-
Posts
7,964 -
Joined
-
Last visited
Content Type
Profiles
Forums
Articles
Posts posted by kye
-
-
2 hours ago, androidlad said:
Sensors cannot distinguish image content and adjust exposure on a per area/per subject basis.
Sony has a working 1/1.55" automobile sensor IMX490 that achieves 20 stops (120dB) dynamic range in a single exposure. Basically a 3um photosite with a 0.9um subpixel. Three readout for every exposure - big pixel + high gain, big pixel + low gain and small pixel, them merged to get an 120dB signal.
20dB would be just fantastic! What bit depths are they talking? No point having 20 stops of DR and then only having 10 or 12 bits - by the time you convert back to a standard gamma curve it would be banding central.
Could you post a link to your source?
-
51 minutes ago, thebrothersthre3 said:
I think this will come down to your color manipulation talent. The pressure is on ?
Indeed it is.. good thing I never claimed that I can do it, only that I would try! ???
My suspicions are that colour matters, but that there are other elements to it as well. @webrunner5 has commented that it looks too clean, and I've heard many other similar kinds of comments around the place too. To me, those comments seem to be similar to the comments about modern 4K cameras being too sharpened, so I have a hunch that the difference between sharpened 4K and P4K RAW is maybe similar to the differences between P4K and BMPCC.
If you spend time looking at film stills they have beautiful colour (very high bit-depth!) and of course are uncompressed, but they're also quite soft in comparison to the sharpness of 4K, even if not the resolution of 4K. It is possible to reduce the sharpness of an image without reducing the resolution, and of course the 1080p RAW from the BMPCC will have less resolution than the P4K in 4K RAW too.
I believe that the softening effect of vintage lenses on sharpened 4K footage is one of the reasons they're so popular on these forums - they help to give the look that many of us enjoy.I'm thinking of things like blurring the footage, blurring the footage and blending it into the original footage, downscaling and the upscaling the footage, as well as adjusting for colour.
My plan is to colour match the footage as much as I can, then pumping it through as many different ways of processing it as I can think of and having them one after another and then posting it to get people's impressions of what works and what doesn't, then trying different things based on feedback. It will be interesting to see what the results are.
In the end I'm hoping that even if we don't match the 'look' of the footage from the older cameras, that we manage to find some things that get part of the way there, and then we can explore what those techniques look like on other cameras like the GH5. It would be great to get to a point where we know the settings to make H264 look a lot more like these classic cameras. Plus if it's a combination of various things, having it in a Powergrade like Juan Melara did to replicate the LUTs will mean that we can refine or disable each adjustment to taste.
@graphicnatured - just a thought, is it worth shooting the P4K in 1080 Prores as well?
If we could make P4K 1080 Prores HQ have the classic look then I'm sure that knowledge would be of real interest to a lot of people here.1080 RAW would be severely windowed so difficult to match framing on, and what we learn from matching 4K RAW and 1080 Prores can probably be combined for processing 1080 RAW if anyone decides to shoot in that mode to extend their lenses or get 120fps.
-
17 minutes ago, androidlad said:
Thanks for posting the links, it would be more sensational to post that BBC study's editorial title "H.265/HEVC vs H.264/AVC: 50% bit rate savings verified", yes but only at lower bitrate. This is exactly what I meant.
For "higher bitrate" tests in the second paper, is that really how you interpret the figures and the study itself? Then please explain how HEVC manages to achieve ~33dB PSNR with a bitrate of 0?
It would be more sensational, but I'll leave that to you, I'm more interested in facts.
In terms of how I interpret the data on those graphs, I do it by looking at the datapoints that I quoted. If you're not sure how to read graphs properly then there are many good online courses from reputable university courses available.
If you disagree with their methodology or results, then I look forward to reading your published and peer-reviewed paper on the subject. Or actually making some kind of sensible criticism. Zooming in to a low resolution graph and assessing the data point that's less than one pixel wide is just stupid.
I blocked you previously because of your disgusting personal attack on someone else here on this forum, but have been reading your comments because I thought that despite having basically no inter-personal skills (or no compassion) you had some technical knowledge to contribute. But if you can't even read a graph properly, or more likely you're deliberately mis-reading it because you value being right more than the facts, then I'm not sure why anyone would trust anything you contribute.
5 minutes ago, thebrothersthre3 said:It has 10bit 4k 60p which the GH5 lacks.
I looked for that mentioned above but couldn't find it. Maybe I missed it?
4K60 10-bit with HLG would really be something if it could do it!
-
On 1/19/2019 at 4:15 AM, Mark Romero 2 said:
You could - instead of using a lut in resolve - use the Resolve color management feature, which is something I highly recommend. You can set it up to effectively transform the gamma and gamut.
Avery peck has a good video on youtube on how to use resolve color management.
+1 - Just use Resolve's integrated conversions.
In fact, there's a big difference between using the conversions in Resolve and a LUT:
- If you use a LUT and the conversion clips any parts of the signal (highlights or shadows) then they're clipped forever and nothing you do after the LUT can get them back.
- If you use Resolves conversions (either in the Clip properties or via the Colour Space Transform plugin) the clipped values are retained within Resolve (as super-whites or super-blacks) and if you adjust the image after the conversion then you can get them back into the normal range without damaging them.
The internet talks a lot about LUTs but that's mainly because the people doing all the talking are selling...... LUTs.
I don't know how the other NLEs work, but I'd imagine they work similarly.
If you have to use a LUT then you can lower the contrast before the LUT to get the output from the LUT within range, but this defeats the purpose of using a LUT in the first place (because your inputs to the LUT now don't match how the camera encoded them) and you may as well just apply contrast or curves to get the look you want and ignore the LUT.
- KnightsFan and Mark Romero 2
-
2
-
9 hours ago, androidlad said:
Sorry but I'm gonna correct it every time I see this. 72Mbps HEVC is identical to 72Mbps H.264. That 50% bitrate saving marketing BS only applies to extremely compressed materials for streaming. For 4K it would be around 10-20Mbps for HEVC to show its strength.
The reason they offer HEVC is because BT.2100 HLG spec stipulates it.
I totally agree on calling out BS, so here goes - your post is BS.
You're right that it's optimised for lower bitrates, but the advantages remain at higher ones.
This paper shows the objective (Peak Signal-to-Noise ratios) and subjective (blind test) comparison of the two: https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7254155
The graphs in the paper test bitrates on UHD 60Hz up to 38Mbps H.264 and 18Mbps HEVC (broadcast bitrates) and they conclude:
QuoteThis shows that for 74 out of the 80 pairs of test points (or 92.5%) HEVC has a bit rate saving compared with AVC that is greater than or equal to 50%.
And what about higher bitrates?
This paper here shows the relatively quality of the two at higher bitrates - up to about 250Mbps: https://pdfs.semanticscholar.org/b0bc/9342d1031250db0c7e2aabd2eeed51beef2e.pdf
Figure 3 shows that the 50% saving seems to extend up to around 6Mbps H264 (where the equivalent HEVC is about 3Mbps) but after that point there is a knee in the HEVC curve.
Figure 4 shows that the HEVC bitrate required to match the H264 goes above 50% up to the point where 40Mbps H264 requires about 30Mbps HEVC.
Figure 5 shows that the HEVC bitrate required to match the H264 goes back closer to 50% where 250Mbps H264 requires about 120Mbps HEVC, and for another clip 200Mbps H264 requires about 110Mbps HEVC.Fact-checking the internet is fun, but it helps if you actually know the facts when you do it.
1 hour ago, Cliff Totten said:Damn.....this is one HELL of a shoddy CODEC, frame rate and compressed audio $hit list! I pre-ordered this camera today but without the paid VLog and 10bit recording licenses,...this camera out of the box is of very little use to me...not like THIS! (HLG forced into 72mb/s h.265? only....compressed audio?..my God, this is bad)
Yeah, some higher bitrates would have been nice.
I came late to the GH5 party so I don't know what the original specs were or what people thought of them, but they did introduce the 400Mbps and 5K Open Gate modes in firmware updates, so maybe there will be updates for this too? I guess time will tell.
- androidlad and Castorp
-
1
-
1
-
The BMPCC4K (Pocket4K / P4K) is a wonderful camera, but some say it looks too clean, or doesn't have the classic look from the previous BMPCC (OG) or BMMCC (MC) cameras.
Considering that the P4K should have either higher quality levels (ie, more pixels) or sufficient quality but different (bit-depth and colour science) than the others I think we should be able to process the P4K in post to match the classic look (or looks) as the older models. Even if we can't, I'm sure that there are things we can learn in the attempt.
Thanks in advance to @graphicnatured who has volunteered to shoot it. I've shot A/B camera tests before and they're a lot more work than they seem like they should be. Assuming we learn anything, we all owe him a drink - he'll need it!
-
5 hours ago, graphicnatured said:
Our weather is crappy through until about Wednesday so I'll start doing some outdoor shots first since they are the easiest. I'll shoot each in succession and pay attention to shooting things that don't have a ton of change from minute to minute. All on 24-70, and closely match focal length. I'll match the f stops and shoot all with and without IR Cut.
Great stuff!
I'll make a new thread so we stop clogging up this one
Edit: done.
-
14 hours ago, tellure said:
Just theorizing here since I don't know a ton about sensor design but if you have a per-pixel ADC then could you, for example, expose for the blown-out windows in an indoor shot and then boost the gain only on the darker pixels to get the equivalent of a bracketed HDR shot?
I think that's the idea.
If I understand it right, some pixels could be at ISO 100 and others at ISO 25000, so your DR would go through the roof. However, if you had a part of the image that was very dark and those pixels had high ISO, then they'd still be noisy.
-
1 hour ago, webrunner5 said:
Crazy price I know, but that rig I would guess is around a $100,000.00. But that is about as good as it gets for sports, wildlife especially if it was the FS7 mk II. And a person ought to be able to re coop your money back with deprecation and cash inflow in a pretty short time. I think in a way it is a bargain. You could do what few can. You can't make money doing what everyone else can do. And I think it would hold up and be a viable solution for quite a few years. I am sure it would be a good investment if you have talent, or learn to become talented LoL.
I have spent a lot more than that on one piece of heavy equipment used even. It takes money to make money. No free rides. That is a actual case of GAS for sure. That would draw a crowd no doubt. ?
Agree.
They said that it takes two people to set it up but then a single person can operate it, which seems ideal. If you were any good as a wildlife photographer then I'm sure you could put that to good use. Anyone who is travelling the world may very well be going with another person anyway, so that's not as big a deal as it sounds. It's also not as much of an investment as it sounds considering that the image quality is very high and the costs of travelling to the exotic locations, getting guides, etc would be considerable, especially over a multiple year timeframe.
I've been contemplating a trip to Antarctica as part of my bucket list and considering the costs involved I'd definitely be taking some serious camera equipment, especially renting some serious glass. Not suggesting that I'd rent that one (!) but those long lenses really are the tool for the job. Personally though, I'd make sure I took multiple camera bodies as a backup, and having two bodies means that you can always have a short lens on one and a longer one on the other, like the pro event stills shooters do. With my GH5 I can also take advantage of the crop factor to turn more reasonable lenses into hugely long telephoto lenses too, saving considerable weight as well!
-
44 minutes ago, hansel said:
Haha, yeah you are totally right. It is more like “you clean your shoes before entering the house” kind of feeling. I think it is pretty much about triggering the stuff to grow or give the spores not enough substance to life of....
One of the articles said that fungus will grow if there's humidity, the right temperature range, and a source of food. So not only do the spores get in-between the layers of glass inside the lens, but particles of food do as well!
No more of those "throw flour everywhere in slow-motion" shoots people!!
-
39 minutes ago, Orangenz said:
NZ is not part of Australia.
True. I'm not sure why you're pointing that out, but ok
39 minutes ago, Orangenz said:NZ is not part of Australia. Those are new homes, and I agree they are getting bigger as the area of land decreases to almost the same size as the house. The existing average would be closer to 70m2. I've lived, and helped build, in both countries so that's what I'm basing my opinion on.
70m2 is tiny!
I lived in an 80m2 two bedroom (it was built as a granny-flat by a friends parents to retire in, but they moved out because it was too small). It was a little larger than it absolutely had to be, with a small office and an ensuite in addition to the bathroom / laundry, but it was smaller than the two-bedroom unit I used to live in.
I can't imagine that there are enough 1 or 2 bedroom places under that size to offset the staggering number of 3 or 4 bedroom houses that fill the suburban areas of every city and town here in Australia. It would be interesting to see some stats on existing dwellings but I can't imagine the average is that small.
13 minutes ago, fuzzynormal said:The goal my wife and I are striving for is winter in SoCal, spring anywhere shooting a film, summer in Poland, and Japan during autumn.
That sounds like a pretty nice rotation! I'm guessing that you have ties to Poland? it's not normally on many people's 'must-see places" lists
-
8 hours ago, DBounce said:
I can't say the new Panasonic's really interest me. However, I am somewhat intrigued by the upcoming Fujifilm GFX 100. 100MP sensor with no crop in 4k.
It's ok... I know we all love to talk about the crop, but don't worry about not being able to talk about it any more - way before Canon does no-crop 4K we'll be talking about the crop in 8K!
7 hours ago, BTM_Pix said:For Panasonic's higher end stills shooters, its a massive leap to go from a G9 to that anyway, let alone when you can't take your lenses with you so I'm not sure how much brand loyalty they can rely on in this instance.
The alliance as a whole isn't just about these two cameras in isolation though so the landscape might look a lot different in a year's time when Sigma's lenses and new SD Quattro's are available with that mount as well as subsequent Leica models.
I agree, when you can't take your lenses with you then all bets are off and everyone is "in the market" for a new system again. It's a pretty significant point for brands to try and capture and "lock in" customers into their ecosystem. Maybe it's one of those "can you afford to do it? true, but can you afford not to do it?" type things for Panasonic to release their own offerings.
7 hours ago, buggz said:I am a hobbyist enjoying the GH5 and have many lens options already secured, anamorphic as well.
I think this kit will last me for a long time, I would have to see a huge revolutionary development for me to think seriously about upgrading.
Also, the cost of the higher pixel model is currently prohibitive to me.
As a happy GH5 owner I definitely agree.
It will be interesting to see what the GH6 offers. The GH5 has few flaws, but if they offered 4K60 10-bit with HLG and H265 all internally that would be a decent step up. Also if they offered a card slot that could do higher speeds in UHS-I then that would be great too. Being able to use Sandisk 90MB/s cards instead of being forced to buy UHS-II cards would be great.
And of course, if they offered the ability to render prores proxies to one card and H265 to the other that would be wonderful.
Or RAW!!
*ahem*
5 hours ago, scotchtape said:I was quite excited when the rumor mill first started, but as more and more details came out I've almost lost all interest.
I've watched the release of all the FF mirrorless cameras with underwhelm. They seem to be chasing the photographer market and mostly offer only scraps of improvement for video users, at the cost of buying extortionately priced lenses. I've probably lost all touch with the stills photography market, but there doesn't seem to be anything really that interesting about these cameras from a stills perspective. If you had a 5DIII then I'm not sure why you're paying thousands and thousands...
4 hours ago, Oliver Daniel said:If true, it confirms they are leaving the fancier video specs to the GH6 and GH6s.
The earlier comments from Panasonic indicated that video was staying in MFT for now, so that makes sense.
Of course, they might get their system established, some more lenses sorted out, and then start cramming video features into the FF range, we'll see.
Out of a choice of an MFT system limited to 6K sensors that's been around for ages with all kinds of strange glass vs a brand-new FF system with 8K sensor and completely new glass or high-end Leica glass, which system do you think they're going to introduce 8K video into?
-
2 hours ago, webrunner5 said:
For all you well heeled sport, wildlife shooters your dream lens is here.
https://***URL not allowed***/canon-50-1000mm-cine-servo-lens-review-yes-1000mm/
The Canon 50-1000 lens:
The best way to make an FS7 look small, and to make your tripod look like a spaceship!
-
On 6/2/2018 at 3:46 AM, Don Kotlos said:
Did the 4K pocket make the competition so desperate?
Two suggestions for @Andrew Reid:
- Unify all the RAW capable cameras into one sub-forum. Between BM making a splash, Kinefinity, ProRes RAW & RAW output capable cameras, there are a lot to cover.
- Turn the lenses pined thread into a sub-forum. Threads can be lens specific and easier to browse through.
I like the Lenses sub-forum idea. So many lenses, so little time!
-
7 minutes ago, Robert Collins said:
I am yet to really understand what Panasonic will bring to FF mirrorless that isnt offered by Nikon, Canon and Sony especially given the big three already have substantial lens line-ups in one form or other (as well as pdaf).
I am just hoping it isnt 'hubris'...
It's difficult to think of anything they are likely to bring to a FF camera, but it's also difficult to think that the company that made the GH series (and its consistent "they listened to us and gave us everything we wanted" track record) wouldn't do a good job.
When is this thing being launched?
-
Another thought, a colour chart would be useful, and maybe something with predictable movement, like a fan.
That way we can verify that any movement-related differences aren't related to shutter angle, and the colour chart would help in matching colours.
-
3 minutes ago, Nikkor said:
The Benefit of raw is that there is no debayering, so there is no Fake Information taking up bandwith.
In raw it's 14bit per Pixel, not per color channel per pixel.
That's a pretty big advantage. Not only the lack of processing to debayer, but also the lack of processing to get data rates down to be manageable. Still, a 12-bit intermediary codec would be pretty nice and skip having to make proxies.
12-bit files with no rendering proxies would be a big drawcard.
-
Is there any chance it will record Prores internally? Or cineform?
There are some very interesting 12-bit 4:4:4 codecs that don't have crazy data-rates:
- 1080: Cineform 12-bit HD Low 130Mbps
- 1080: Cineform 12-bit HD Medium 160Mbps
- 1080: Cineform 12-bit HD High 195Mbps
- 4K Cineform 12-bit UHD Low 445Mbps
- 4K Cineform 12-bit UHD Medium 515Mbps
- 4K Cineform 12-bit UHD High 630Mbps
For many the ability to record 12-bit 444 internally to an SD card would mean they wouldn't care about RAW, surely?
-
12 minutes ago, BTM_Pix said:
And the difference is definitely clear in the ungraded footage.
Definitely adds to the weight of needing movement in the test.
+1 for movement.
The difference might have been clear, but the x-factor was completely missing. I'm not sitting and looking at the images going "wow those cameras are great", I'm thinking "these make terrible photo cameras"
We also need to be sure to nail focus too. I'd suggest stopping down to mostly eliminate it as a variable. I'm thinking there's mojo in these cameras even without fast lenses, so stopping down isn't going to prevent us from seeing the magic.
-
2 hours ago, Eric Calabros said:
You get lower DR at high ISO cause the amplification needed to increase the brightness of darker areas is too much for already bright areas. With this new pixel design, it could be possible to amplify only those pixels that need to be amplified. So you get near ISO 100 DR with very low noise at say ISO 12800.
Other than fabrication complexities, there is big problem in profiling the output, because it's not going to be linear, I guess.
There might be a clever way to keep it aligned, for example if the gain on each pixel only had a few settings to choose from, but each was 2X or 4X the last. Then you read the value of the pixel and only need to do a bit-shift to the value to align it with the rest of the image. Audio DACs are 16 or 24-bit, and each bit is worth double the previous one, so it wouldn't be that challenging to also make the amplifier for each one have a few "bits", and I'm sure there are some simple circuits that can use the amplifier setting to shift the bits in what the ADC is putting out.
-
1 hour ago, thebrothersthre3 said:
I might be interested, they are saying its DFD auto focus, which I doubt will be good for video. If it has internal 10 bit though, that would be interesting.
I just assumed that it would be at least as good for video as the GH5. If it didn't have 10-bit then I'd be stunned.
-
36 minutes ago, graphicnatured said:
I don't have a gh5, unfortunately, I have access to the three cameras mentioned as well as an Ursa Mini Pro, but that isn't part of this conversation. I have Kino Flos and Astras, bounces, etc.. A adorable eight year-old I can use as a model or can find someone to help, and I live in a pretty beautiful place (San Diego) I'm looking forward to this. If anyone in the forum is from around here and has a GH5 they want to bring, I'm all for it. I think so many are really interested in matching the new look to the old look though. Shooting simultaneously makes sense unless we light it, then I don't think that matters as long as we get the movement to be close. My lenses are all Nikon mount: I have (nikon) 85 1.8, 24-70 2.8, 70-200 vr1, 35mm f2, Sigma 50 1.4, Tokina 11-16, Tokina 100 macro, Helios 44-2, jupiter 9, helios 40, slr magic variable nd 2, hoya ir cut --- Both OG Pocket and Micro have Speedboosters, P4k has Viltrox. I would love to collaborate if anyone in Socal wants in on this.
Awesome!
These camera tests are a lot more work than they seem, so including whatever other cameras (and helpers to come with them!) is a lot easier than trying to repeat the test with those missing cameras, or to try and shoot something equivalent under different conditions.
I'm totally fine for someone on social to get in on this too, I'm after knowledge not fame
37 minutes ago, graphicnatured said:I think so many are really interested in matching the new look to the old look though.
This is my primary drive, so exact colours and whatever don't matter as much to me.
When you look at old BMPCC footage that "x-factor" is apparent, so I don't think we need identical movements. It's a "we'll know it when we see it" kind of thing, and that's why I think so many are interested in it but no-one can really describe what "it" is. Thus why we invent new words like Mojo
17 minutes ago, BTM_Pix said:I don't think using the exact same lens on each camera is optional and its where a lot of these comparison videos on YouTube are doomed before they start.
My vote is for matching the "soul" of the cameras so I don't care if the movement is identical, so now we just have to pick the most cinematic lens... lol.
....
So, artificial lighting, adorable 8-year old, Nikkor 24-70 f2.8, adapters, cameras and a tripod.
In terms of the setup, how do we shoot it so that the character of each camera is most emphasised? I'm thinking more dramatic lighting and lots of movement perhaps? Should we use mixed lighting?
What really makes the BM cameras sing? -
2 hours ago, thebrothersthre3 said:
Best not to argue with the director ?
True!
Sometimes for the wrong reasons, but still best not.
-
35 minutes ago, graphicnatured said:
I can shoot the same scene with a P4k, a OG Pocket and a Micro Cinema if you want. Tell me what you want shot and I'll do it and share the raw files.
Awesome!
I'm thinking we should do this carefully and properly, unlike all the other BS camera comparisons we all justifiably hang shit on.
I think there's two approaches
- Shoot all of cameras simultaneously, which would mean no differences in what they're filming and we can sync them up for direct comparison. The downside to this is that you can't use the same lens on each of them, unless you can borrow some duplicates from somewhere, and the angles would be slightly different, but with a long focal length this is passable.
- Shoot them sequentially and have a completely controlled scene, so 100% artificial lighting (already warmed up), etc. The upside is we can get the same angle and lens, the downside is that any human movement or poses won't be identical.
In terms of what to shoot, we absolutely need movement, we absolutely need skin tones, and we absolutely need to have a range of highlights/shadows and different colours.
I'm not going to be able to match the colours exactly like @Sage has done with the GHa, I'm more interested in making the footage from the P4K look like it was shot with either of the other two cameras, ie, to be believable. I'm with @DBounce that motion plays a critical role here. I'm also thinking that resolution and softness are in the mix somehow too.
And if we can get a GH5 in there too that would be absolutely brilliant!
There really is something special about the BMPCC and some of these old cameras, and maybe @webrunner5 is right, but maybe not, and i'm not sure if anyone has given it a serious go or not.
What does everyone think about how to shoot it?
I'm going to need guidance from everyone else on this project for sure
Fstoppers moved to Puerto Rico. Have you had the desire to move?
In: Cameras
Posted
Very nice!
We already know - she's becoming a rapper!
Step One: get that million dollar contract
Step Two: we'll all be waiting for your hot track
???