Pardon? What did I say that was ignorant? My point wasn't whether Panasonic will produce a $3K micro 4/3 camera, but given that pieces of the technology -- whether it's from them or someone else -- already exist at cheaper price points in vDSLRs and consumer video cameras such that there will be downward pressure on prosumer/low-end professional price points is valid. You can consider it unrealistic if you accept the pricing models and product-design approaches manufacturers have been using. By putting HD video -- if even with its aliasing, moire, and rolling-shutter artifacts -- on vDSLRs, camera manufacturers have created a growing demand for larger imagers among a sizeable segment of video camera consumers. In the not-too-distant future, the $5-10K range will give way to a lower price range. Look at what Red's doing. And if they hadn't decided to go modular, the resultant higher economies of scale may have yielded a $3K 3K camera.
Sony's answer to the new Panny...
- Thread starter ATL Media Group
- Start date
Barry_Green
Moderator
Yes, but more megabits doesn't necessarily mean more quality. Look at the Canon 7D... 50 megabits, but doesn't compare to the 21 megabits of PH mode. You can't just look at the numbers and make a decision, there's more to it than that.
Point is, PH mode is proven, it's better/more robust than XDCAM-EX, it's nothing to discount.
Barry_Green
Moderator
We don't know. Only the engineers know.
Again, does it matter? Results matter, but if they choose a method that delivers flawless fantastic super-sharp results, does it really matter whether they did it through downscaling, through line skipping, through osmosis, through ionization, or through a Ouija board?
Emanuel
Cross-Examiner
Yes, I think that's the question.
your looking at it at a totally different perspective.
look at the 1d mark IV, the thing costs 5k cuz its a highend still cam. compare that to a t2i the stills feature... fook me! t2i has 2 more megapixels than a 1d mark iv! yeah but look at the spec sheet 1D mark iv is completely professional
now lets do that analogy on the gh1 -> af100
Barry_Green
Moderator
Really?
Why?
Aren't you more interested in the results, rather than the method used to achieve the results?
I think a three-chip system would produce results superior to a Bayer chip, but Red (and Canon, and Panasonic) have chosen to go with a single Bayer chip, and people seem to be satisfied with the results from the Red One, don't they? However, I don't think you'll find a credible engineer on this planet who would argue that a Bayer chip is a superior imaging approach as compared to three chips with a prism! Bayer is a compromise, one that costs at least 20% to 30% of the resolution a chip is capable of, it limits the resulting image to about 4:2:0 color sampling, and yet -- where's the "outrage" against Bayer? Fact is, it's a reasonable choice, given that a primary design parameter of digi-cine cameras is that they be able to use film lenses, which rules out three-chip designs... so, it can be said that empirically, Bayer sucks. But subjectively, looking at the results, it does just fine.
So why all the obsession over something that nobody (outside of a couple of Panasonic engineers in a factory in Osaka) can answer? Why not just look at the results, and see if they deliver on their promise of full 1080p with no aliasing or moire? When there are results, or the ability to test the unit, you can bet that we'll dig into it and find out what type of quality it delivers.
If it delivers an image as sharp, crisp, and clear as an EX1, but with the shallow DOF of a 7D, are you seriously concerned whether it does it through line-skipping, tightrope-jumping, juggling, sleight of hand, smoke and mirrors, or skullduggery?
Because if line-skipping is "bad", then everyone who's declaring it as bad must immediately condemn the DSLRs.
And if the DSLRs are good, then y'all have to say line-skipping is good.
There's no middle ground.
(and with all that said, nobody knows whether the AF100 uses line skipping or not. It very well may not.)
Last edited:
First, throw away some pixels! The sensor is a 4:3 format, but you want a 16:9 picture. So you don't even scan the top and bottom lines of the sensor, basically throwing away 25% of the chip. This should also help improve your frame rate.
Then you de-Bayer, losing another 20% or so of the pixels. Then you scale the remaining pixels down to 1920x1080. And voila!
I have a good idea how the camera is made -- probably as well as you. But you ignored the comparison to the 150, which has essentially the same construction and features (at least the ones mentioned for the AF100 so far), except for HD-SDI, but which has expensive glass and a $2K lower price.
By the way, Jan Crittenden said the AF100's imaging chip is the same as the one in the GH1 (here): "Okay several questions here. First the sensor is the same as the GH1 Lumix camera, then we optimize it for HD recording."
Yes, and my comparison to the HMC150 covers that.
Barry_Green
Moderator
And the HMC150 probably has a market at least 4x as large as a dedicated digital cinema camera. And you don't know what all the AF100 does. There might be all sorts of things in there that we don't know about yet.
And besides, since when does what it costs to make something, bear *any* relationship to what that product sells for? Manufacturing cost has little to nothing to do with sales cost. Marketing, distribution, R&D, all these things are far more costly on a smaller-market product.
But none of that matters. All that matters is whether you would buy it or not. If you won't, then go ahead and exit the thread. If you will, stick around and let's discuss it.
And besides, since when does what it costs to make something, bear *any* relationship to what that product sells for? Manufacturing cost has little to nothing to do with sales cost. Marketing, distribution, R&D, all these things are far more costly on a smaller-market product.
But none of that matters. All that matters is whether you would buy it or not. If you won't, then go ahead and exit the thread. If you will, stick around and let's discuss it.
im guessing a lens will come with it and it will be a lot less than 6k. i remember hmc150 being announced at 4500 then on release it was 3200.
yes i know jan said that but somebody gave a link to a japanese website that says its new sensor.. thats why i said unconfirmed
neway
you cant just compare features and base the price on that. its not lego where you just take one piece and compare it to another!
heres my opinion on the price
you're right, it could very well be 2-4k on producing/marketing/distribution the camera and the rest would be profit but we still want panasonic to profit dont we?
lets move back on topic
Last edited:
mikkowilson
Steadi-Guru
And they aren't saying. I asked them.
And I fully agree with Barry. Who cares how they do it if it works? It's a brand new type of camera, so even if we did know how it did it; it would be impossible to try and predict results as we have no prior data to go on.
- Mikko
Steve Castle
Well-known member
On the subject of how the AF-100 could lower its resolution from 12MP to 1080P. Obviously the easiest is line-skipping. Another would be to capture the whole frame and then scale it, the third way is pixel-binning.
Pixel-binning operates very differently from a CCD and CMOS sensor.
For CCDs, they operate in a binned mode natively. Charge from adjacent pixels in a column are summed in the horizontal register. Charge from adjacent columns are summed at the output node. The effective resolution of the CCD is decreased (in this case to 1080p) depending on how many pixels are summed in each direction.
Also, the S/N (signal-to-noise) ratio is increased by the square root of the number of pixels binned if the noise is dominated by shot noise and the S/N is increased linearly with the number of pixels binned if the noise is dominated by read noise. In a CCD, unlike a CMOS, the binning operation is noiseless since summation takes place in the charge domain. This is a major benefit for CCDs over CMOS sensors. However, CCDs have difficulty scaling-up, something the CMOS is good at, and where pixel-binning would be most beneficial.
Now, if pixel-binning is to happen without a CMOS sensor, there needs to be a fundamental operational change. The sensor would need to integrate an active-pixel array with a passive memory cell array to permit frame-transfer operation. This is so that charge integration amplifiers at the bottom/horizontal of each column of the sensor can allow for binning of the pixel during the read-out process.
So if this sensor is going to do pixel-binning it will require additional passive memory on the sensor to be able to do it. However, I'm not sure the cost-to-benefit ratio is that significant (it would be worth it for video, but not for stills). Hitherto, these sensors have been designed for stills first, with video as an afterthought. Obviously, as the market for FourThirds sensors has shifted, so does how its designed. So there is a possibility that these sensors can be read-out fairly fast with video in mind. Obviously, Panasonic is trying to find a middle-ground between video/stills for these FourThirds sensors.
Also, Panasonic uses an N-type MOSFET sensor (NMOS), and not much is openly published on it, but it seems to be a fairly flexible resolution sensor judging by past iterations, and the aliasing issues haven't been incredibly bad. I'm sure they've made significant strides for a video-centric device AF100 even if it line-skips. At this point its wide-open, but there are a lot of options that Sony/Panasonic (and Canon hopefully soon) could take.
Pixel-binning operates very differently from a CCD and CMOS sensor.
For CCDs, they operate in a binned mode natively. Charge from adjacent pixels in a column are summed in the horizontal register. Charge from adjacent columns are summed at the output node. The effective resolution of the CCD is decreased (in this case to 1080p) depending on how many pixels are summed in each direction.
Also, the S/N (signal-to-noise) ratio is increased by the square root of the number of pixels binned if the noise is dominated by shot noise and the S/N is increased linearly with the number of pixels binned if the noise is dominated by read noise. In a CCD, unlike a CMOS, the binning operation is noiseless since summation takes place in the charge domain. This is a major benefit for CCDs over CMOS sensors. However, CCDs have difficulty scaling-up, something the CMOS is good at, and where pixel-binning would be most beneficial.
Now, if pixel-binning is to happen without a CMOS sensor, there needs to be a fundamental operational change. The sensor would need to integrate an active-pixel array with a passive memory cell array to permit frame-transfer operation. This is so that charge integration amplifiers at the bottom/horizontal of each column of the sensor can allow for binning of the pixel during the read-out process.
So if this sensor is going to do pixel-binning it will require additional passive memory on the sensor to be able to do it. However, I'm not sure the cost-to-benefit ratio is that significant (it would be worth it for video, but not for stills). Hitherto, these sensors have been designed for stills first, with video as an afterthought. Obviously, as the market for FourThirds sensors has shifted, so does how its designed. So there is a possibility that these sensors can be read-out fairly fast with video in mind. Obviously, Panasonic is trying to find a middle-ground between video/stills for these FourThirds sensors.
Also, Panasonic uses an N-type MOSFET sensor (NMOS), and not much is openly published on it, but it seems to be a fairly flexible resolution sensor judging by past iterations, and the aliasing issues haven't been incredibly bad. I'm sure they've made significant strides for a video-centric device AF100 even if it line-skips. At this point its wide-open, but there are a lot of options that Sony/Panasonic (and Canon hopefully soon) could take.
Are you referring to the Red Scarlet or the AF100? Because I'd hardly call the AF100 a dedicated digital cinema camera. It's essentially the HMC150 with a bigger imager. Its codec is far from ideal for digital cinema.
Fair comment. And if it comes with a fairly fast 8-10x zoom, then I'm more kindly disposed. Or 50-Mbps data rate and 10-bit, 4:2:2 subsampling (High 4:2:2 Profile?). Or 1080p60, like Panasonic's 3MOS 700-series consumer cams.
As is manufacturing. But how do you figure the AF100 will have a smaller market that the HMC100's? Solely by the size of the imager? Other than that it sounds like they have pretty nearly the same specs, although I suppose the HMC150's 3.9- to 51-mm zoom would be lighter than an equivalent (and as yet hypothetical) 14- to 182-mm micro 4/3 zoom.
Still, I'd say the market for the AF100 stands to be at least as large as the HMC150's, especially if Panasonic can either keep the price lower than $6,000 or demonstrate distinguishing features (other than a bigger imager).
How would I know if would buy it? I'd have to hear a lot more about it before I could make a decision.
Last edited:
Uh, isn't that what we were both doing (magnesium vs. plastic, etc.)? Isn't comparing features and build quality how, at least in part, we determine parity? I mentioned that the HMC150 and AF100 were in many ways similar. I don't think it's unreasonable to ask how, given that similarity, the AF100 might be worth $2K more. Yes, a larger sensor, but a less complex (and possibly cheaper to manufacture) optical system (no prism block and no lens). And I don't believe that the two will have significantly different economies of scale that would make the AF100 more expensive to manufacture, market, or distribute.
I agree that no matter how they avoid aliasing, it's gonna be a game changer.
Still, it is very important how they do it. If line-skipping, sensitivity and noise will be worse due to lower fill factor (not using entire sensor area). Also, if oversampling was used, the MTF (sharpness) could benefit, due to OLPF design (higher spatial Nyquist frequency).
My bet is they're gonna use line-skipping. When Jan said they avoid moire by using OLPF, she probably meant that.
I hope it's gonna have less rolling shutter than GH1, though. All in all it's gonna cost 10x (ten times!) as much. Also, Jan said they're working on optimizing the sensor for video. I know it's the market that sets the price, but I keep in mind it's Panasonic - they've always offered great value. Charging additional $5400 for replaced OLPF, different enclosure, XLR input and better chip/software (uncompressed output, AVCCAM codec) would make me hesitate. Oh, yes - I forgot to mention the handle
combatentropy
Veteran
Graeme Nattress.
Not to pick a fight, and in context you were saying that Bayer is fine. But for its compact allowance of large sensor designs and simpler lenses, mature color decoding algorithms, and efficient use of total pixel counts, Bayer is best.
Last edited:
Hunter Hampton
Carbonite Member
Its all speculation at this point, but regarding line skipping; they could use it on this camera and still get an amazing image that would look as if it was captured from the whole sensor and scaled to 1080p.
Here is why; just because dslr cameras only grab 1080 lines or 720 lines to get HD- doesnt mean that they will do the same and then just throw a heavy low pass filter over the sensor. On this new video camera, they could grab many more lines than 1080 and scale that image to 1080 making extremely sharp 1080p without alaising (from oversampling the image- but still line skipping on the sensor)
Make sense? Line skipping doesnt have to mean a bad image- these aint dslrs...
Here is why; just because dslr cameras only grab 1080 lines or 720 lines to get HD- doesnt mean that they will do the same and then just throw a heavy low pass filter over the sensor. On this new video camera, they could grab many more lines than 1080 and scale that image to 1080 making extremely sharp 1080p without alaising (from oversampling the image- but still line skipping on the sensor)
Make sense? Line skipping doesnt have to mean a bad image- these aint dslrs...
Barry_Green
Moderator
I know Graeme's position on Bayer very well, and I've said that Bayer was the ONLY choice that made sense for what Red was doing.
But I feel quite confident in saying that Graeme would agree with me, that three equal chips with a prism would deliver better imagery than a single chip with a Bayer mask. In fact, it's just plain silly to argue otherwise. One system delivers 100% resolution with 4:4:4 color, and the other system delivers about 78% resolution with around 4:2:0 color.
In that thread you quoted, Graeme is not comparing like to like, he's comparing three smaller chips against one larger Bayer chip. I'm saying three equal chips with a prism, vs. one with a Bayer.
This isn't really even in dispute, this is long-since settled science. There's a reason every broadcast camera uses three chips. The only one-chip designs are either those in cheapo consumer cameras, or those that specifically want to use film lenses (including very expensive cameras, of course, such as the F35).
But I feel quite confident in saying that Graeme would agree with me, that three equal chips with a prism would deliver better imagery than a single chip with a Bayer mask. In fact, it's just plain silly to argue otherwise. One system delivers 100% resolution with 4:4:4 color, and the other system delivers about 78% resolution with around 4:2:0 color.
In that thread you quoted, Graeme is not comparing like to like, he's comparing three smaller chips against one larger Bayer chip. I'm saying three equal chips with a prism, vs. one with a Bayer.
This isn't really even in dispute, this is long-since settled science. There's a reason every broadcast camera uses three chips. The only one-chip designs are either those in cheapo consumer cameras, or those that specifically want to use film lenses (including very expensive cameras, of course, such as the F35).
Last edited:
Ben_B
Veteran
The HMC-150 has a powered zoom lens which this camera won't have unless they do some kind of broadcast lens for it. This alone makes the HMC-150 make more sense for ENG, some kinds of documentary/event filming, and other applications. It is a self contained unit that does not require much lens switching, and does not suffer from problems associated with flickering varifocal zooms. The AF100 is best suited to digital cinema or other applications where a cinema look is desired.