Press Release: Panasonic unveils HMC150 pricing and ship date

[David Heath]

Originally Posted by Bob Diaz

Panasonic has said that the pixel offset increases the resolution by 1.5 times. If so, that says that the 960 x 520 offset sensors are like a 1,440 x 810 image sensors. My own guess is that the 1.5x factor is a bit high, but whatever it is, the 1280x720 images from the camera appear to be sharp enough. I consider the 1920x1080 images good enough, but this is subjective and others may differ.

Close, but it's more accurate to say that pixel offset may increase the *luminance* resolution by *up to* 1.5 times. The 1,440/810 figures are where resolution must have fallen to zero, the point beyond which it is physically impossible for any more detail to be resolved. In practice no system resolves detail at 100% all the way to a certain figure, then nothing, in practice it tails off as shown in mtf graphs. Red/blue contribute about 50% off the luminance information, green about the other 50%, treat them separately for resolution purposes (the whole basis of pixel shift) and it's easy to see why the mtf for fine detail will be much lower for such a sensor than for three 1MP sensors - all else equal. Maybe even more important is that the 1MP sensors will show far less aliasing.

This is even assuming monochrome images (the best case scenario for pixel offset), and true 1440x810 sensors would be capable of chrominance detail to the same figure as luminance. In practice, resolution figures are normally quoted to the figure where the mtf has fallen to a certain level. That's why (as you guess) pixel shift is more normally considered to improve performance by 1.2-1.3x, and 1.5 is, well, optimistic.

How important this all is in the real world is another matter, but it should give a foundation for realising that claims of "full horizontal resolution 1920x1080 images" from 960x540 sensors just are not backed up by theory. Not even close. And practical tests back up the theory quite closely.

The old tube cameras did have a rolling shutter, but the real problem was "Comet Tails" from bright lights in the picture.

At the risk of being flippant, Bob, I'd say there were quite a few problems with tube cameras, apart from comet tails! :) Top of my list would be registration errors, followed by low-light lag, shading errors...... But tell all this to the youth of today........

But my point was that nobody even commented on rolling shutter issues, let alone complained about them, they were there, but...... And nobody applauded, even noticed when they went away with CCDs. The absence of comet tails etc was applauded, the advent of vertical streaks was not appreciated, but rolling shutter issues? It's why I'm surprised it's receiving so much attention come the EX.

[Kevin Shaw]

Originally Posted by Tim Polster

So what it comes down to for me is get an HPX-500 and two HMC-150s or an HPX-500, HMC-150 and an EX-1.

Keying - HPX-500 vs EX-1
Weddings - HMC-150 vs EX-1

In general use, I don't think there will be that much difference between the HMC-150 and the EX-1, especially for the web.

But the price difference between the two including memory costs is about $5,000+ dollars.

It doesn't make sense to compare the HMC150 to the EX1, since the latter is a different class of camera targeted at different users. What many wedding videographers are buying is the Sony Z7U, which offers both tape and CF-based recording at affordable prices. And in general it makes sense to buy cameras from one manufacturer with as few different models as possible, so your option of mixing three cameras from two manufacturers doesn't sound too promising.

A fairer comparison might be an HPX500 with two HMC150s versus a Sony S270 with two Z7Us.

[Tim Polster]

Thanks for your reply Kevin, but I have disagree in a way.

The other Sony camera you mention costs the same as the EX-1, so it would be in a different price class as well.

But the main thrust of my post was to get opinions about the effectiveness of the models compared to each other for the two specific tasks , keying and wedding use.

I don't mind which class the cameras are considered to be in, just the performance/price ratio.

I believe we have to compare the cameras, it is part of the purchasing process.

[Kevin Shaw]

Originally Posted by Tim Polster

I don't mind which class the cameras are considered to be in, just the performance/price ratio.

My point was that the EX1 is in a higher performance class than the HMC150, so trying to compare their prices doesn't make much sense. But if you must compare the EX1 to the HMC150 then consider several key benefits of the former to determine whether the extra cost is justified:

Larger sensor with better depth of field control
Four times the resolution at the sensor level
Faster memory cards with more recording options and quicker transfer speeds
More advanced lens and focusing options
More established and efficient workflow on today's computers

I suspect the HMC150 will be popular with wedding videographers due to the low combined cost of camera and memory, but in that price range the most relevant comparison will be to the Canon XH-A1 and Sony Z1U. The latter aren't solid state cameras but are cost-effective, proven performers for wedding and event work. For those who want affordable solid state recording, the HMC150 will best be compared to the Sony Z7U.

[Tim Polster]

I guess I am working from the point that the HMC-150 is a HVX-200a or an HPX-170 with a different recording codec.

Those cameras are considered to be in a class as well.

I don't like the camera vs. camera idea either, but in all reality, it is really difficult to make camera purchase decisions.

I will be shooting at 720p for wedding work.

Will the EX-1's image be $3,000 better than the HMC-150 at this resolution?

[Kevin Shaw]

Originally Posted by Tim Polster

I will be shooting at 720p for wedding work.
Will the EX-1's image be $3,000 better than the HMC-150 at this resolution?

That depends on how the two cameras compare in low-light situations, and how you feel about the mellower depth of field of the EX1 due to the larger sensor - which puts the EX1 in a class by itself for HD cameras under $10K.

If money isn't a significant concern the EX1 should be an easy choice over the HMC150, but the latter is definitely cheaper. If price is a concern then the logical competitors to the HMC150 are other cameras in the same price range, at which point you're mainly choosing between solid state and tape-based workflow.

[Tim Polster]

Well since I will probable go with an HPX-500 as my main camera (produced projects ect...), the role of this camera choice is for weddings and second & third camera for the HPX-500.

Which is why the HMC-150 - EX-1 debate is going on in my head.

It is some way a compliment to the HPX-500, but will also be used as the main camera in a wedding and small project situation.

Which is why I thought one of each - HPX-500 & HMC-150 pair and EX-1 & HMC-150 pair for weddings. (HMC-150 as a second camera for the ceremony only). All three for large events.

Or the best case in terms of color matching is two HMC-150s (if the image quality is not that much of a tradeoff from the EX-1) and an HPX-500.

So I am not concerned about camera class, I am approaching this from a pure business use/quality point of view.

Complex, but that is the state of buying cameras today.

[Bob Diaz]

David,

It's always fun reading your messages, like me, you have been around a few years and remember the "good old days" of 20 minute 3/4" Umatic "portable" decks that I think were around 70 or so pounds with a camera. Oh joy...


The rolling shutter issue never came up with tube cameras because the shutter was a 360 degree shutter at about 1/60th of a second. A whip pan would be mostly a blur with that long of a shutter. Camera Flashes appeared part on the first field and the rest on the second field. Plus some after image may have held for a short while.

Today's CMOS cameras, like the Sony V7, appear to have a faster speed for the rolling shutter. Thus the lean caused by a whip pan is very hard to see. However, camera flashes appear different on CMOS then from a tube camera. Flashes appear as bright horizontal strips at random points in the frame.

At this point, we are getting into a subjective area, because what bothers one person, may not bother another. To me the bright horizontal strips appear unnatural and bother me. However, because this is subjective, I understand why someone can say, "It doesn't bother me."



As far as offset pixel resolution on the HMC-150: Yes, in theory, if the color in the detail area is (0 -->100% Green AND 0% Red AND 0% Blue) OR (0% Green AND 0 --> 100% RED AND 0 --> 100% Blue), the resolution would be 960 x 540.

However, I've worked enough in Photoshop to know that this is more of an exception, than typical for images. For example, I've taken a photo I shot the other week of Motorized Scooters in a parking lot. No matter what point (color) I select in the image Red, Blue, and Green always have some component greater than 0 and changes at different locations in the image. In the photo I've attached, the Red of the Scooter also shows up as some green and blue. I've tried other colors in the image and get the same thing, at least some red, some blue, and some green.


Depending on the model one wishes to use, the increase in resolution from the 960 x 540 sensors could be form 1.2x to 1.5x. The number is debatable; however, to me the increase is enough to produce a reasonably sharp 1280x720 image.


Bob Diaz

[Paulo Teixeira]

Tim Polster,

Have you ever thought about pairing up the HPX500 with the HPX170? This way, the picture quality will be much more similar to each other and not to mention the codec being the same.

[Tim Polster]

Hey Paulo,

Your correct, but I am a cheepskate and also like the long recording times of the HMC-150!

[Kevin Shaw]

Originally Posted by Tim Polster

Or the best case in terms of color matching is two HMC-150s (if the image quality is not that much of a tradeoff from the EX-1) and an HPX-500.

Yes, color matching and overall image "look" will be easier sticking with one brand of cameras, and if you don't care about 1080 resolution then the advantages of the EX1 are less. So the HMC150 fills a previous price/performance gap in Panasonic's lineup which corresponds with your situation, and you'd probably be fine with it - provided the HMC150 is good enough in low light for weddings.

Another option to consider would be to get one EX1 and two Sony Z7Us, but it sounds like you're inclined to go with Panasonic.

[Noa Put]

Originally Posted by David Parks

[url] B&H has it for $3,495.00. Good price.

I Saw the first webshop with a price and availability end September in The Netherlands and it was 4200 euro which is 6080 dollar, is the price mentioned at B & H the regular price on the otehr side of the ocean? In that case we Europeans are getting sc***ed again. No wonder you guys can afford those expensive camera's, everything is 50% cheaper were you live. :)

[David Heath]

Originally Posted by Bob Diaz

....... like me, you have been around a few years and remember the "good old days" of 20 minute 3/4" Umatic "portable" decks .... Oh joy....

Absolutely, as a recordist then, and we always had the additional weight of a sound mixer then! But as you find the comments fun........

As far as offset pixel resolution on the HMC-150: Yes, in theory, if the color in the detail area is (0 -->100% Green AND 0% Red AND 0% Blue) ..........

However, I've worked enough in Photoshop to know that this is more of an exception, than typical for images. ...... No matter what point (color) I select in the image Red, Blue, and Green always have some component greater than 0 and changes at different locations in the image.

Depending on the model one wishes to use, the increase in resolution from the 960 x 540 sensors could be form 1.2x to 1.5x. The number is debatable; however, to me the increase is enough to produce a reasonably sharp 1280x720 image.

Yes, I agree with most of that in principle. The only thing I'd add is to stress how I don't think you can have a simple number to represent resolution, it's the whole principle of mtf - how does the system deal with a change from 100% white to 0% black - how low a percentage mtf are you prepared to accept to say the system is still resolving that detail? And practically it will tail off with finer and finer detail - at what percentage do you say "that's the limiting resolution"?

And I suspect that for a given fineness of detail, the modulation factor will be a maximum if it is black-white, go down as the image becomes more coloured, and only go away in an extreme case, highly saturated colours. Hence I'd say your Photoshop observations seem to predict pixel shift nearly always having some effect on extending resolution, though unlikely to be 1.5x, certainly for any decent mtf.

Practically though, I'll agree that these chips yield a reasonably sharp 1280x720 image - it's the likelihood of any more than that I'll dispute. That's not to dispute the claims of those who say the 1080 mode of such as the HVX200 is sharper than the 720 mode - DVCProHD subsamples to 960x720 in 720p mode, 1280x1080 in 1080 mode - and I suspect what is being seen there is nothing to do with the 720/1080 factor as such, rather the 33% increase in horizontal resolution.

Which raises another interesting thought. The HMC150 and AVC-HD in 720 mode doesn't subsample - it records the full 1280x720 raster. Hence I suspect that the difference between 720 and 1080 with this camera will be far less marked than with DVCProHD. In fact, for a given bitrate, 720 may even look BETTER than 1080 as the compression will be less.

[David Heath]

Originally Posted by Bob Diaz

Panasonic has said that the pixel offset increases the resolution by 1.5 times. If so, that says that the 960 x 520 offset sensors are like a 1,440 x 810 image sensors. My own guess is that the 1.5x factor is a bit high, ........

Hello Bob! Sorry to revisit this, but I've been given (almost by accident) what I now consider the definitive answer, and I'm afraid neither of us was quite right!

The 1.5x figure I'm told is reasonable (for luminance) for any pixel shift system, but it applies to the system in total. Hence, for 960x1080 chips with horizontal pixel shift, it's equivalent to luminance chips of about 1440x1080.

The confusion with the Panasonic implementation is because it applies in 2 dimensions. The 1.5x figure is still true - but shared between the dimensions. Now, the 50% gain corresponds to an effective increase in the number of equivalent pixels for luminance - from 0.5MP to 0.75MP - and because they are shared between h and v, the effective increase IN EACH DIMENSION is the square root of 1.5 - 1.22x.

Hence it's reasonable to consider the effective luminance resolution as that of a chip with dimensions about 1170x660. Well, pretty close to 1280x720.....!? I believe some measurements have shown resolved detail a bit higher than those figures, but looking at zone plate results, I'm pretty sure they are aliases.

[Dwain Elliott]

Originally Posted by Tim Polster

Hey Paulo,

Your correct, but I am a cheepskate and also like the long recording times of the HMC-150!

Tim, I must agree with Paulo. The HPX170 seems to be the PERFECT solution to your stated objective, but only you can decide if it's worth the additional cost over the HMC150.