. To obtain correct exposure: - when shooting scenery use apperture variations from f4 till f11 + ISO variations. Overall this give me 5 stops of tollerance which is more than enough. - when shooting people and apperture is locked at f4 or f2.8 can change only ISO. Changing lighting is most of the time not an option. In both cases when using the VariND filter will use the 3 stops it gives me first before apperture and ISO. Most modern camera have huge lattitude - up to 6 stops with Panasonic, so underexposing a stop or over exposing a stop or two are almost never a problem I use also only 3 white ballance points - 3200 typicall for night shots, 5600 for daylight and 4400 for something inbetween. And it works quite well for me in both RAW and 10bit 4:2:2.
  6. My approach is similar to the one projectwoofer described. It consist of two main steps 1. Color correction and conversion from Panasonic V-Log to REC 709 or another color space if step 2 will be executed 2. Color grading creating certain look. For step 1 tried: - Davinci Color Space transform (CST) for Panasonic V-Log to REC 709 - Panasonic provided conversion LUT from V-Log to REC 709 - 3rd parties conversion LUTs in my case from 55Media, which is similar to S1Alex conversion LUTs from V-Log to Arri LogC I can get similar results from all 3 methods but the easiest one and the one giving me slightly better results are correction LUTs from Media55 (Panasonic V-Log to Arri LogC). Also have in my Davinci Resolve node tree, nodes for White Balance, Exposure Correction (primaries), color saturation, contrast and stabilization. Most of the times just perform step 1 and convert from V-log to Arri LogC to REC 709, adjust individually white balance, exposure, saturation etc. on the individual clips if needed and that's good enough for me. Processing all clips this way is easy, fast and gives me good and consistent results. For step 2 tried - some LUTs - Dehancer which works only in Davinci Resolve and gives you great film emulations. Tried some of them on few small projects and so far like the results very much. There are reviews comparing Dehancer to Film Convert and they say Dehancer is a little bit better. For sure you will be able to get great results with Film Convert too. - using some film emulation in Davinci Resolve and more specifically Kodak 2383 print film, which I also like. Juan Melara has LUT and Power Grade for Kodak 2383, one day I may get them. - using this simple trick in Resolve to steal looks from movies or clips that I like. Later may try to replicate the looks with power grades. For most of my projects 70-80% just do color correction / conversion and that's good enough for me. Is it better than Natural Profile in Panasonic S5/S1/S1H ? It looks so for me. Better dynamic range and better control on colors, saturation and ability to use a look / grade. For the rest 20-30% use 2-3 looks that I like using Dehancer, Kodak 2383 and eventually power grades that am working on. Having another 2-3 special looks in future is maybe my limit. Using an individual look / grade for every projects, that's for well funded commercial projects and Hollywood 🙂 Having 3 to 6 looks and using them would be good enough for me. Hope this helps
  7. Film making is a broad term. There are two main flows and sources of material to learn from - fellow one man or small company video maker - cinema - both fictional and documentary In cinema there is a pool of people specialized in different areas and having different roles. In one man / small companies roles from cinema are done by one person or fewer persons. There are also examples in cinema when one person was working in different roles but those are rather exceptions. IMHO the best educational material comes from people working for cinema. Here is a list of those roles, it is for sure not complete: producer director actor DP, camera man Editor Color Correction / Grading screenplay, writer composer audio special effects and graphics Which one you are interested in and want resources for ? Walter Murch is an editor. Are you looking for resources related to editing and cutting ?
  8. @Mark Panasonic S5 (+ MC-21) clearly uses both in body and lens stabilization. That's what it says in the menu - lens stabilization for up / down left / right movements and in body stabilization for body roll. When I switch lens stabilization off from the switch on the lens on LCD screen symbol for stabilization turns off. At 10mm and even 18mm video looks super smooth don't see warping in the corners.
  9. Yes Panasonic S5 + MC-21 works with Canon EF-S 10-18mm and switches automatically to APS-C mode. Auto focus and everything else works. Unfortunately you can't switch manually to Full Frame mode. It is well known fact that Canon EF-S 10-18mm covers full frame from 14 to 18 mm. Corners are far from perfect on full frame, but for video we cut them anyway. Mine however is with modified mount, it can also be mounted on a full frame Canon body. Not sure if the original APS-C mount which protrudes out will be able to mount on MC-21. It looks camera Panasonic S5 + MC-21 doesn't recognize EF-S 10-22 as APS-C lens and thinks it is Full Frame lens. Well then switch to APS-C mode manually (should be possible) and vignette will disappear.
  10. Have both BMPCC 4K and Panasonic S1/S5. For me both are great cameras for video/cinema. Historically bought BMPCC 4K first 2-3 years ago and resisted buying Panasonic S5/S1 mainly because want to shoot in RAW. Image quality has the highest priority for me when choosing a camera. Panasonic cameras initially didn't have RAW video recording capability. Recently Panasonic added ProRes RAW and BRAW as external recording options. RAW video plus very good prices second hand tipped the scales for me. Didn't test them extensively side by side. From the limited tests that did can say both BMPCC 4K and Panasonic S1/S5 for me have excellent image quality in BRAW. 10bit 4:2:2 h264 internal from Panasonic is probably a notch down but very close. 10bit internal is now OK for me to use in some projects. Initially that was not the case. My plan was to use only BRAW external from Panasonic. When first reviews of Panasonic S1H started to appear, downloaded several clips in 10bit 4:2:2 V-Log and graded them. Was looking mostly for skin colors. Was not satisfied, they didn't look as good and as easy to grade as BMPCC 4K BRAW. Log picture profile that both have is like negative. Until we develop the negative, color correct/grade it in post, it is not possible to talk about color. What kind of color you get in the end is more skills in color correction / grading then camera color science. Agree with Kye on this. As said, I downloaded several clips shot with Panasonic S1/S5/S1H series of cameras internally in V-Log 10bit 4:2:2. Color corrected them in Davinci. Some shots and more specifically skin tones didn't look as good as they look on BMPCC 4K. It is all subjective we have to admit this, but that was my impression and assessment. For both cameras - BMPCC 4K and Panasonic S1/S1H/S5 used the same method (Color Space Transform) to convert from Log profile and their respective color spaces to REC 709. Since both BRAW and Davinci Resolve are Blackmagic products, getting great colors with simple CST is easy with BMPCC 4K. This method however does not gives always satisfying result and colors for Panasonic internal 10bit V-Log. So I tried Panasonic V-Log to REC 709 conversion LUT and 3rd parties Color Space Conversion LUTs (55Media). They gave me better results and better starting point. Adding some simple corrections, contrast, saturation, balancing exposure etc. after the conversion LUT now gives me better colors and skin tones. There are some test footage from both cameras on interenet. This one has downloadable source BRAW (BMPCC 4K) and MP4 (Panasonic S1H) files. https://www.youtube.com/watch?v=4Al-tIFXFuY Among the sample footage there is one clip for checking skin tones. Lighting is quite challenging and constantly changing and shots are in different time, he shot with one camera then the other. Still it is a good exercise to try. When I tried it first almost a year ago with CST used for both cameras, didn't like the colors from Panasonic. Doing it now with conversion LUTs get better colors from Panasonic and can match closely both cameras. More examples and tests here: https://www.youtube.com/watch?v=rghbaHoVRsk https://www.youtube.com/watch?v=wB2ZkRaPWjY https://www.youtube.com/watch?v=rU4eydFl1hM https://www.youtube.com/watch?v=fJawO5eeWr8 While I agree with Kye that getting good colors is mostly skills in editing program of your choice, don't think that getting a good starting point, correct colors or pleasing colors is that difficult. Usually use 4-5 nodes maximum for this task. It takes a lot of time and skills to build a certain look, to finesse every take, emphasize some lighting etc. This requires a lot of knowledge and practice. But basic color grading/correction should not be that difficult. With Panasonic 10bit 4:2:2 V-log internal you have to try different methods, correction LUTS, etc. It may take some additional efforts and time. Once you have the new workflow, it will be easy to apply it and it will be more or less consistent so at the end you get colors you like. I usually stop at this point and do not go for a specific look like teal / orange etc. One day maybe but for the moment it is OK. A good example of what am saying is Zeek. His videos shot on EOS M + ML RAW for the last 2 years have consistent and great colors. Yet his color correction / grading workflow in Resolve is very simple. The difference and reason why we are saying we like colors from camera brand A better that Camera brand B is mainly because our workflow to develop colors from Log profile and camera color space to REC 709 is giving us colors we like. Once we change workflow or color correction/grading is adjusted for camera B that we didn’t like, problem is solved. We can again get colors we like. Am not saying those are correct colors. About LUTs mimicking Alexa. Of course, they are not 100% close and in some lighting condition may be off. It would be important only if you try to match Alexa footage with BMPCC 4K in one video for example. When we talk Alexa colors again we confront the problem how did we get those colors. Even if we download conversion LUTs from ARRI website (which I did), there are different LUTs for different Arri Camera models and color space transforms. Depending which one you use you get slightly different colors. So what are after all Alexa colors? For me those LUTs simply are good starting point, a reference. That's it. Advantages and disadvantages of BMPCC 4K over Panasonic S1/S5/S1H is another topic. You have to try both in order to make the best decision for you. As said many times - it is personal; everybody should try himself or herself. For me Panasonic S1/S5/S1H have few advantages, so tend to use Pana S5/S1 more.
  11. Fixed magnetic ND filters https://fstoppers.com/reviews/review-haida-nanopro-magnetic-nd-filters-570861
  12. Depending on lighting using fixed ND could be OK or major pain in the neck. If light doesn't change often, there is no need to change ND filter as well. But if it does change or you move often from bright sunlight to deep shadows then screwing and unscrewing filters becomes really annoying. At least for me. There are magnetic fixed ND filters as well. They are more expensive but IMHO they will greatly improve usability of fixed NDs. Am willing to try them
  13. I have used both VND and fixed ND. VNDs are OK when shooting RAW as it is relatively easy to correct the color shift. For fixed NDs always use 180 degree shutter when shooting at 24fps and depending on the available light choose an ND which is a little bit stronger and adjust (usually raise) ISO for correct exposure. Or vary with aperture if possible. For examaple if I shoot at f8 may raise it to f11 or lower it to f5.6. If aperture is open and can't close it for esthetic reasons (blurred background) then ISO is the only option. You don't have to worry about base ISO because all modern cameras have huge latitude. Several stops (3-5) above and under the base ISO. Even if you overexpose 1-2-3 stops or under expose 1-2 stops you don't lose information or dynamic range. See for example Panasonic S1/S5 tests in 10bit 4:2:2 V-Log internal on slashcam.de or this test Normally use no more than 3 NDs and sometimes even 2. 3, 6, 8/9 stops, or only 3, 6 stops and then by stacking both 3 and 6 stops ND get 9 stops. This works with most lenses but not all. Vignetting at the corners is not a problem with most lenses because shoot at 2:1 ratio for video and corners are usually cut out of the picture.
  14. It does only if you need: - external monitor - better, faster codec for editing - 4K 60p unlimited In terms of picture quality can't see any difference. Both internal 10bit 4:2:2 h264 and external 10bit 4:2:2 ProRes have the same dynamic range and colors. S5 can shoot 5.9K internal but with time limit and in 10bit 4:2:0 h265. This codec is difficult for editing, so I never use it. 4K is good enough for me. I had the Ninja and found the round trip from Final Cut to Resolve time consuming and cumbersome. So after Panasonic gave us BRAW external, sold the Ninja V and bought Blackmagic Video Assist 12g 5''. In BRAW color science is slightly different , Panasonic provides second color space conversion LUT additional to the one that does V-Log , V-Gamut -> Rec 709. Colors are slightly better, richer and you get the ability to set White Balance in post. For me 10bit internal recording is good enough most of the times. But have this maximalist kind of "Only the best is good enough" mentality. 😀 So I often use Video Assist and like it very much. First because I like having a bigger external monitor with cage and top handle. It makes my shots and camera movements smoother, better and easier. And second I like the extra flexibility in post (set white balance in post mostly). If I have to be nimble and don't want to draw too much attention, Blackmagic Video Assist stays in my bag and there is only a small cable hanging from the camera. Or simply shoot internally. If you have the Ninja V, just try if a small rig with Ninja V on top, or just the monitor on top is OK for you. And ideally rent Video Assist for a weekend if you can. I have found that trying equipment myself is critical and has no substitute. No amount of reviews or advice from colleagues can help me make a decision better, then trying it myself. We are different and have different needs. Found both Panasonic S1 and S5 second hand at a very good prices (around 1100 Euro per camera) So now have both and like them both 😀
  15. No mechanical shutter. Wow, missed that. Brave move from Nikon.
×
×
  • Create New...