The brochure that's making the rounds on the web looks like a low-resolution PDF which trashes the quality of the images. ( also some of the images in the brochure are recycled from other Panasonic brouchures that are quite old now )

This still looks like a mock-up of the final brochure to me.

I've seen these measurements (17.3mm x 13mm) on brochures and official Panasonic websites for the GH1 as well, despite the fact it uses the larger multi-aspect sensor. I think these are just the default specs that they write for all their micro 4/3rds products.

The important thing here is to look at the aspect of the AF100, which is 16:9 native. It is probably derived from a 4:3 chip and my best guess going on the footage I've seen, is it's a very similar sensor to the GH1 ( if not the same ). So the actual size will be a lot closer to S35, than to S16.

If you're planning on shooting 2.35 with lenses that will cover S16, you might still have problems with vignetting, as has already been pointed out. I guess you could always look through some of the image galleries on Flickr or similar, where there are many samples from people shooting m4/3 cams with c mount lenses. Of course, a lot of these old 16 lenses wouldn't cover S16, but it might give you more of a feel for how it would work.

Oh dear uncertainty creeping in again.

My take is If we have to make a choice between what the brochure and Panasonic representatives have said IE the sensor is cropped from a 17.3 x 13mm frame and what others think but is the opposite. Really unless we hear otherwise and if we have been misinformed but we should accept what the brochure says.

For all of us though this is frustrating and unless someone who knows for sure speaks up then who can seriously buy a £4000 camera if they don't know what they're getting?

Mark, this link might be some help. It's one of the best explanations IMO, of how the over sized sensor in the GH1 works, especially in relation to other m4/3 cameras such as the G1.
The biggest Four Thirds sensor yet (Four Thirds User)

Obviously this doesn't help with S16 coverage, but I would imagine that you would be lucky to get full coverage at all focal lengths. I'm hoping to rent an AF100, sometime in the new year, when they are more widely available and this might help decide if it is worth the extra outlay over the GH1 for me.

Adrian

Yep nice description and shows how the four thirds sensor could be different sizes as we already discussed. For 16/9 on the 101/100 though I'm 75% sure NOW the sensor is cropped!

From what i garnered from the IBC Vimeo clip, It's 4/3, 4:3 aspect ratio imager. I thought that's where I'd heard it! Go ahead and listen, carefully, to the description of the accommodation of the 16:9 "slice".

I'm glad that I do know the difference between the convention of a 4/3 and a 4:3 ratio, and here, unless somebody would like to correct me, on what I'd heard, was where I correctly got the comparison being made.

Listen carefully, it's around 9:38, on this Vimeo link, and starts as a result of a discussion around the ISO settings and 3200 which leads into this further discussion around the actual pixels being employed, listen carefully:-

IBC Vimeo clip

Now, for me, the real question still remains, is this 16:9 "slice" being achieved from the WHOLE 4:3 physical surface and then being electronically arranged as a 16:9 slice? Or is this 16:9 "slice" a physical masked area of the whole chip?

My opening post was about getting to grips about what I'd heard from this clip. And I'm still none the wiser as to which of these options holds for this camera, which, has to be said, from what I've seen makes amazing footage.*

I may not know much on this topic, but I do know what I don't know, and will continue to ask.*

Best regards

Grazie

Hi Grazie

When is a four third sensor not a four third sensor!

The four thirds system is not the shape or size but the spec and the size and shape can change within that spec. Specifically the DIAGONAL which can be 22.5mm on the 17.3 x 13mm which is a 4/3 aspect ratio. So if you physically lower that DIAGONAL 22.5mm to create a 16/9 the sensor on paper is wider and thats what they build.

However the suspicion is they have simply cropped a 17.3 x 13mm sensor and not made a wider but either a shorter sensor or just not using all the sensor area, or the 22.5mm diagonal spec. From all the evidence so far from Panny and reps it seems that's what we have.

Maybe they could make an anamorphic mode on it that could use the whole of the cropped sensor for max quality and would be compatible with a larger range of lenses than increasing the sensor size for 16/9. Wow that'd be something.

You don't need to physically mask the sensor, as you would a 35mm motion picture gate, they could just read out the 16:9 slice and leave the rest of the sensor unconnected, so that there isn't a full 4x3 option.

You'd need an anamorphic lens to use the full 4x3 as a 16:9 aspect ratio.

Well, I listened ... and I read your comments Grazie.

First, let me point out that Jan Crittenden is for all intents and purposes 'the horse's mouth' -- she has a firm grasp of her product and engineering, and though her answer was actually to a question a little different than the one you pose, I have no doubt that her answer was informed.

Second, and related -- Jan is a very active forum participant at another site; if you were to ask a technical question she would answer.

Third, I don't have a problem reconciling her comment with the slightly generic explanation offered in the Panasonic site I referenced earlier -- the one that shows a 4/3 sensor delivering 4:3, 16:9 and 3:2 from the same chip. If someone were to summarize that as offering a '16:9 slice from a 4/3 sensor' I'd say that was accurate ...

Fourth, I still don't really understand your question. Are you positing that the area of the sensor as used is NOT 4/3 standard? That would have implications for the lenses -- it would mean that the image in terms of depth of field, angle et cetera would not be the same in an AF100 when compared to a shot from a GH2 in 16:9 mode -- a suggestion that the camcorder would not follow the 4/3 standard. You seem to be asking if the camcorder is somehow anamorphic -- using the 'whole 4:3' sensor ... but that's just impossible. You can't have an anamorphic sensor like that without also having an optical lens creating a matching anamorphic image.

So I find I'm still where I was posts ago: A little unclear on your question, and suspecting the answer can found in the Panasonic brochure I linked.

Cheers,
GB

Quote by R Geoff Baker

the one that shows a 4/3 sensor delivering 4:3, 16:9 and 3:2 from the same chip. If someone were to summarize that as offering a '16:9 slice from a 4/3 sensor' I'd say that was accurate

I think there could be confusion in calling it a 4/3 sensor. Could be better to call it a four thirds sensor. That is what I wanted to know it is a slice from a four thirds sensor that has a 4/3 aspect ratio 17.3 x 13mm and not a 16/9 four thirds sensor.

Brian
Yes your right You'd need anamorphic lenses.

Yes, I did and have read her credentials.

Would you care to share which site that is?

It's the word "slice" that I'm confusing myself over.

I'm still confusing myself as to how one gets a 16:9 area from this sensor? How is it done? If you, Pannsonic and the rest of the team here have got it - good. I'm asking simple HOW this is done and still use the full area - or doesn't it? I used the word mask, one of our colleagues here says that it isn't a mask, rather that in those areas outside the 16x9 shape, the receptors wouldn't be connected. Sure, I can understand that, no problem.

Previously and in an attempt to understand this I was considering that ALL of the area was being used and that somehow the data from that sensor was being adjusted to fit a 16x9 shape. I was trying to fit a PINT into a HALF PINT pot. Can sensors be arranged in different ways to provide various shape configurations? Stranger things have happened.

So, and following the line of discussion, this would mean that not ALL of the sensor is being used, but what IS being used fulfills the HD convention, which for this sensor is what?

Grazie

Some Digibeta cameras use different areas of the sensor for shooting 4x3 and 16:9. The former doesn't use either side of the 16:9 sensor, the result being a narrower field of view when shooting 4x3.

This was a feature that was used during the transition from 4x3 to 16:9, now everything is shot using the latter.

OK. Take a piece of paper and draw a circle. Then draw a rectangle such that the corners are just outside the circle. The circle represents the coverage area of the lens, the rectangle represents the whole sensor area.

Now, you can draw a whole series of rectangles of different shapes (4:3, 16:9 etc) but all with their corners touching the circle. These represent the photosites that need to be read out to give the desired aspect ratio - no need for masking, it's just a case of selectively reading or not reading from the chip. The most obvious point that follows is that the diagonal dimension is ALWAYS the same - regardless of aspect ratio. It also should be obvious that some photosites will always be used (the ones nearest the sensor), others never will (the ones at the corner, outside the circle).

Not quite that simple in real life, but it gives the general idea. Most immediate fact is that for the 16:9 rectangle encloses about 12 million photosites. It's extremely unlikely that all are being read every frame, more likely one in three. In a DSLR, that's what gives the aliasing, in the AF101 a good optical low pass filter stops the aliasing. It's not as good as a purpose designed video sensor (as with Sonys F3), but is a lot better than a DSLR, whilst making use of off-the-shelf components.

I have posted this link earlier, it is Panasonic's own explanation, it includes a diagram. Perhaps if you indicate what in this specific explanation fails I'd have a clearer idea where you are going.

DMC-GH2 | PRODUCTS | LUMIX | Digital Camera | Panasonic Global

The Flash diagram that steps through three different aspect ratios includes an example of how Panasonic does it, and the 'wrong' way to do it.

It seems rude to Chris for me to identify the 'other' forum -- a Google of her name quickly turns up her posting habits.

Unfortunately you are continuing the confusion in your own post:
4/3 = four thirds, not four to three
4:3 = four to three, not four thirds

A 4/3 sensor can be configured to have any aspect ratio, including 4:3, 3:2 & 16:9 -- the lens creates a circle, the aspect ratios are selected from an area within that circle.

HTH
GB

I beg to differ 4/3 or 16/9 have been used historically as ways that describe the 4:3 16:9 Aspect ratio. Maybe wrongly but never the less this is used through history. Certainly here in the UK.. NOW a new spec(Four thirds) for a digital sensor is using the same desctriptive term of 4/3 that many may or could associate with 4:3 and causing CONFUSION.

Just trying to help.

FOUR THIRDS SPEC
Rotate a DIAGONAL LINE 22.5mm with a hole in the middle rotate so the outside edge makes a circle Draw a 4:3 aspect ratio on its edges
NOW draw a 16:9 aspect ratio on its edges
Look at what youve drawn and you can see the 16:9 is wider.

HOWEVER
The camera uses the 4.3 aspect ratio and then cuts off the top and bottom So you are not using the full sensor and it is NOT close to being the same size as 35mm film It is about halfway between Super 16mm and Super 35mm. Panasonics claim that it is almost the same as 35mm frame is WRONG.

This is how the AF100 Manual describes the chip:

My apologies if there is a regionalism at work here, though in my years in Hitchin, Hertfordshire it seemed to me the convention was followed -- a ratio was expressed with a : or occasionally (vernacularly) with an x as in 4x3 read as 'four by three' ... a / represented a divisor, so:
4/3 was four divided by three or four thirds
4:3 was four to three or four units compared to three units
4x3 was four by three, typically four somethings by three somethings -- so four inches by three inches as an absolute measure.

It is true that the American photographic convention of 8x10 was reversed to 10x8, but the American convention didn't follow through all examples so a small print was a 3x4 more often than a 4x3 for example.

But never, except in error, was a 4:3 ratio written as 4/3.

Again, if your corner of Herts was different my apologies, though I'd suggest the time has come to stand on the convention that expresses the least confusion. The 4/3 format is a fraction, as in four thirds of an inch, and may use 4:3 as a ratio, or 16:9 or 3:2 or, though I know of no examples, any other ratio as they will all fit within the circle cast by a lens designed to cover a 4/3 of an inch.

Cheers,
GB

You express yourself with clarity, but I see no reason to believe accurately. Panasonic's documentation for the GH2 is very clear in this regard, and at odds with your description. In the GH2, the same imaging chip uses more pixels in width for the 16:9 than for the 4:3 -- the correct way.

As for 'using the whole chip' I can't see how this is important or significant. The size of a 4/3 imaging chip at 16:9 is an 'absolute' -- if the designers choose to derive it from something larger, that seems neither here nor there, so long as they get it right.

As for the proximity to Cine 35 in size, there does indeed seem to be some hyperbole here. The real numbers are these:
Super 16 is 12.4mm x 7mm if a 16:9 ratio is used;
4/3 is 17.8mm x 10mm in 16:9 mode;
Cine 35 is 21.2mm x 11.9mm in 16:9 mode.
So 4/3 is twice the area of Super 16 but only seventy percent of the area of Cine 35.

Cheers,
GB

Eh?

Isn't a 1/3 receptor a surface area spec? The same as 1/3 of a square inch. And 1/4 is one quarter of a square inch. 2/3 is two thirds of a square inch area. None of this describes the shape, purely surface area. We need convention ratio to then define the shape. Here that would be 4:3 or 16:9 .

So a 4/3 is four thirds of a square inch area, which I could say was one square inch PLUS another third of a square inch?

Right or wrong?

Grazie

Well, wrong.

The convention is to describe a diagonal corner to corner length, so a 1/3 inch chip is one third of an inch, diagonally. A 4/3 is four thirds on an inch, corner to corner. You can see that this means the dimensions in real terms are different for every aspect ratio -- which is why I again refer you to Panasonic's diagrammatic example. It is clearly demonstrated, with pictures, how it works.

In no case is format size using 'area' -- in video the standard is diagonal width. So in fact, as demonstrated in the picture and stated here -- the format describes a circle. The aspect ratio is a 'slice' from that circle, and each aspect ratio uses a slightly different slice ...

Please look at the diagram I have linked repeatedly, and ask questions in reference to that so we can speak the same language ...

Cheers,
GB

Yes, it's not the standard 35mm cine frame.

In practise, since this seems to be what everyone is getting excited about you should be able to achieve a DOF pretty similar to most 35mm productions by opening your lens an extra stop over to that used on most 35mm motion pictures. So if you use a stop on the AF100 of between f1.4 and f2.4 you should be in a quite common DOF zone.

You're right Grazie.. Heh heh But then in old money 2/6 might mean something else. The / can be a way to join something as a part of and in relation to. Sorry just my interpretation.

4:3 aspect ratio (4/3 = 1.33, medium shape)

I'm happy to agree to disagree Some may see clearly that 4/3 or 16/9 has no relation to 4:3 or 16:9 Indeed the former makes sense to me I have many times though seen the aspect ratio written as 16/9 Specifly in TV shops a few years back. However experiences may differ.

As regards the the GH2

QUOTE
As for 'using the whole chip' I can't see how this is important or significant. The size of a 4/3 imaging chip at 16:9 is an 'absolute' -- if the designers choose to derive it from something larger, that seems neither here nor there, so long as they get it right.

That statement could be confusing to those of us who stupidly and ignorantly confuse 4/3 with 4:3 and was the interpretation I wanted to help be understood by some not up to speed by saying it might be best to call it four thirds.
QUOTE
As for the proximity to Cine 35 in size, there does indeed seem to be some hyperbole here. The real numbers are these:
4/3 is 17.8mm x 10mm in 16:9 mode;

Numbers are aproximate my guess is its 17.3 x 13 and cropped.

You'd have to stand further back to get the same field of view if using equivilent lenses though.

Ok, this is helpful:Image sensor format

Grazie

Yes, the field of view at a particular focal length will be narrower than standard 35mm. It's really no big deal in the longer term, this can also happen when you go between different 2/3" cameras (some of which are nearly Super16), you do get used to it.

Perhaps it's more of an issue at the wider end, but that would depend on the lenses you have available, although there are a number of possibilities .

OK I hope this does not complicate things more.

In the drawing below, on the left is a 24mm x 36mm full 35mm frame, just for reference.

On the right are measurements (in mm except where shown in ").

You will notice that a 4/3 standard chip is not 4/3" in diameter. It is .804" in diameter, in this case (using measurements from the AF100 manual). This is because chip size is an old standard from the Tube days when tubes were measured to the outside of the tube, not the imaging area.

The outer circle shows the actual 4/3" or 1.33333" diameter.

It is confusing to look at and think about, but those are the standards, they are shown in the table in the second image (source Image sensor format - Wikipedia, the free encyclopedia).

Agreed, close to Cine 35 but not quite there -- Sony's APSC choice for some of their cameras is in fact closer.

What is significant I think is that it is 4/3 standard -- the same lens on any 4/3 camera using 16:9 would deliver the same result.

Cheers,
GB

The problem is your not getting the same angles either. Say you shoot an OTS you have to stand back a lot further to get the same dof. This might be a problem in a small room for example and is a compromise and not the idea that's being sold. It's not the same similar or close to S35. But none the less will give you something in between 16mm and 35mm and that will match many more 16mm lenses and turbo charge what they can do in terms of dof.

Note that the chart is not standardized to 16:9, so the numeric values are 'off' as they seem to assume 4:3 -- again, I suggest readers look at the Panasonic pictures I have linked in this thread as they include examples and values for 16:9

Cheers,
GB

I'd also like to know if the camera line skips.

Most film productions aren't shot on Super 35, although it's becoming the size for digital sensors rather than the old standard 35mm.

I think you can only regard a format for what it actaully is and quite few people were asking for this particular format in the early RED days, so I suspect there's a demand for it.

Personally, I tend to work by angles of view, it tends makes more sense in story terms. rather than obsess over working to a very shallow DOF for a shot. I'd worry more over this if the character(s) feel isolated and alienated, which can to be the effect in story terms of a very shallow DOF. Certainly the AF100 does allow a useful DOF option range, which people can creatively use.

In the wide shots, framing, composition and action are usually more important.

I "think" the sensor size is perfect for me if as I suspect my 16mm superspeeds will work with this.

However most will need to buy dedicated lenses as ordinary still 35mm lenses will have a 2x zoom making wides difficult.

My problem is possible line skipping and 8 bit out make it all a little questionable for me. I really need more info on this.

Agree.

For me shallow dof is useful to create a pretty C/U (Ooops did I just use the /) ... and to separate the background and define the intent and focus the eye on the target. I was watching the news last night and far more interested in what some kids were doing in the clean background rather than the presenter! Anyway it's just a tool but it's the tool that up till now has separated Pro's from consumers. and many consumers in their long long wait are over compensating. But a Camera is only a small part of making a film and a good DP can use a camera without shallow dof and still make a masterpiece.

So there mushing (technical term) a 12MB matrix into a 1920 x 1080 2MB matrix, as with DSLR. But apparently with a heavier low pass filter, as well as other electronic magic. The camera has the advantage of not needing to make very high res still photos.

It may not be the exact chip in the GH2, but I bet it's very close.

It makes me wonder how good a purpose built chip will be? Does a purpose built chip even need to be bayer? Is the AF100 bayer?

You can be pretty certain it does. Referring back to the GH2 link (and it's likely to be the same or very similar sensor) - DMC-GH2 | PRODUCTS | LUMIX | Digital Camera | Panasonic Global - it refers to it being able to have a burst speed of 5fps at 16 megapixel and 40fps at 4 megapixel. That's logical enough - it can get all the data off the chip 5 times per second, but not 40 times per second. Move to true video, and up to 60fps, and the same limits are likely to apply - it's only likely to be able to read a quarter or less of the total number of pixels. (I actually suspect it will be reading a third of the 12 megapixels in the 16:9 centre section, again 4 megapixel.)

Why does it worry you? With the right OLPF, pixel skipping is not the problem it is in DSLRs. The worst implication is likely to be reduced sensitivity compared to if it wasn't used. I suspect that is why it doesn't seem to be anything like as noise free as the F3 seems to be. And that is likely to mean the lack of 10 bit doesn't really matter, 10 bit only means a lot in a very noise free camera, so whilst the F3 may well benefit from it, the AF101 probably won't.

David

My current workflow is Sony EX1 Letus adapter and 35mm still lenses using the 10 bit out to a Ki Pro editing in vegas by proxy and colour correcting and rendering in 32 bit DPX files in After effects. This maintains the highest quality.

My interest in what the AF100 can do is would it be a step up or a step down and so I'm trying to evaluate what it can do here in this forum as I expect are many others as I and most people cant afford to just buy one and then decide.

Mark, I use an EX1R and an EX3 with NanoFlash. I also Have both AJA Kona and Matrox MX02 in my edit suites.

I have captured 10 bit into my Kona from SDI in the EXcams, for chromamkeys. And now that I have been using my NanoFlash for more than a year, I have not noticed any huge difference in chroma keys 8bit (NF) vs 10bit (Kona). I think the EXcams are fairly noisy so the benefit of 10bit is IMHO not great. The 422 color space is a significant upgrade though.

I don't think you will really have an answer if the AF100 is an upgrade to your SDOF until we have some more footage available. And more testing has been done. I do think it will be a lot more ergonomically workable than a ground glass adaptor on an EXcam.

I have ordered the AF100 and I will be doing a lot of comparos EXcams to AF100. And I will post my findings here.

I hope personally to be able to mix EXcams and the AF100 in my productions the way I have done with my 5DmkII and 7D. I really see the AF100 not as a replacement for my EXcams but a much better 7D. I think even with all its problems the 5DmkII is still in a class of its own. I love shooting with it, the 7D not so much.

These are my personal feelings and observations, your "smilage" may vary depending on your shooting style and clients.

Thanks Olof

I'm looking forward to seeing some of your findings!

Best

Mark