My wish for a different FZ1000, Call it FZ8000

Waltersmatthews

Forum Enthusiast
Messages
307
Reaction score
61
The sensor in the FZ1000 is 20 MP on a 1 inch sensor. I own one, it is a great camera and I love it so lets get that out of the way from the beginning.

When we shoot 4K with the FZ1000 it uses only 8MP from the center of the sensor and the signals from those 8MP's feed directly into the processor. Therefore the light procession capability of the camera, at 4K is the same as it is at 1080P. The signal to noise level at 4K is teh same as it is at 1080P which is to say OK but not really very good. These days ISO 1600 is pretty marginal for both stills and for video.

If Matsushista, AKA Panasonic were to design a 1 inch 8MP sensor and use that "reduced resolution" sensor in the current camera with the current electronics of the FZ1000(obviously there would have to be detail signal level clean up) several things would happen.

First the low light capability of the camera for Still Photography and for Video would improve dramatically. Each pixel would be 2.5X larger than in the current FZ1000 and that would translate into dramatically improved low light performance.

The current lens and how it operates (other than turning off the lens retraction after 15 seconds which I personally want) would work perfectly with the 8MP 1" sensor.

With improved low light performance we could shoot at a minimum of 2 stops or perhaps 3 stops less light and still get great photos which would allow us to shoot at, for example, 250 or 500 of a second where we now have to use 60 or 125 of a second. That would translate into sharper photos and better quality video.

Sony did just this in their Alpha 7S and that camera delivers remarkably good video and stills with very low light capability.

Finally, if they were to design and have manufactured an 8 MP 1" sensor the geometries of the sensor would be significantly bigger which would allow for less error in its parametric performance and less sensitivity to defect levels which would translate into higher yield.

Now, I know the pixel peepers will scream that they don't want to give up their 100% looks at the images so, hey, don't buy the camera but to those of us who make our living shooting images and making videos the camera would be near perfect. The 4K video feed from such a camera could have very fast shutter speeds and therefore a lot more useful video feed for 8MP stills. Moreover, in shooting more than 1000 weddings, 500 of them digitally, there has never been a situation, other than the formal family photos, where I have had any need for a higher pixel count than 8MP.

Unfortunately I believe Sony supplies the FZ1000 sensor to Panasonic but who knows, Matsushista could make such a sensor if they decided to. They have several high functioning and well engineered manufacturing areas that could turn out such a chip.
 
OK, round 24, FIGHT!
If Matsushista, AKA Panasonic were to design a 1 inch 8MP sensor and use that "reduced resolution" sensor in the current camera with the current electronics of the FZ1000(obviously there would have to be detail signal level clean up) several things would happen.

First the low light capability of the camera for Still Photography and for Video would improve dramatically. Each pixel would be 2.5X larger than in the current FZ1000 and that would translate into dramatically improved low light performance.
This is not true: http://photography-on-the.net/forum/showthread.php?t=706255

And there are plenty of examples that you can see for yourself in dpreview's studio comparison tool that demolish this myth as well.

DxO Labs DxOMark database is another resource you can peruse to the same end.

In both cases all you need to do is compare sensors of the same size and technology generation that have different pixel counts at the same display size (or in the case of DxOMark look at the normalized SNR values).
Sony did just this in their Alpha 7S and that camera delivers remarkably good video and stills with very low light capability.
Actually there's very little difference between the 12MP A7s and the 36MP D810 at any ISO you're likely to be using once you normalize the print/display resolution: http://www.dpreview.com/reviews/ima...=-0.005675368898978415&y=0.004920621921385182

DxO Labs test results confirm this: http://www.dxomark.com/Cameras/Compare/Side-by-side/Nikon-D810-versus-Sony-A7S___963_949

We're talking less than a stop.

But just as important as anything else is detail resolution. When you have more detail in a photo courtesy of having captured an image with a greater number of pixels, noise reduction software can do a better job of reducing noise while preserving that detail for a given print size. In other words more resolution can offset the higher read noise that is sometimes generated by greater pixel densities.

 
How bout a (1") 24 mpx sensor for still shooting and then using groups of 4 for 8mpx video ???

But I want the lens to start @ 20mm WA, (w/ a UWA convertor to get to 12-14mm equivalent).

And starting @ f/1.4 with f/2.8 at least through 200mm towards 400 @ f/4. (To match/better the RX-10.)
 
LOL! Sounds familiar no?
 
How bout a (1") 24 mpx sensor for still shooting and then using groups of 4 for 8mpx video ???
I'm pretty sure that the main reason that no-one is doing 4K by downsampling full high resolution sensor output, or via 2x2 pixel binning or whatever (which still requires that every pixel be "read"), is because of sensor read-out speed limitations (and probably thermal issues as well). But if this wasn't the case what you're suggesting would be pretty awesome I think, at least for videography enthusiasts.

Of course if the FZ1000 was actually a video camera rather than a stills camera with video functionality then it might indeed make sense to utilize a native 8MP 1"-type sensor.
 
Last edited:
You're asking too much for that price range, feel free to buy a RED or its competitors, oops $15-30K.
 
Sony did just this in their Alpha 7S and that camera delivers remarkably good video and stills with very low light capability.
>>Actually there's very little difference between the 12MP A7s and the 36MP D810 at any ISO you're likely to be using once you normalize the print/display resolution: http://www.dpreview.com/reviews/image-

Ah yes, this time lets compare apples to peaches rather than to oranges.

Lets start with the DPR comparison

http://www.dpreview.com/articles/4613822764/high-iso-compared-sony-a7s-vs-a7r-vs-canon-eos-5d-iii

and further lets look at Sony versus Nikon in the same generation of camera manufacturing. As we march forward in time each successive generation of sensor from specific vendors continues to improve in low light capability and specifically with a Nikon versus Sony comparison where both vendors use the Sony sensor, Nikon consistently beats Sony in low light performance. Therefore when you take a full two years of development with Nikon and factor in the "fact" that Nikon consistently does better in low light than Sony, your choice of what to compare is bogus. The comparison in the DPR review of the Sony 7S to the 7R with the Canon 5DIII thrown in for good measure was done with cameras of nearly the same generation and the Sony to Sony comparison is, in my opinion the best, indicator of comparative low light performance.

I have no idea what you are looking for in a camera but for my purposes in wedding photography, where I often have to shoot in poorly lit sanctuaries, where the local flash Natzi aggressively forces me to not use flash and marginally lit reception halls, low light performance-that is high-ISO performance, without flash, is of critical importance. While you can set the FZ 1000 to ISO 3,200; 6,400; and even 12,800; my evaluation of those settings has suggested to me that they are just too noisy to be of much use.

In your reply you acknowledge a possible 1 stop of improvement and for my needs one stop would be great. The situations I have described can be shot with good results at ISO 3,200, which is where I shoot but not with the FZ1000. My Nikon DSLR's perform very nicely at that ISO and an occasional move to ISO 6,400 even works out pretty nicely albeit with a lot more work in post processing.

The DP review comparison of the 7S, the 7R and the Canon 5DIII suggest that there is a 2 stop advantage in very low light situations and nothing you referenced suggests that would not be true of a "larger sensor" compared to a "smaller sensor" which is what you would get with a 8MP 1" sensor compared to the significantly reduced size 8MP area sampled on the LZ1000 when recording 4K video.

My LZ1000 was being shot in 4K by my assistant while I did what I call the formals. i.e the group shots of family and bridal party. I used the video feed to generate individual frames and the results are nothing less than spectacular. Perfect focus, all eyes open and all but one smiling which was better than the stills I made with the D800E. The LZ of course had lower resolution but a lot better timing.

So in summary, I believe the existing data shows that an 8MP 1" sensor would produce better overall performance in low light, probably ISO 3,200 or even 6,400 and it would be particularly true for 4K video which is where my primary interest lies. Moreover 8MP images from a 1" sensor are more than good enough for my needs for still photography. Indeed they would be a lot lower resolution that wnat my D800E or Pentax 645D or Canon 1DsIII can deliver but none of my customers could even see a difference and even if they could they would not care. A 6 MP camera can produce stunning 18x24 images.
 
Sony did just this in their Alpha 7S and that camera delivers remarkably good video and stills with very low light capability.
>>Actually there's very little difference between the 12MP A7s and the 36MP D810 at any ISO you're likely to be using once you normalize the print/display resolution: http://www.dpreview.com/reviews/image-

Ah yes, this time lets compare apples to peaches rather than to oranges.

Lets start with the DPR comparison

http://www.dpreview.com/articles/4613822764/high-iso-compared-sony-a7s-vs-a7r-vs-canon-eos-5d-iii

and further lets look at Sony versus Nikon in the same generation of camera manufacturing. As we march forward in time each successive generation of sensor from specific vendors continues to improve in low light capability and specifically with a Nikon versus Sony comparison where both vendors use the Sony sensor, Nikon consistently beats Sony in low light performance. Therefore when you take a full two years of development with Nikon and factor in the "fact" that Nikon consistently does better in low light than Sony, your choice of what to compare is bogus. The comparison in the DPR review of the Sony 7S to the 7R with the Canon 5DIII thrown in for good measure was done with cameras of nearly the same generation and the Sony to Sony comparison is, in my opinion the best, indicator of comparative low light performance.

I have no idea what you are looking for in a camera but for my purposes in wedding photography, where I often have to shoot in poorly lit sanctuaries, where the local flash Natzi aggressively forces me to not use flash and marginally lit reception halls, low light performance-that is high-ISO performance, without flash, is of critical importance. While you can set the FZ 1000 to ISO 3,200; 6,400; and even 12,800; my evaluation of those settings has suggested to me that they are just too noisy to be of much use.

In your reply you acknowledge a possible 1 stop of improvement and for my needs one stop would be great. The situations I have described can be shot with good results at ISO 3,200, which is where I shoot but not with the FZ1000. My Nikon DSLR's perform very nicely at that ISO and an occasional move to ISO 6,400 even works out pretty nicely albeit with a lot more work in post processing.

The DP review comparison of the 7S, the 7R and the Canon 5DIII suggest that there is a 2 stop advantage in very low light situations and nothing you referenced suggests that would not be true of a "larger sensor" compared to a "smaller sensor" which is what you would get with a 8MP 1" sensor compared to the significantly reduced size 8MP area sampled on the LZ1000 when recording 4K video.

My LZ1000 was being shot in 4K by my assistant while I did what I call the formals. i.e the group shots of family and bridal party. I used the video feed to generate individual frames and the results are nothing less than spectacular. Perfect focus, all eyes open and all but one smiling which was better than the stills I made with the D800E. The LZ of course had lower resolution but a lot better timing.

So in summary, I believe the existing data shows that an 8MP 1" sensor would produce better overall performance in low light, probably ISO 3,200 or even 6,400 and it would be particularly true for 4K video which is where my primary interest lies. Moreover 8MP images from a 1" sensor are more than good enough for my needs for still photography. Indeed they would be a lot lower resolution that wnat my D800E or Pentax 645D or Canon 1DsIII can deliver but none of my customers could even see a difference and even if they could they would not care. A 6 MP camera can produce stunning 18x24 images.
But what is wrong with my suggestion of a 24mpx w/ pixel-binning (groups of 4) to get the 8mpx ???

For still images there could be an option to select either the (lower noise) 8mpx or full-24mpx for "poster" size enlargements.

HTC of course tried your concept on cell-phones with their "super-pixels" and it does offer some advantages -- but overall it's camera performance does not match the higher mpx cameras on other cell-phones.

--
Thanks for reading .... JoePhoto
( Do You Ever STOP to THINK --- and FORGET to START Again ??? )
 
Last edited:
lets look at Sony versus Nikon in the same generation of camera manufacturing
Technically that's not what you're doing. The Sony sensor in the A7r is essentially the same sensor that was in the Nikon D800 that went on sale in early 2012. The only real difference is that the newer 36MP sensors do not have an optical low-pass filter (OLPF) in front of them.

In other words, the sensor in the a7s is newer technology.
In your reply you acknowledge a possible 1 stop of improvement and for my needs one stop would be great.
Nope. It's definitely less than a stop, even if you compare the A7s to the A7r. And you don't get it for free either. In "tuning" the 12MP A7s sensor for low-light performance, Sony had to sacrifice dynamic range and color sensitivity in the ISO100-400 range (check the DxO measurements for yourself). And there doesn't appear to be any particular reason why you couldn't make the same sort of trade-off with a higher megapixel sensor.
and nothing you referenced suggests that would not be true of a "larger sensor" compared to a "smaller sensor" which is what you would get with a 8MP 1" sensor compared to the significantly reduced size 8MP area sampled on the LZ1000 when recording 4K video.
Sure, an 8MP 1"-type sensor would be better for video. But then you're sacrificing stills resolution. The FZ1000 is a stills camera first, and if anything needs to be compromised, it shouldn't be stills performance. There are plenty of high-end 4K camcorders if that truly is your priority.
So in summary, I believe the existing data shows that an 8MP 1" sensor would produce better overall performance in low light, probably ISO 3,200 or even 6,400
For video, sure (because of the effective sensor size issue). For stills, absolutely not.
and it would be particularly true for 4K video which is where my primary interest lies.
Again there are plenty of high-end 4K camcorders available. Leave our stills performance alone.
 
Last edited:
>But what is wrong with my option of a 24mpx w/ pixel-binning (groups of 4) to get the 8mpx ???

Nothing and I don't know the details of what is going on with the LZ100 but I suspect that they are maxed out with what they can process with direct feed of the reduced field 8MP part of the sensor and while pixel binning would certainly accomplish the same thing, You have added a lot of processing overhead.

>For still images there could be an option to select either the (lower noise) 8mpx or full-24mpx for "poster" size enlargements.

Indeed and with still images they are not under such a data crunch per unit time constraint and by reducing the FPS, they could accomplish the data crunch task. I would imagine it could be done with a firmware upgrade as well.

>HTC of course tried your concept on cell-phones with their "super-pixels" and it does offer some advantages -- but overall it's camera performance does not match the higher mpx cameras on other cell-phones.

Hard to compare images from vendor 1 with vendor 2 without knowing what they really are doing.
 
>Nope. It's definitely less than a stop, even if you compare the A7s to the A7r. And you don't get it for free either. In "tuning" the 12MP A7s sensor for low-light performance, Sony had to sacrifice dynamic range and color sensitivity in the ISO100-400 range (check the DxO measurements for yourself). And there doesn't appear to be any particular reason why you couldn't make the same sort of trade-off with a higher megapixel sensor.

That dynamic range loss is more than made up by much higher DR at higher iso's that the A7r

Quoting from DP review on the comparison..."For example, at ISO 409,600 the shadows and darker midtones of the A7S show a nearly 2 EV advantage over the 5D Mark III and A7R - where, at web resolutions (5 MP), the noise levels of the A7S look visually similar to ISO 102,400 on the other cameras.*3 This advantage drops to 1 EV or less in brighter tones, which are recorded as higher signals at the sensor and are, therefore, less dependent on the electronic performance of the sensor."

>Sure, an 8MP 1"-type sensor would be better for video. But then you're sacrificing stills resolution. The FZ1000 is a stills camera first, and if anything needs to be compromised, it shouldn't be stills performance. There are plenty of high-end 4K camcorders if that truly is your priority.

You are ignoring the projected high ISO improvement you would get for stills.

>Again there are plenty of high-end 4K camcorders available. Leave our stills performance alone.

At what cost. An 8K sensor cold be slipped into the FZ1000 and it could be done as a cost reduction to the current camera as the sensor yield would go up, no other real modifications would have to be done and therefore the overall cost down (albeit a very small cost reduction as they are already at very high yield on that size sensor.) In very low light the modified LZ1000 would be a better performer than what we have now. Why not let us have a choice.
 
You are ignoring the projected high ISO improvement you would get for stills.
You're missing the point.

1) the sensor in the a7s is newer technology

2) the sensor in the a7s is "tuned" for high ISO performance at the expense of low ISO performance, and there doesn't seem to be any reason why you couldn't do the same thing with a higher megapixel sensor

3) at any ISO the average person is likely to be shooting at (~6400 ISO or less) performance is near enough to identical anyway, particularly when you factor in the benefits of having more pixels on your subject

The bottom line is that a lower resolution sensor would not produce better results at extremely high ISOs because it was a lower resolution sensor. Regardless of the resolution it would produce better results at extremely high ISOs if you traded away dynamic range and color sensitivity at low ISOs in order to get there. It's no accident that the Nikon D4s makes the exact same compromises.
An 8K sensor cold be slipped into the FZ1000 and it could be done as a cost reduction to the current camera as the sensor yield would go up, no other real modifications would have to be done and therefore the overall cost down (albeit a very small cost reduction as they are already at very high yield on that size sensor.) In very low light the modified LZ1000 would be a better performer than what we have now.
For video, sure (because of the effective sensor size issue). But not for stills.
Why not let us have a choice.
How is this a choice? You want to force a particular compromise because you care more about video than stills, and more about performance at extremely high ISOs than performance at low ISOs (which are the ones most people try to shoot at). What about everyone else?

Again, if you want a 4K video camera that puts video performance above everything else then you can have one. But it's not the FZ1000. And it shouldn't be.

Of course this is all moot anyway as the FZ1000 is already here and it is what it is ;)
 
Last edited:
OK, round 24, FIGHT!
Thought you'd appreciate seeing that PetaPixel is helping to perpetuate this "myth". Here's a blurb from PetaPixel's report on the Photokina announcement of the new LX100:

And, finally, the larger sensor and low resolution helps in the low-light department, boosting the max ISO to 25,600 compared to the 12,800 of its predecessor, the LX7.

 
OK, round 24, FIGHT!
Thought you'd appreciate seeing that PetaPixel is helping to perpetuate this "myth". Here's a blurb from PetaPixel's report on the Photokina announcement of the new LX100:

And, finally, the larger sensor and low resolution helps in the low-light department, boosting the max ISO to 25,600 compared to the 12,800 of its predecessor, the LX7.

I wonder if by "low resolution" they mean lower than a "normal" 4/3" sensor? People made the same mistake regarding the LX7 (and presumably other LX models) thinking that Panasonic had put a "special" low resolution (10MP) sensor in it to keep noise down when in fact what they were really doing was utilizing a smaller portion of a 12MP sensor (making the effective pixel density higher than it seemed at first blush).

Same thing here. The LX100 doesn't have a special "low resolution" 12.8MP 4/3" sensor. It utilizes a smaller portion of the same 16MP 4/3" sensor that Panasonic use in their MFT camera line, so the pixel density is effectively exactly the same.
 
>Thought you'd appreciate seeing that PetaPixel is helping to perpetuate this "myth". Here's a blurb from PetaPixel's report on the Photokina announcement of the new LX100:

So, first, you when you use the term myth, you et.al.are suggesting that both Sony with the Alpha7is and Panasonic with the LX100 really don't know what they are doing and are just selling marketing hype when they make lower resolution, and therefore larger pixel size, sensors and in doing so advertise they have enabled higher ISO shooting as a consequence. Even in the face of the DP review of a comparison of the 7s to the 7r and the Canon 5DIII the larger pixel better low light performance is a "MYTH?"

Next, what is the explanation of why the Canon 5DIII at 21MP is at least 1 stop better than the Nikon D800 or the D800E at low light photography. Please don't quote me external reviews that might say something different as I own both cameras and my direct experience in many weddings tells me that when the light is low and I have to get a photo without flash the 5DIII is a lot better than the D800 and they are of the same generation technology, albeit from different manufacturers. Perhaps because Canon is just better at signal to noise processing than Nikon but I doubt that because in comparisons of the low end Rebel series to the low end Nikon DSLR's of the same generation where MP counts are comparable, Nikon consistently trumps Canon in low light performance.

In any event, arguments about how many angels can balance on the head of a pin it looks like Panasonic might be ready to deliver what I asked for albeit in a different camera format. What I can't tell yes is whether they are generating the 4K video from the entire sensor area which would be great or if they are staying with the 8MP crop they used in the LZ1000. It looks like it will be a very worthy competitor and answers many of the complaints about the LZ, namely size. I'm not sure I want to give up eye level viewing but who knows....?????
 
>Thought you'd appreciate seeing that PetaPixel is helping to perpetuate this "myth". Here's a blurb from PetaPixel's report on the Photokina announcement of the new LX100:

So, first, you when you use the term myth, you et.al.are suggesting that both Sony with the Alpha7is and Panasonic with the LX100 really don't know what they are doing and are just selling marketing hype when they make lower resolution, and therefore larger pixel size, sensors and in doing so advertise they have enabled higher ISO shooting as a consequence. Even in the face of the DP review of a comparison of the 7s to the 7r and the Canon 5DIII the larger pixel better low light performance is a "MYTH?"
Um, no - I quoted the word "myth" because I was responding to cainn24, who I've already sparred with along these lines. The fact that I quoted the word should have been a fairly strong hint that I am not yet fully convinced and I am not in any way claiming what you presuppose above.

And as we see, PetaPixel may have been wrong anyway as the new LX100 uses a standard 4/3 sensor but uses portions of it to support the LX's normal multi-aspect ratio feature, not necessarily larger pixel sizes. Making assumptions based on resolution specs without understanding the underlying technology can easily lead to this type of error. To be fair, it's a brand new camera just announced today, so full disclosure is yet to come.

--
Bruce
You learn something new every time you press the shutter
 
Last edited:
So, first, you when you use the term myth, you et.al.are suggesting that both Sony with the Alpha7is and Panasonic with the LX100 really don't know what they are doing and are just selling marketing hype when they make lower resolution, and therefore larger pixel size, sensors and in doing so advertise they have enabled higher ISO shooting as a consequence.
First of all the LX100 doesn't have a lower resolution sensor: http://www.dpreview.com/forums/post/54388502

Second of all no-one is denying that the Sony a7s set a new benchmark for low-light performance at extremely high ISOs. What I was trying to explain to you was that they didn't get there simply by increasing the size of the pixels on the sensor. Low ISO performance was also compromised a result.

I'm not sure if you realize this or not but the real reason that Sony put a 12MP sensor in the a7s was so they could have 1:1 pixel mapping when shooting 4K video with a 16:9 AR.

The new trend in video is to avoid line skipping and pixel binning to prevent moire and aliasing and you simply can't do that with a high resolution sensor.
Even in the face of the DP review of a comparison of the 7s to the 7r and the Canon 5DIII the larger pixel better low light performance is a "MYTH?"
The better performance isn't a myth. It's real (as is the degraded performance at low ISOs). It is simply your interpretation of how the performance gains are achieved that is in question, and also the real-world relevance given the ISO range that most people shoot within.
In any event, arguments about how many angels can balance on the head of a pin it looks like Panasonic might be ready to deliver what I asked for albeit in a different camera format. What I can't tell yes is whether they are generating the 4K video from the entire sensor area which would be great or if they are staying with the 8MP crop they used in the LZ1000. It looks like it will be a very worthy competitor and answers many of the complaints about the LZ, namely size. I'm not sure I want to give up eye level viewing but who knows....?????
The best case scenario is that they do things the same way they do them on the GH4, at least in terms of mapping sensor pixels to video pixels.
 
Last edited:
Next, what is the explanation of why the Canon 5DIII at 21MP is at least 1 stop better than the Nikon D800 or the D800E at low light photography. Please don't quote me external reviews that might say something different as I own both cameras and my direct experience in many weddings tells me that when the light is low and I have to get a photo without flash the 5DIII is a lot better than the D800 and they are of the same generation technology, albeit from different manufacturers. Perhaps because Canon is just better at signal to noise processing than Nikon but I doubt that because in comparisons of the low end Rebel series to the low end Nikon DSLR's of the same generation where MP counts are comparable, Nikon consistently trumps Canon in low light performance.
I have to respond to this.

If you get better results out of your 5D then good for you. I'm not going to argue. How can I? But it's not because the Nikon has a noisier sensor when the print/display size is normalized.

http://www.dxomark.com/Cameras/Compare/Side-by-side/Canon-EOS-5D-Mark-III-versus-Nikon-D800___795_792
 
If you were to shoot both in very low light I suspect you would come to the same conclusion. Besides, I think you can actually fit 4 angels rather than 3 on the head of that pin.

I think the LX100 will be a very very good camera.
 
If you were to shoot both in very low light I suspect you would come to the same conclusion.
If I was shooting JPEG only I probably would. The 5D Mark III has some very aggressive NR applied to its JPEGs at high ISOs compared to the Nikon, as is evident from dpreview's studio shots. But that's got nothing to do with raw sensor performance.
Besides, I think you can actually fit 4 angels rather than 3 on the head of that pin.
I'll probably stop fretting about those angels right about the same time you do ;)

(or perhaps more accurately, I'll go back to fretting about other angels on the heads of other pins instead!)
I think the LX100 will be a very very good camera.
No argument there.
 
Last edited:

Keyboard shortcuts

Back
Top