The Blackmagic LF Camera - Is Full Frame ever coming?

I know.... The tower, shot with a model on a table-top. It was done a few times after Scorsese as well, but I can't remember who now. BTW that Vertigo shot looks a whole lot wider than 50mm, but since it's VistaVision, maybe it is....

I know.... The tower, shot with a model on a table-top.

Actually that's the best remembered use but it first appeared in the opening moments when the policeman falls to his death after trying to save James Stewart and then later when he's trying to show Barbara Bel Geddes he's coping with vertigo. The most famous homage is Jaws, when Roy Schieder realises it's bad You might be right about the focal length (I think there was a 35mm too) but it certainly was common to have regular deep DOF in Hollywood's heyday.

Ah, yes, the Zolly Effect. Dolly and Zoom at the same time, but in opposite directions. Ever since I got parfocal zooms I've wanted to give it a try and still haven't. Might be time to do it.

I always forget that technically Vertigo was shot on VistaVision. Especially due to the dolly zoom effect I always think it was film running vertically though the camera for 1.85:1 cropped off the 4-perf 35mm film. I have Vertigo in 4K HDR through iTunes, might be time for another watch now. Same with Goodfellas since it too is 4K HDR, and I also own it.

There are a lot of great uses of the dolly zoom. The best ones are the more subtle ones. But it also can be used for disorienting effect like in Nightmare on Elm Street. Spielberg used it in Jaws, and Scorsese used it in Raging Bull to great effect. I also love the use in Wall Street from Oliver Stone.

Either way, it's how you use the tools to tell the story that matter. And, all the movies I've named used this one in camera effect really well at various points in the stories they were telling.

AFAIK, in film literature it's called the Vertigo effect.

Vertigo Effect, Dolly Zoom, Zolly... it goes by many names.

I always forget that technically Vertigo was shot on VistaVision. Especially due to the dolly zoom effect I always think it was film running vertically though the camera for 1.85:1 cropped off the 4-perf 35mm film. I have Vertigo in 4K HDR through iTunes, might be time for another watch now. Same with Goodfellas since it too is 4K HDR, and I also own it.

What-is-VistaVision-VistaVision-aspect-ratio.jpg (59.19 KiB) Viewed 13601 times

We might see a lot of this kind of advertising in the coming weeks and months It's well worth a revisit Tim, especially with the painstaking restoration of Harris/Katz in the 90's (Anyone who thinks 35mm film is the best archival for digital should reflect on the condition they found it after only some 30 years or so). The other thing that strikes me about the 2 stills I posted is that Vegas seems much closer to what we now perceive as cinematic in terms of colour, yet it's almost monochrome compared to Vertigo's Technicolor and that's (Vegas) what I see with a lot of film emulation luts used in grading too. Vertigo has deep saturated reds and pink skin tones, but also strong blues and green too. Perhaps not a popular look now but one I like for sure.

filmpolice-vertigo-film-still.jpg (504.09 KiB) Viewed 13574 times

Another still from the film to show that FF might not always mean or demand shallow DOF and a detailed shot well composed, won't confuse the audience as to where to look.

Could well be. It it was probably a studio setup anyway, with the exterior being a rear projection.

We might see a lot of this kind of advertising in the coming weeks and months It's well worth a revisit Tim, especially with the painstaking restoration of Harris/Katz in the 90's (Anyone who thinks 35mm film is the best archival for digital should reflect on the condition they found it after only some 30 years or so). The other thing that strikes me about the 2 stills I posted is that Vegas seems much closer to what we now perceive as cinematic in terms of colour, yet it's almost monochrome compared to Vertigo's Technicolor and that's (Vegas) what I see with a lot of film emulation luts used in grading too. Vertigo has deep saturated reds and pink skin tones, but also strong blues and green too. Perhaps not a popular look now but one I like for sure.

I remember watching North by Northwest at SMPTE 2010 Conference in LA projected on the (at the time) new Christie 4K projector. They spoke about scanning it in 6K for restoration and preparation for the 4K projection. Ironically that's the one Hitchcock film I haven't seen updated to 4K HDR on iTunes in all the time that the others have been updated. Considering I saw it in 4K over 11 years ago now that's a pity. I own it too on iTunes.

The Star Wars Special Editions also came with discussion about how messed up the negatives were after 20-years. I remember reading about Lucas talking about that, or seeing it in a documentary. Either way, film preservation is key. The Godfather's recent 50 year restoration for 4K HDR also brought news about the condition of the film negative being not as great.

Another thing to remember is post 1953 films that were in color were heavily saturated in order to showcase the colors since that's when films switched from Black & White to Color permanently in order to combat television. That's also why we went from full screen to widescreen. Stereo 3D also had it's small time then, with Hitchcock shooting Dial M for Murder (1954) in 3D. But in the Widescreen vs 3D battle Widescreen won.

Another thing to remember is post 1953 films that were in color were heavily saturated in order to showcase the colors since that's when films switched from Black & White to Color permanently in order to combat television. That's also why we went from full screen to widescreen. Stereo 3D also had it's small time then, with Hitchcock shooting Dial M for Murder (1954) in 3D. But in the Widescreen vs 3D battle Widescreen won.

I think it was more to do with Technicolor imbibition process, which produced heavily saturated colours since the late 30's, Wizard Of Oz, for example. By the 50's, 3 strip cameras were no longer used, the negative being single pack Eastman, but the imbibition process of directly printing dyes onto the release prints remained. And there were plenty of Hollywood B+W features made after 1953, such as perhaps my favourite film of all time, Lonely Are The Brave, 1962.

I have the UHD HDR 4K Blu-ray of Vertigo, I think it was remastered in Dolby Vision, and to be honest, I don't feel it's entirely successful - it looks muted in comparison to the Rec. 709 Blu-ray, no matter how I fiddle with the DV 4K player or TV. Same with Rear Window too. I know the TV and player are good, since My Fair Lady (hate musicals but love the cinematography on this and West Side Story, the original ) looks stellar remastered, so do a lot of well graded DV Netflix films and series.

Maybe it is just me, but to me that does not look like the vertigo effect, but that background and front ground are separate.

no, it was a real shot with vertigo effect

Screenshot 2022-04-08 at 13.51.44.jpg (246.63 KiB) Viewed 13454 times

Responding here... My understanding is that the Yedlin response is mostly correct in terms of matching a look, however, I wanted to address some of the comments starting with sensor size not affecting circles of confusion.

My understanding is that every CoC calculation takes sensor size into account to determine the maximum blur circles that can be achieved. It's in the formula. There's other things in the formula, but the size of the sensor or film is in there.

Now if you want to argue people can't see that, ok, but it should not be stated that it doesn't affect CoC because... It does ... So says math. Maybe certain people can see this under certain conditions? Maybe it's just they see a difference in how gear is used to achieve certain shots? I dunno, but math is math.

Secondly I would agree about there being more cinema lenses for s35 currently, but things change. I recall similar discussions about film v digital, 4k v HD acquisition... List goes on. I don't think looking backwards is really a good way of looking toward the future but I could be wrong. I think trends are important to watch for and considering the amount of inexpensive glass already on market now and how use of larger sensors cameras seem to have picked up recently, much like the older digital v film arguments, I think this argument against larger sensors is only going to get weaker.

There's a lot of glass out there being used on both s35 and 135 regardless of if it's a cinema housing or not. There's a hardly a lack of good glass there. I mostly sit at f2.8, so personally I'm pretty happy with getting f2.8 on one of many good 24-70's offered instead of using shorter, faster prime lenses or the more limited zoom range f1.8 zooms I used to use to achieve a similar look.

I also wanted to point out that it's a little annoying to read discussions like this because of the assumptions made about people... like saying they're just using certain cameras because they're paid to do it or they got duped by marketing. You don't know that and the accusations are silly and not the kind of thing I expect professionals to say about others in the community.

You don't care about LF or 135 formats, fine, but you don't have to go get personal about it questioning other's motives, intelligence, knowledge, or integrity. Good luck to all of you.

Responding here... My understanding is that the Yedlin response is mostly correct in terms of matching a look, however, I wanted to address some of the comments starting with sensor size not affecting circles of confusion.

My understanding is that every CoC calculation takes sensor size into account to determine the maximum blur circles that can be achieved. It's in the formula. There's other things in the formula, but the size of the sensor or film is in there.

The CoC is different for different sensor sizes, but here's what happens when two sensors -- full frame, and Canon APS (crop factor 1.52) are compared, using the crop factor to create identical (we hope) images. Note here the 50mm FF is reduced to 33 and the f4 stop to f2.63 on the APS sensor. The small differences noted in the DOF readouts (last line) are rounding errors:

ff..JPG (181.01 KiB) Viewed 13367 times

aps.JPG (185.46 KiB) Viewed 13367 times

I think trends are important to watch for and considering the amount of inexpensive glass already on market now and how use of larger sensors cameras seem to have picked up recently, much like the older digital v film arguments, I think this argument against larger sensors is only going to get weaker.

The Sigma FP is really inexpensive right now and BRAW through the Video Assist really helps wrangle the color in Resolve to a usable place for me. It's my BMD LF for now, only because it's so tiny. I don't think it represents a unique visual look apart from the Pocket 6k on the whole, there are some extreme examples like the 14mm 1.8 lens mentioned above. I think you'd need a 9 or 10mm F1? But at this point it's interesting that there are a lot of MFT lenses for an even smaller format that are 0.95 and 0.85.

I think the question that's being debated is whether or not it represents a distinctive visual language or style that can't be replicated with a smaller format.

As John P's showing above as long as you have a lens that can match on the smaller sensor, the visual language or 'look' remains the same.

Responding here... My understanding is that the Yedlin response is mostly correct in terms of matching a look, however, I wanted to address some of the comments starting with sensor size not affecting circles of confusion.

My understanding is that every CoC calculation takes sensor size into account to determine the maximum blur circles that can be achieved. It's in the formula. There's other things in the formula, but the size of the sensor or film is in there.

Now if you want to argue people can't see that, ok, but it should not be stated that it doesn't affect CoC because... It does ... So says math. Maybe certain people can see this under certain conditions? Maybe it's just they see a difference in how gear is used to achieve certain shots? I dunno, but math is math.

Diffraction limits related to COC, diameter of an airy disk are roughly the same between FF and S35, an extreme case being the U12K 2.2 um pixel size, diffraction limits start to kick in between about F/4.0 and F/5.6, and this is about the same as FF would be for the same pixel size, the main difference being there are no 12K FF sensors and the pixel sizes are much larger so diffraction limits are higher. Taking this in relation to Yedlin's comments, I don't think this difference is very significant in terms of matching a look and thus Yedlin is correct in my opinion.

https://www.cambridgeincolour.com/tutor ... calculator

The Sigma FP is really inexpensive right now and BRAW through the Video Assist really helps wrangle the color in Resolve to a usable place for me. It's my BMD LF for now, only because it's so tiny.

I've been thinking about the Sigma FP as a cheap LF cine cam, which I think is very useful on gimbals and conspicuous enough for street work like in LA. Since you can do BRAW and/or Prores RAW with an external recorder, why not? Even the 12-bit DNG is pretty decent, I think, and should be able to match with the UM4.6K or Pockets. I have to do more research but at $1600 including a 45mm 2.8 lens, it's not a bad deal as an LF to have. It's firmware and menu system seem to be well designed too. With the PL adapter, that gives options.

Which is obviously not what I was referring to, so...

And yet that is the particular quote, in full, you referenced and so I thought you meant it all untrue. My apologies Rakesh, we are at cross purposes then.

Yeah, it's easy to be mislead by a wayward quote in forums.

BMD's forum emoticons aren't silly enough.

Responding here... My understanding is that the Yedlin response is mostly correct in terms of matching a look, however, I wanted to address some of the comments starting with sensor size not affecting circles of confusion.

Uh, no -- I said that the PHOTOSITE size doesn't affect the circle of confusion. There's a misperception that a larger sensor equates to larger photosites which is also false, because photosite size is a function of sensor design and not of sensor size. The AlevIII sensor's photosites are all the same size regardless of format, because all variants of the AlevIII sensor intentionally use exactly the same sensor design.

My understanding is that every CoC calculation takes sensor size into account to determine the maximum blur circles that can be achieved. It's in the formula. There's other things in the formula, but the size of the sensor or film is in there.

In reality however, the CoC is determined by the optics -- mainly depth of field, magnification, and focus point. Change nothing but the sensor, and the CoC will remain the same, period.

What the sensor size DOES affect is how much you blow up the image AFTER you capture it, which therefore influences how large a CoC is acceptable to the viewer, since blowing up an image obviously will also blow up the CoC.

That is why an image can look tack sharp on a small camera screen, but blurry on a computer.

Now if you want to argue people can't see that, ok, but it should not be stated that it doesn't affect CoC because... It does ... So says math. Maybe certain people can see this under certain conditions? Maybe it's just they see a difference in how gear is used to achieve certain shots? I dunno, but math is math.

And the math shows that the CoC has no relationship to the size of the sensor or the photosites on it. The only thing that the CoC measures (and it IS a measurement, there's absolutely nothing abstract about it) is how large an image you get on the imaging plane from a single point source of light. At the plane of focus, it's a point. As you move away from the plane of focus, it gets bigger. At some point, it gets big enough to look perceptibly out of focus.

That's quite literally it. There's nothing more to it. The size of the sensor has nothing to do with, the size of the photosites has nothing to do with it, and it's easy to prove, since the light passing through the lens does the same thing no matter what you put behind the lens. A box, your eyeball, your cat... the light does what the light does.

You don't care about LF or 135 formats, fine, but you don't have to go get personal about it questioning other's motives, intelligence, knowledge, or integrity. Good luck to all of you.

When they're clearly giving disinformation in order to promote themselves or a product, well...

The reality is that no one here is saying that they don't care about larger formats. They're just pointing out that that nearly everything everyone has said about why you SHOULD spend the extra money to move up to larger formats is mostly false.

The Sigma FP is really inexpensive right now and BRAW through the Video Assist really helps wrangle the color in Resolve to a usable place for me. It's my BMD LF for now, only because it's so tiny.

I've been thinking about the Sigma FP as a cheap LF cine cam, which I think is very useful on gimbals and conspicuous enough for street work like in LA. Since you can do BRAW and/or Prores RAW with an external recorder, why not? Even the 12-bit DNG is pretty decent, I think, and should be able to match with the UM4.6K or Pockets. I have to do more research but at $1600 including a 45mm 2.8 lens, it's not a bad deal as an LF to have. It's firmware and menu system seem to be well designed too. With the PL adapter, that gives options.

The Sigma fp also has great color rendition. The VA approach does add some functionality though, like support for DCI frame rates and aspect ratios, that enable it to match more easily with other cinema cameras. I don't like using cDNG on it, but only because it's so unwieldy. Look wise, it's very good, in spite of its linear encoding. Sigma did a LOT right with that camera.

I've been thinking about the Sigma FP as a cheap LF cine cam, which I think is very useful on gimbals and conspicuous enough for street work like in LA. Since you can do BRAW and/or Prores RAW with an external recorder, why not? Even the 12-bit DNG is pretty decent, I think, and should be able to match with the UM4.6K or Pockets. I have to do more research but at $1600 including a 45mm 2.8 lens, it's not a bad deal as an LF to have. It's firmware and menu system seem to be well designed too. With the PL adapter, that gives options.

I think you should rent the FP before you buy it. I rented it for a week of tests a couple of years ago during all the initial hype, and was completely disappointed. The camera can give a decent image with the proper conditions, but man, does it have poor dynamic range. The image is nice, but not that special, nobody is going to mistake the image with an Alexa. It's also god awful to use in an ergonomic sense. The autofocus is also bad. I don't think it's worth the trouble in this day and age to shoot with this camera.

I've pretty much given up on finding that magical camera that's inexpensive but has an Alexa-like image. I think the Pockets are the closet, then followed by the Nikon Z6/Z7 and then the Sony A7SIII/FX3. I own all those cameras and although I like them, I'm never fully happy because not matter what, the image they render doesn't look like an Alexa.

I've been thinking about the Sigma FP as a cheap LF cine cam, which I think is very useful on gimbals and conspicuous enough for street work like in LA. Since you can do BRAW and/or Prores RAW with an external recorder, why not? Even the 12-bit DNG is pretty decent, I think, and should be able to match with the UM4.6K or Pockets. I have to do more research but at $1600 including a 45mm 2.8 lens, it's not a bad deal as an LF to have. It's firmware and menu system seem to be well designed too. With the PL adapter, that gives options.

I think you should rent the FP before you buy it. I rented it for a week of tests a couple of years ago during all the initial hype, and was completely disappointed. The camera can give a decent image with the proper conditions, but man, does it have poor dynamic range. The image is nice, but not that special, nobody is going to mistake the image with an Alexa. It's also god awful to use in an ergonomic sense. The autofocus is also bad. I don't think it's worth the trouble in this day and age to shoot with this camera.

I've pretty much given up on finding that magical camera that's inexpensive but has an Alexa-like image. I think the Pockets are the closet, then followed by the Nikon Z6/Z7 (w/Ninja V ProRes) and then the Sony A7SIII/FX3. I own all those cameras and although I like them, I'm never fully happy because not matter what, the image they render doesn't look like an Alexa.

I think you should rent the FP before you buy it. I rented it for a week of tests a couple of years ago during all the initial hype, and was completely disappointed. The camera can give a decent image with the proper conditions, but man, does it have poor dynamic range. The image is nice, but not that special, nobody is going to mistake the image with an Alexa. It's also god awful to use in an ergonomic sense. The autofocus is also bad. I don't think it's worth the trouble in this day and age to shoot with this camera.

I've pretty much given up on finding that magical camera that's inexpensive but has an Alexa-like image. I think the Pockets are the closet, then followed by the Nikon Z6/Z7 (w/Ninja V ProRes) and then the Sony A7SIII/FX3. I own all those cameras and although I like them, I'm never fully happy because not matter what, the image they render doesn't look like an Alexa.

Good advice, thanks! I don't think you'll ever find anything that will look like the Alexa. You can come close to it but that's as far as you go. Renting is pretty much most of my direction these days, aside from the BMD cameras and Panasonics that I already owned. I sold my Red Komodo recently and am waiting to see what BMD will show up at NAB. I was just thinking about the FP for run-n-gun gimbal and drone c-cam footage, being a FF sensor and RAW abilities... but I think right now I might revert my thoughts for it.

I think you should rent the FP before you buy it. I rented it for a week of tests a couple of years ago during all the initial hype, and was completely disappointed. The camera can give a decent image with the proper conditions, but man, does it have poor dynamic range. The image is nice, but not that special, nobody is going to mistake the image with an Alexa. It's also god awful to use in an ergonomic sense. The autofocus is also bad. I don't think it's worth the trouble in this day and age to shoot with this camera.

I've pretty much given up on finding that magical camera that's inexpensive but has an Alexa-like image. I think the Pockets are the closet, then followed by the Nikon Z6/Z7 (w/Ninja V ProRes) and then the Sony A7SIII/FX3. I own all those cameras and although I like them, I'm never fully happy because not matter what, the image they render doesn't look like an Alexa.

Good advice, thanks! I don't think you'll ever find anything that will look like the Alexa. You can come close to it but that's as far as you go. Renting is pretty much most of my direction these days, aside from the BMD cameras and Panasonics that I already owned. I sold my Red Komodo recently and am waiting to see what BMD will show up at NAB. I was just thinking about the FP for run-n-gun gimbal and drone c-cam footage, being a FF sensor and RAW abilities... but I think right now I might revert my thoughts for it.

(Anyone who thinks 35mm film is the best archival for digital should reflect on the condition they found it after only some 30 years or so).

All depends on how it's stored. A lot of film restorations are from piecemeal badly stored negatives. Like any archiving, you have to actually store them with care, not just in the bottom of a closet somewhere.

Separation masters are the best, but not a lot pay for that.

And yet...we forget film looks better today than it probably did when originally screened.

Film re-matering means we can watch the Godfather and it looks better than it probably did when it was released BECAUSE we have the original negative.

I bet we won't be able to say that about episode 1 of Star Wars in 30 or 40 years....

https://www.moviemaker.com/film-preserv ... efinitely/

JB

I've pretty much given up on finding that magical camera that's inexpensive but has an Alexa-like image.

G2 / 12K.

I shoot them alongside Alexa all the time for this reason. Take any of my recent shows (Aside form Morning Show) and generally 20-30% of the final shots are G2 / 12K. When people ask, I usually say, 90% of an Alexa for 10% of the price....

Sigma FP is a huge disappointment, and I think Sigma also screwed up the BRAW part as well....

JB

Responding here... My understanding is that the Yedlin response is mostly correct in terms of matching a look, however, I wanted to address some of the comments starting with sensor size not affecting circles of confusion.

Uh, no -- I said that the PHOTOSITE size doesn't affect the circle of confusion. There's a misperception that a larger sensor equates to larger photosites which is also false, because photosite size is a function of sensor design and not of sensor size. The AlevIII sensor's photosites are all the same size regardless of format, because all variants of the AlevIII sensor intentionally use exactly the same sensor design.

My understanding is that every CoC calculation takes sensor size into account to determine the maximum blur circles that can be achieved. It's in the formula. There's other things in the formula, but the size of the sensor or film is in there.

In reality however, the CoC is determined by the optics -- mainly depth of field, magnification, and focus point. Change nothing but the sensor, and the CoC will remain the same, period.

What the sensor size DOES affect is how much you blow up the image AFTER you capture it, which therefore influences how large a CoC is acceptable to the viewer, since blowing up an image obviously will also blow up the CoC.

That is why an image can look tack sharp on a small camera screen, but blurry on a computer.

Now if you want to argue people can't see that, ok, but it should not be stated that it doesn't affect CoC because... It does ... So says math. Maybe certain people can see this under certain conditions? Maybe it's just they see a difference in how gear is used to achieve certain shots? I dunno, but math is math.

And the math shows that the CoC has no relationship to the size of the sensor or the photosites on it. The only thing that the CoC measures (and it IS a measurement, there's absolutely nothing abstract about it) is how large an image you get on the imaging plane from a single point source of light. At the plane of focus, it's a point. As you move away from the plane of focus, it gets bigger. At some point, it gets big enough to look perceptibly out of focus.

That's quite literally it. There's nothing more to it. The size of the sensor has nothing to do with, the size of the photosites has nothing to do with it, and it's easy to prove, since the light passing through the lens does the same thing no matter what you put behind the lens. A box, your eyeball, your cat... the light does what the light does.

You don't care about LF or 135 formats, fine, but you don't have to go get personal about it questioning other's motives, intelligence, knowledge, or integrity. Good luck to all of you.

When they're clearly giving disinformation in order to promote themselves or a product, well...


The reality is that no one here is saying that they don't care about larger formats. They're just pointing out that that nearly everything everyone has said about why you SHOULD spend the extra money to move up to larger formats is mostly false.

- Point taken on photosites, thanks for clarifying.

- In reality, the size of the sensor/film is used to determine CoC and the CoC DOES in fact change with all other things equal but the size of the sensor/film, which I think you just conceded by saying it is influenced by the sensor/film size. You of course blow the image up for presentation, don't you?

https://www.photopills.com/calculators/coc
https://en.wikipedia.org/wiki/Circle_of ... 20of%20arc.

Notice a pattern in the wiki chart and increasing film/Sensor size? Notice when you leave everything the same and only change the sensor size of the camera for the calculator it changes the CoC?

Explanation?

I am not a mathematician and maybe you are. I find it odd that if the size of the sensor/film has nothing to do with CoC that the results show the opposite. I am willing to accept the idea that, in the grand scheme of things, that one or few will notice this visually when you account for "the look" with your optics, however, saying it has NOTHING to do with a math calculation that clearly shows it does... makes no sense to me. If you know for a fact its got nothing to do with it, I think you ought to write a paper and send it to the mathematicians to change the formula.

- Who's "they?" and how do you know "they" are doing this? Isn't this just you speculating and then personally attacking their character/integrity?

- The reality here is that I am not arguing against clarifying things and pointing out flaws in arguments. What I am arguing against is the clearly personal attacks being made on people's character/integrity or intelligence. It is possible to have discourse on this and not take it personally or attack others personally who are not here to defend themselves.

What the sensor size DOES affect is how much you blow up the image AFTER you capture it, which therefore influences how large a CoC is acceptable to the viewer, since blowing up an image obviously will also blow up the CoC.

That is why an image can look tack sharp on a small camera screen, but blurry on a computer.

Yes.

It's a subjective number, because it is affected (subjectively) by the display size.

In the film days, when focus pullers used DOF calculators on a slide rule, you used different CoC depending on how "precise" you wanted the numbers to be.

Of course focus isn't "in" or "out" it changes from IN to OUT and the COC is helping you determine WHERE it's IN or OUT..

The three lines in the image of the DOF calculator are the different COC the focus puller could assess..

https://www.fdtimes.com/2018/03/07/dept ... -repealed/

JB

What the sensor size DOES affect is how much you blow up the image AFTER you capture it, which therefore influences how large a CoC is acceptable to the viewer, since blowing up an image obviously will also blow up the CoC.

That is why an image can look tack sharp on a small camera screen, but blurry on a computer.

Yes.

It's a subjective number, because it is affected (subjectively) by the display size.

In the film days, when focus pullers used DOF calculators on a slide rule, you used different CoC depending on how "precise" you wanted the numbers to be.

Of course focus isn't "in" or "out" it changes from IN to OUT and the COC is helping you determine WHERE it's IN or OUT..

The three lines in the image of the DOF calculator are the different COC the focus puller could assess..

https://www.fdtimes.com/2018/03/07/dept ... -repealed/

JB

If it were argued as Yedlin points out, in terms of Blur Circles, with a distinct explanation as to why Circles of Confusion's formula is not necessary for the purposes of lens blur in cinema, then I am good as I agree since it is demonstrable.

Just saying that CoC doesn't account for film back/sensor and it doesn't change depending on the film back/sensor... when in fact the formula and results do.... is just demonstrably wrong.

It seems more realistic and a better argument in my mind to just say, "CoC may differ according to the math, but its irrelevant for our purposes here due to Lens Blur/Blur Circles formula and demonstrations that prove the case."

Responding here... My understanding is that the Yedlin response is mostly correct in terms of matching a look, however, I wanted to address some of the comments starting with sensor size not affecting circles of confusion.

My understanding is that every CoC calculation takes sensor size into account to determine the maximum blur circles that can be achieved. It's in the formula. There's other things in the formula, but the size of the sensor or film is in there.

The CoC is different for different sensor sizes, but here's what happens when two sensors -- full frame, and Canon APS (crop factor 1.52) are compared, using the crop factor to create identical (we hope) images. Note here the 50mm FF is reduced to 33 and the f4 stop to f2.63 on the APS sensor. The small differences noted in the DOF readouts (last line) are rounding errors:

ff..JPG

aps.JPG

Please understand I am simply talking about the math and what those numbers say about CoC when using that actual formula. I get it you can match the background blur regardless of what those numbers say, I've done it and taught it, but the formula is there with film back/sensor size in the thing and you get different results for CoC based on changing it and leaving all other variables constant. I didn't make that up, it is what it is.

If you want to use a different formula that makes more sense for lens blur, which I do (Yedlin's Blur Circles instead of CoC explanation for example), then do it and maybe argue that's the better way to demonstrate or calculate matching looks than worrying about the CoC calculation.

- In reality, the size of the sensor/film is used to determine CoC and the CoC DOES in fact change with all other things equal but the size of the sensor/film, which I think you just conceded by saying it is influenced by the sensor/film size. You of course blow the image up for presentation, don't you?

No, I didn't. You're conflating the blur circles that Steve Yedlin is referring to with circles of confusion. The former are related to the latter, but they are not the same. That's where people get so confused; they don't account for the reality that sensor and the final image aren't the same thing.

And as Steve Yedlin also pointed out, one of the often ignored factors is that larger coverage lenses are typically also slower than smaller; that's why my Schnieder Super-Symmer was a blazing fast f/4 for 4x5, and my 300mm Fujinon-C, one of the fastest on the market, was a whopping f/8.

Notice a pattern in the wiki chart and increasing film/Sensor size? Notice when you leave everything the same and only change the sensor size of the camera for the calculator it changes the CoC?

Explanation?

I studied physics, including optics. The sensor size doesn't have any effect on how light passes through the lens, or where a photon is relative to the lens axis when it reaches the image plane. Since that's the case, there's no possibility that the sensor size can affect anything; implying otherwise is the same as implying that changing the size of the sensor changes the optics of the lens or the phyics of light, both of which are clearly, patently false -- as should be glaringly obvious to anyone who stops to take a minute to think about it.

I am not a mathematician and maybe you are.

I'm not a mathematician, I'm a physicist.

I find it odd that if the size of the sensor/film has nothing to do with CoC that the results show the opposite.

You shouldn't since the CoC is a characteristic of a lens, its magnification, refraction, aperture, and focus plane. You're basically implying that cropping a sensor would change the optics, which is obviously not the case, so...

- Who's "they?" and how do you know "they" are doing this? Isn't this just you speculating and then personally attacking their character/integrity?

Just look around; it's everywhere. There's a discussion about a "masterclass" whose basis is a clear fallacy -- i.e. the claim that a large format sensor provides more room for actors to move around the frame than a smaller format sensor does.

That's false.

- The reality here is that I am not arguing against clarifying things and pointing out flaws in arguments. What I am arguing against is the clearly personal attacks being made on people's character/integrity or intelligence. It is possible to have discourse on this and not take it personally or attack others personally who are not here to defend themselves.

That's why no one was referring to them by name. But that said, such marketing claims are all over the place, and they are the primary driver behind people hoping for a larger format sensor from Black Magic.

And no one is saying that there's anything WRONG with a larger sensor, only that the "look" isn't among them... and why.

The cost for larger sensor cameras and the lenses required to cover them however ought to make it pretty clear why people who have them want to market them as having a special look that can't be matched with a Super35 sensor... and no matter how much anyone tries to claim otherwise, that really isn't true.

The cost for larger sensor cameras and the lenses required to cover them however ought to make it pretty clear why people who have them want to market them as having a special look that can't be matched with a Super35 sensor... and no matter how much anyone tries to claim otherwise, that really isn't true.

+1000 to this

The cost for larger sensor cameras and the lenses required to cover them however ought to make it pretty clear why people who have them want to market them as having a special look that can't be matched with a Super35 sensor... and no matter how much anyone tries to claim otherwise, that really isn't true.

+1000 to this

Adding another +1000 to this too. It’s all claims, no truths.

There is the theories and there is the reality. If size would not matter we would have phones with interchangeable lens system, and would not need anything else, except maybe heavy gage to make the dynamics of operating it more as we are used to.

There is the iPhone 13pro that can create a faux 'cinematic' look, complete with shallow DOF and many many people will not miss, with it, interchangeable lenses. Broadcasting uses big cameras that can be operated as 'we are used to' but still have 2/3" chips because it is more practical for live focussing and huge parfocal zoom reach.

Until the Canon 5DMKII, people were making films with small chips and fixed lenses. Then accidentally someone found this camera could make video images that 'looked like film', even Canon didn't realise that and then it took off. Consumer demand drove the DSLR, for better video codecs and resolution, and later the more traditional prosumer larger body form, large format revolution.

Everyone suddenly needed shallow DOF, they must have it, even the broadcasters, with sport now, preferably with a set of primes, that look and operate like cine primes, because that is more 'cinematic', even though perhaps unnecessarily impractical, outside of a full crew. When this was never a factor with real feature film making - they just happened to be shooting on 35mm film, because that was the main format and hence the frame size and they didn't have zooms early on. They only shot on larger formats for a bigger better negative, that could blow up better to a bigger screen - not because of any special mystical qualities in the one format or the other - as Yedlin, others here and I have tried to show with film stills.

Full frame demand has come from still cameras, via that 5DMKII, not because VistaVision is almost identical. We have already reached equivalence, in the Super 35 sensor format, with 95% of what has been shot in the cinema since Charlie Chaplin first flickered on the silver screen, over 100 years ago. Yes there are theories and there is reality.

Consumers are demanding FF, it is not being foisted on us by manufacturers, but it is cheap for them and us, to churn out endless quasi still DSLRs that take video. And so you can be rest assured and needn't worry, FF and a probably a box camera are coming at some point, I would think, from the likes of BMD, despite (or should that be in spite?) what any of us say here.

"They only shot on larger formats for a bigger better negative" That is still very much true today, now it is just about pixel quality. Also dust etc. is less visible on big format.

But yes, I do not think there is many actually needing FF, would be just nice to fully utilise the FF lenses many of us have without speed booster etc. but then again those still photography lenses are not that great for video...

Again Kim, this has been shown not necessarily to be true: the Alexa LF is 4.5K, the Alexa 65, 'only' 6.5K - they are overwhelmingly the most represented cameras currently shooting high budget features. Pixels and silver halide film grain are different things. Dust being visible is about cleanliness, large f stops and very wide close focusing, not the format - I've never seen dust on 1/3" or 2/3" chips.

No one can argue with your 2nd point, it makes sense if those are lenses you already have but that is not the same as others claim, FF has a 'better' more 'cinematic' quality than Super 35.

It is just a fact that with same technology a larger pixel will get more light and result less noise.

Yes it is but that maybe at the expense of resolution - you can't have both big pixels and very high rez. The Alexas chose the former at 8 microns and therefore less of them.

maxresdefault.jpg (72.62 KiB) Viewed 12943 times

And speaking of uTube, look at this picture too, of this guy who has been cruelly lampooned there, and compare it to the interior deep DOF Vertigo still I posted early. They are both from roughly the same camera format size. I would argue 24fps is far more the signature of the 'film look' than DOF could ever be.

With FF you can have both, big pixels and plenty of them.

You can have bigger pixels than the equivalent pixel count at a lower sensor yes, but not necessarily the biggest and you might not have noise problems already with properly exposed Super 35. With respect Kim - and that always sounds patronising - I mean it sincerely you are trying to justify the advantages of FF because it's going to work for you but that is not the same thing as the main reason being propounded out there - that FF is more 'cinematic' than Super 35 and has inherent mystical qualities that you simply can't get with a smaller sensor.

30p doesn't look anything like 24p to me. 50/60p no matter the sensor size looks like 'video' to me. I love 24/25p, it has the cadence of 'film' and that perhaps is just as silly as the FF argument I am criticising, but there you are.

Incidentally this FF revolution has been accompanied by the rise in gimbals, stabilisation and autofocus, because of the very lack of tripods and impossibly shallow DOF of FF. I've worked in BBC Natural History years ago and nobody hand held on telephotos or anything else for that matter. Yet there are a legion of young folks making wildlife films this way. And I, perhaps being old and cranky, want to shout out 'get it on a mofo tripod, you clown!'. I could never use autofocus, it seems the antithesis of filmcraft to me. But then I come from a different era

I could never use autofocus, it seems the antithesis of filmcraft to me. But then I come from a different era

Personal view on this is that it's not necessarily a generational thing, but because of the rise of solo shooter/operators and current production expectations of less people doing more jobs.

Personal view on this is that it's not necessarily a generational thing

Thanks Chris, I agree. I have worked on many British reality shows where the self shooting PD's use the de facto FS7 handheld mostly - it's the wrong camera for the job. They would be much better off with the Canon XF305s they used to use. They get 2 days training normally on these cameras and even after building up a fair amount of experience - IMV, they are the wrong camera for the job, the audience don't care that reality is shot on big sensors and it's all still HD anyway in UK broadcast.

Conversely I am currently working on a long running studio/VT insert series, that uses also FS7 and FX9s, in both cases, but they are used by proper cameramen and the the rushes look great. Right hands and right tools and horses for courses

.. 30p doesn't look anything like 24p to me. 50/60p no matter the sensor size looks like 'video' to me. I love 24/25p…

Steve, have you tried shooting 30p with a 216 degree shutter angle so the duration of blur approximately matches 24/25p? That would also solve the issue of projection on a 60Hz monitor.

Steve, have you tried shooting 30p with a 216 degree shutter angle so the duration of blur approximately matches 24/25p? That would also solve the issue of projection on a 60Hz monitor.

I live in Pal land Rick, there's no practical use for 30p here. I don't have any problems with 24/25p that Kim mentioned. 24/25p look fine to me on Netflix, on both my 60Hz screen laptop as they do on my multi-frequency 4K TV.

However bigger sensor has many advantages that come just simply from physics.

Yer a hopeless case Kim There are no compelling advantages due to physics where I can agree with you, as has been demonstrably shown throughout this thread. But as I say you want FF and a box camera, neither of which appeal to me, but I am convinced they will happen for commercial reasons, so you'll get what you want, regardless of what I say

A fact is something that can be proved. So where is it?

Steve, have you tried shooting 30p with a 216 degree shutter angle so the duration of blur approximately matches 24/25p?

Rick I forgot to add - post is my main business - Do NOT use 30p if you intend to distribute to other than 'NTSC' countries - it cannot be converted to anything really useful, other than 60psf, it will be unacceptable for international broadcast and streaming.

That is technically better image. Is it more cinematic,

not a given at all....

There's a whole lot to be said for degradation in theatrical exhibition. It's involving in the way pristine presentation isn't.
The irony is, folks who need the grit/grain of presentation the most -- namely, the no-budget realm -- seems to want it the least.

Meanwhile, the 'Dune' DP scanned his Alexa LF footage to 35mm, and then back to digital, to take the edge of perfection off.

Thanks Chris, I agree. I have worked on many British reality shows where the self shooting PD's use the de facto FS7 handheld mostly - it's the wrong camera for the job. They would be much better off with the Canon XF305s they used to use. They get 2 days training normally on these cameras and even after building up a fair amount of experience - IMV, they are the wrong camera for the job, the audience don't care that reality is shot on big sensors and it's all still HD anyway in UK broadcast.

Conversely I am currently working on a long running studio/VT insert series, that uses also FS7 and FX9s, in both cases, but they are used by proper cameramen and the the rushes look great. Right hands and right tools and horses for courses

Haha, this hit close to home. I've been pushing the company I work for to use the Canon XF605 for ENG/documentary-style events coverage, but no dice. Frequently wonder why they bother to hire video professionals when management with no industry experience dictate which what equipment gets bought and used, but oh well.

… 30p doesn't look anything like 24p to me. 50/60p no matter the sensor size looks like 'video' to me. I love 24/25p, it has the cadence of 'film’…

Steve, I should have quoted this more complete excerpt in my previous post. This implied to me you were familiar with both PAL and NTSC frame rates. So I was suggesting part of the problem with 30p was that the duration of the blur was too short compared to the blur of 24/25p video so the cadence of 30p felt off. I totally understand 30p isn’t relevant in PAL countries.

Steve, I should have quoted this more complete excerpt in my previous post. This implied to me you were familiar with both PAL and NTSC frame rates.

Thanks Rick, I understood. I am very familiar with with both - we do international versions quite regularly. 24 and 25p can cross borders, so too can 50i/p and 60i/p - 30p, if you'll forgive me, is the barsteward of all formats, it can only sit in 'NTSC' land or uToob, it has no passport All professional platforms, whether terrestrial or streaming would reject it out of hand, because of this and because it's not particularly 'cinematic' either. Interestingly you guys and us, have grown up with different notions of what is acceptable, when movies are shown on TV. You had a stutter that would have been unacceptable to us due to pulldown, and we had a pitch +4% change from 24 to 25, that would have been totally unacceptable to you. 24p is a world standard now we can all enjoy equally.

as another glass is raised to the super16 6K of the URSA 12K thinking how would this do in a pocket.
hey Rick, to get more blur to 30p try upping the shutter angle to 225 (the ratio of 24 to 30.)

@Steve, thanks for the additional background. Certainly agree with your conclusions and I think it would have been good if 25 fps was worldwide.

@Jeffrey, yes I’m thinking I’ll give that a try. I’ve used 360 degree shutter angle with 30 fps, sometimes works when movement is minimal or for dreamy dancers. But most times I use 180 degrees for 30 fps. I may follow your suggestion and go with 225 degrees shutter angle for my theatrical shoot next month. I’ve also shot 24 fps for theatrical and music events. My clients live in NTSC land and view video on computers and television; no one has complained about the use of either frame rate.

What I don’t like is having shutter angles less than 150/120 degrees unless the effect is desired.

Sigma FP is a huge disappointment, and I think Sigma also screwed up the BRAW part as well....

The camera can give a decent image with the proper conditions, but man, does it have poor dynamic range. The image is nice, but not that special, nobody is going to mistake the image with an Alexa. It's also god awful to use in an ergonomic sense. The autofocus is also bad. I don't think it's worth the trouble in this day and age to shoot with this camera.

I've pretty much given up on finding that magical camera that's inexpensive but has an Alexa-like image. I think the Pockets are the closet, then followed by the Nikon Z6/Z7 (w/Ninja V ProRes) and then the Sony A7SIII/FX3. I own all those cameras and although I like them, I'm never fully happy because not matter what, the image they render doesn't look like an Alexa.

I don't like using cDNG on it, but only because it's so unwieldy. Look wise, it's very good, in spite of its linear encoding. Sigma did a LOT right with that camera.

The DNGs do open up for me after doing a transform from Linear to REDWideGamutRGB / Red Log3G10. I do feel that it has about 12 stops, if that's enough? . . . most situations it seems to match the Pocket 6K with a little less noise at similar ISO, the rolling shutter seems similar too. I usually have to go 2 stops over with the 4.6K, I suppose I'm rating it at 3200 ISO at that point to get more perceived DR over the FP.

But also I do hope I would see a camera like this full frame soon. This one BTW is with already existing BMD lens mounting system.

It does seem like there is an opportunity to partner with Sigma for an L mount BMD Pocket LF with an FP sensor or something like the Dual ISO sensors in the Pocket 6K. Maybe more likely than the 12k from a cost standpoint.

I'd buy a Micro Pocket 6K to replace the FP too.

Leica owns “L” mount. It’s up to them.

Why would BMD want to use a sensor that sigma designed ?

The FP is disappointing.

JB