AF-100 Another DSLR $9,000 4:2:0 is a solution?

[Otis Grapsas]

The very ergonomics of weight distribution is not good. Especially with a heavy HJ22 zoom lens. I just don't like it. It's archaic.

The weight and weight distribution is typical for a camcorder of this type. It's aproximately on the center of the camera without a lens. It must have been impossible for them to balance the camera with a heavy lens. The center of gravity would have to be way back, very close to the battery, the camera would be 45cm long or more. In order to achieve this I build a camera 40cm long and although it balances fine with a 700gr monitor on the front and everything, that's only with a lens up to 600gr (the small primes I use). Anything heavier and the right (holding) hand gets weight and perfect balance is lost. With a viewfinder you can push it to the back to rebalance though. The 2700 is a very good design. A prosumer handheld setup is not more balanced that a professional shoulder mount camcorder. It's not as stable with unstabilised lenses in telephoto. The weight in hand instead of on the shoulder? Or front heavy in one of those laughable shoulder mount adapter accessories with the pipework, the ridiculous rods that end up pressing on your chest and move when you breath and the ones that require your hands to be too far from the body? Are we shooting or searching for water with a stick?!

 

[Duke M.]

Fast glass can transform any camera. It's where good light gathering and shallow DOF comes from.

Even if we ignore shallow DOF, one can see how ISO affects a sensor of this type in dynamic range here:
http://www.dxomark.com/itext/review_panasonic_gh1/dynamic_range.jpg

The color precision here:
http://www.dxomark.com/itext/review_panasonic_gh1/color_sensitivity.jpg

The noise levels here:
http://www.dxomark.com/itext/review_panasonic_gh1/snr_18.jpg

If it's ISO100 f2 vs ISO400 F4, it's not that bad. But if it's ISO400 f2 vs ISO1600 f4 or ISO3200 f5.6 it's world of difference.

Interesting that the fall off is that linear in all three categories.

 

[Otis Grapsas]

Interesting that the fall off is that linear in all three categories.

That means there is no extra processing added. It's the natural degradation of a sensor system, half the light vs the same noise. It is the same in all good designs and in all integrations I have tested and on which I had total control on what was being applied to the signal.

 

[Barry_Green]

DOF is a key selling point of Af100 though.

Yes, but there are many other compelling features about the AF100 besides super-hyper-shallow DOF, like variable frame rates in 1080/60p.

If DOF wasn't important, one could get a less expensive CMOS Panasonic.

Yes, partly -- understanding that those less expensive units have an effective ISO of about 50, so it's not exactly a direct comparison.

Olympus 14-35 f2 and 35-100 f2 are very nice because f2 allows AF100 to get shallow DOF that matches a 7D with f2.8 lenses

The DOF difference between an AF100 and a 7D isn't a full stop! It's more like 1/4 of a stop.

Low light is also important. f2 allows 2 to 3 stops lower ISO vs the 14-140 zoom. ISO400 instead of ISO1600 and ISO3200. In modern sensors ISO400 is very clean while ISO3200 has high noise and much lower dynamic range, and in the GH series, horizontal stripes from electrical interference. I can see these stripes in Bloom's ISO3200 sample also.

EDIT: So the question is: Would you pay 4,000 euro extra to improve the shallow DOF, the low light, and the dynamic range performance in low light by 2 to 3 stops? I would.

As said before, if someone's main priority is low light performance, then they should be getting the faster lens, obviously. These are two different products with different priorities. If someone shoots daylight, the 14-140 has numerous substantial advantages.

I mean, how can they really be compared at all? One is a 10x zoom for $700, the other is a 3x zoom for $2500.

Again, it all comes down to what your priorities are, and what each lens gives you. One costs $2500 and gives you a faster f-stop. Period. The other gives you 3.3x as much zoom range, silent autofocus, optical image stabilization, parfocal zoom, focus and zoom readout in the LCD, barrel distortion correction, chromatic aberration correction, and probably more.

Frankly, I think someone would be foolish for NOT getting the 14-140. It just adds too much capability to the camera. It turns it into a full-fledged video camera, minus power servo zoom.

Nobody says the 14-140 has to be your only lens. I'm just saying I think it offers too much convenience and capability to not have it.

 

[Yak]

Yes, but there are many other compelling features about the AF100 besides super-hyper-shallow DOF, like variable frame rates in 1080/60p.

Barry do you know if there are any external recorders that can handle " variable frame rates in 1080/60p" or even 29.97 in 1080P.

 

[Barry_Green]

I doubt external recorders will be able to record the variable frame rates beyond 30fps. I don't know if the HDMI even transports any progressive rate beyond 30... I believe that for 12 to 30, it's embedded on the SDI stream within a 60i signal, so any recorder should be able to handle that.

This is speculation, however.

 

[Yak]

I doubt external recorders will be able to record the variable frame rates beyond 30fps. I don't know if the HDMI even transports any progressive rate beyond 30... I believe that for 12 to 30, it's embedded on the SDI stream within a 60i signal, so any recorder should be able to handle that.

This is speculation, however.

Jan if you see this then maybe you could inform us has how the HPG-20 P2 recorder will interface with the AF-100 as relates to variable frame rates as well as how it will handle 24P and 29.97 in 1080P.

 

[Otis Grapsas]

The DOF difference between an AF100 and a 7D isn't a full stop! It's more like 1/4 of a stop.

If 4/3 is 18mm wide and since it's 4,000 effective pixels, a reasonable effective 3840x2160 crop for video output on an optimal 2:1 ratio would be 17.28mm wide. 7D is 22.3mm, so it's 1.67x more light, 2/3 of a stop. I haven't seen a Panasonic publication on the sensor size for Af100 so I'm assuming something that would work well. And 5DII full frame would be about a 2 stop difference.

1 stop is 2.000x the light and 1.414x the aperture diameter
1/2 stop is 1.414x (pow(2,1/2)) the light and 1.189x the aperture
1/3 stop is 1.260x (pow(2,1/3)) the light and 1.122 the aperture
1/4 stop is 1.189x (pow(2,1/4)) the light and 1.090 the aperture

The AF100 would have to be an 20.46mm effective width in video mode to be 1/4 stop different than a 7D.

The Panasonics with the low effective ISO come with very fast lenses. f1.6 and f1.8 I believe. This allows them to compete with f4-5.8 of higher effective ISO, they do not need high ISOs to match the 4/3" sensor f4-f5.8 exposure.

The 4/3" GH1 effective ISO is no more than ISO100 anyway in all tests I have seen, as the one I listed above. The higher ISOs are usable gain which compromise performance as measured in these graphs, 1 stop per 2x ISO increment starting from ISO200. If we compare native to native, the cheaper Panasonics win. If we accept that the usable ISO is 2 stops better, the cheaper Panasonics still compete due to the fast fixed zoom apertures.

A common complain of the GH1 customers is the noise, because they actually need to use ISO1600 to shoot a birthday party with the f4-5.8 lens. Even a consumer f1.8 camcorder gathers 5x more light, so it only needs ISO320 to match the GH1 ISO1600, in ISO320 they might still be usable even if their native ISO is lower by a stop. And a EX1 and a DVX can shoot that at 0dB gain. ISO1600 and ISO3200 are not much use if you actually need them any time you need reasonable low light because you are limited by the slow zoom. That's why I believe in the fast lenses. They will allow operating the AF100 at its optimum image quality settings and not degraded by 3 and 4 stops in everything. Why turn a 11stop ISO200 camcorder into an 8stop ISO1600 camcorder when you can avoid it? I wouldn't recommend using these high ISO as though they were native/low noise/0dB ISOs.

 

[agcohn]

I doubt external recorders will be able to record the variable frame rates beyond 30fps. I don't know if the HDMI even transports any progressive rate beyond 30... I believe that for 12 to 30, it's embedded on the SDI stream within a 60i signal, so any recorder should be able to handle that.

This is speculation, however.

The Cinedeck website claims to support the following HD Formats:
720p50, 720p59.94, 720p60, 1080PsF23.98, 1080p23.98, 1080PsF24, 1080p24, 1080PsF25, 1080p25, 1080PsF29.97, 1080p29.97, 1080PsF30, 1080p30, 1080i50, 1080p50, 1080i59.94, 1080p59.94, 1080i60, and 1080p60

However, they don't specifically say it records in all of those formats.

And some of those formats are confusing like 1080i60 in addition to 1080i59.94?

 

[Barry_Green]

If 4/3 is 18mm wide and since it's 4,000 effective pixels, a reasonable effective 3840x2160 crop for video output on an optimal 2:1 ratio would be 17.28mm wide.

No need to jump through those math hoops. It's 18.9mm. And I've verified the FOV against a 7D; a 100mm lens on the 7D matches the FOV of an 85mm on the GH1, almost perfectly. So the crop is 1.18 a vs. the 7D, and the DOF difference is about 1/4 stop.

 

[BobDiaz]

Whatever the resolution of the image sensor, chances are it's either an integer multiple of 1920x1080 or if a fractional amount, it's like 2 & 1/2. The reason I think it's something like that is that when converting from the higher native resolution to 1920x1080, it's easier to be a multiple of 2 or at lease an integer value. Fractions make for a messy conversion, not impossible, but just messy.

A 2x2 --> 1 is very clean and easy, but a 2 & 1/32 x 2 & 1/32 is much harder and messy.


Bob Diaz

 

[BitMaestro]

Not that it's worth "jumping through math hoops" for a fraction of a stop, but since I'm procrastinating...

If 4/3 is 18mm wide and since it's 4,000 effective pixels, a reasonable effective 3840x2160 crop for video output on an optimal 2:1 ratio would be 17.28mm wide. 7D is 22.3mm, so it's 1.67x more light, 2/3 of a stop. I haven't seen a Panasonic publication on the sensor size for Af100 so I'm assuming something that would work well.

The standard 4/3 sensor is 17.3mm wide, not 18mm. And when shooting 16:9, the GH1's multi-aspect sensor is 4,352 pixels (and ~18.8mm), not 4,000 pixels. Similarly, the GH2's multi-aspect sensor (which is widely assumed to be the basis for the AF100's sensor) goes from 4,608 pixels (4:3) to 4,976 pixels (16:9) wide.

And while it's true that Pany hasn't said anything specific about the AF100's sensor, Jan did mention that the full resolution is ~12 MP. If we also abide by Bob's theory and use a relatively non-messy 2.5 ratio, that would be 4,800x2,700 which, given the above, works out to a crop factor of 1.22 relative to the 7D. And that--by the calculations in your post: "1/2 stop is ... 1.189x the aperture"--is just over 1/2 stop.

Mark

P.S. re: "A common complain of the GH1 customers is the noise, because they actually need to use ISO1600 to shoot a birthday party with the f4-5.8 lens."--IMO they actually need to use the 20mm f1.7 pancake lens.

 

[Otis Grapsas]

No need to jump through those math hoops. It's 18.9mm. And I've verified the FOV against a 7D; a 100mm lens on the 7D matches the FOV of an 85mm on the GH1, almost perfectly. So the crop is 1.18 a vs. the 7D, and the DOF difference is about 1/4 stop.

1.189 is 1/2 a stop.

 

[Otis Grapsas]

Not that it's worth "jumping through math hoops" for a fraction of a stop, but since I'm procrastinating...



The standard 4/3 sensor is 17.3mm wide, not 18mm. And when shooting 16:9, the GH1's multi-aspect sensor is 4,352 pixels (and ~18.8mm), not 4,000 pixels. Similarly, the GH2's multi-aspect sensor (which is widely assumed to be the basis for the AF100's sensor) goes from 4,608 pixels (4:3) to 4,976 pixels (16:9) wide.

And while it's true that Pany hasn't said anything specific about the AF100's sensor, Jan did mention that the full resolution is ~12 MP. If we also abide by Bob's theory and use a relatively non-messy 2.5 ratio, that would be 4,800x2,700 which, given the above, works out to a crop factor of 1.22 relative to the 7D. And that--by the calculations in your post: "1/2 stop is ... 1.189x the aperture"--is just over 1/2 stop.

Mark

P.S. re: "A common complain of the GH1 customers is the noise, because they actually need to use ISO1600 to shoot a birthday party with the f4-5.8 lens."--IMO they actually need to use the 20mm f1.7 pancake lens.

Panasonic should have the active pixels/sensor size in 16:9 video on the pdf of the Af-100 when they release it.

Home users always get the cheap zoom only and expect miracles

 

[Nick Navaro]

So I guess most of you have seen Philip Bloom's review of the AF100

http://philipbloom.net/2010/10/19/af100/

What did you think? He points out the gripe with the lenses and AVC 24mb codec not being great as most people were moaning about on here. Like I said, I'm in the field and don't want to attach extra bits for better quality....

Anyone know anything about the new Sony mini cinealta 35mm cmos chip cam due to come out next March?? That sounds like it would trump the AF100 easily... Apparently, someone said, not me, Sony wouldn't be as careless as Panasonic leaving an unfinished product in to the hands of others. I only want to buy ONE camera for now, and Philip's review has put me off buying now. Mainly due to the lenses issue.

And Barry, if you read this, this 0.95 lens you recommended is vapour ware apparently. Can't find one anywhere and experienced people have been saying you probably won't get one. :-(

x2 crop factor on the AF100 sucks IMHO. :-(

 

[Homunculus]

So I guess most of you have seen Philip Bloom's review of the AF100

http://philipbloom.net/2010/10/19/af100/

What did you think? He points out the gripe with the lenses and AVC 24mb codec not being great as most people were moaning about on here. Like I said, I'm in the field and don't want to attach extra bits for better quality....

Anyone know anything about the new Sony mini cinealta 35mm cmos chip cam due to come out next March?? That sounds like it would trump the AF100 easily... Apparently, someone said, not me, Sony wouldn't be as careless as Panasonic leaving an unfinished product in to the hands of others. I only want to buy ONE camera for now, and Philip's review has put me off buying now. Mainly due to the lenses issue.

And Barry, if you read this, this 0.95 lens you recommended is vapour ware apparently. Can't find one anywhere and experienced people have been saying you probably won't get one. :-(

x2 crop factor on the AF100 sucks IMHO. :-(

AF100 does not have x2 crop factor nor any lense problems. the crop factor on AF100 to cinema 35mm film is about 1.2 which is almost negligible. it has x 2 crop factor for full frame PHOTOGRAPHY 35mm negative. are you buying the AF100 for PHOTOGRAPHY or for shooting videos/movies?

to say AF100 has x2 crop factor means you're also saying Red One and Arri 535 has x2 crop factor as well which is true because AF100 Red One and all 35mm film cameras from Arri to Panavision have about the same size sensor/negative

 

[Ian-T]

x2 crop factor on the AF100 sucks IMHO. :-(

Here we go again. LOL...what ....do you plan on shooting stills with it or something? There is no x2 crop. So I guess real cinema cameras like the Red One or other 35mm cams sucks also?

Ha...Homunculus I responded before actually reading yours.

 

[Barry_Green]

So I guess most of you have seen Philip Bloom's review of the AF100

No, we've seen Philip's comments on a barely-functional prototype. 75% complete. Not a review, as the unit isn't done yet.

He points out the gripe with the lenses and AVC 24mb codec not being great as most people were moaning about on here.

And he also said it was a "stonking great" camera too.

He shouldn't have said anything about the codec, because it's not done. Panasonic has asked users to not post footage from the native codec yet, as it's in a preliminary stage. Even so, the codec is a darn sight better than the codecs people have been using, such as those found in DSLRs, HDV cameras, and XDCAM-HD or XDCAM-EX cameras. The codec is fine.

Like I said, I'm in the field and don't want to attach extra bits for better quality....

You won't likely have to. I would wager 95% of the buyers will never bother with external recorders.

Anyone know anything about the new Sony mini cinealta 35mm cmos chip cam due to come out next March??

What we know is that they've said it'll be $50,000 and come out in April.

And Barry, if you read this, this 0.95 lens you recommended is vapour ware apparently. Can't find one anywhere and experienced people have been saying you probably won't get one. :-(

The Voigtlander 25mm f/0.95 is not "vaporware", it's just not out yet.

 

[Ian-T]

Anyone know anything about the new Sony mini cinealta 35mm cmos chip cam due to come out next March?? That sounds like it would trump the AF100 easily... Apparently, someone said, not me, Sony wouldn't be as careless as Panasonic leaving an unfinished product in to the hands of others. I only want to buy ONE camera for now, and Philip's review has put me off buying now. Mainly due to the lenses issue.

But why would you want to compare a camera that will cost 10x as much as the AF100? If it doesn't kick the AF100's behind at that price point....then I'll be the first to say the Sony stinks. In others words....I expect it to run all over the AF100.

 

[Nick Navaro]

No, we've seen Philip's comments on a barely-functional prototype. 75% complete. Not a review, as the unit isn't done yet.


And he also said it was a "stonking great" camera too.

He shouldn't have said anything about the codec, because it's not done. Panasonic has asked users to not post footage from the native codec yet, as it's in a preliminary stage. Even so, the codec is a darn sight better than the codecs people have been using, such as those found in DSLRs, HDV cameras, and XDCAM-HD or XDCAM-EX cameras. The codec is fine.


You won't likely have to. I would wager 95% of the buyers will never bother with external recorders.


What we know is that they've said it'll be $50,000 and come out in April.


The Voigtlander 25mm f/0.95 is not "vaporware", it's just not out yet.

What is your source for the $50k Sony? I heard $20k. Which for what it can can do, sounds amazing. And March release.

Well if the lens isn't out yet, then it's currently classed as vapor ware, ala the scarlet and epic... I can't wait years (as people have been with scarlet and epic) for a bloody lens! I need to shoot now!

And the people at crew.tv shot some OK footage with a $20k lens, apparently. What's the point of a $4.5k priced cam, but you need a tens of thousands of dollars lens to get nice footage??

And why are people bandying about it's "75%" done. It's been "75%" done for about 2 months! In 2 months they've not moved a single percent with a test product??

The x2 crop factor is real and exists, not just because of the smaller than 35mm sensor, but also because there are no suitable video lenses for this camera available at all??

"Apparently" "someone" is making a suitable wide lens. Why no more details than that??

Even if you're pro-panny like Barry. Surely you've got to admit the "clumsiness" of the whole set-up of this cam. As someone put it.

If it was a good straight-forward video camera with all the accessories, then I doubt you'd have a thread this long with all the pitfalls.

I want to believe, I really do.