HPX370 - First Look
by Barry Green

What do you get when you take a brilliant but flawed camera (such as the HPX300) and fix all the flaws?

The HPX300 made a pretty big splash when it was introduced, becoming a mainstay at television stations and the main news camera for NBC's owned & operated stations... but it garnered a few complaints too, primarily the amount of skew/wobble in 1080/24p mode. It could also exhibit some black speckly noise (which was largely cleared up in a firmware update). I had the opportunity to test a prototype HPX300 prior to it being released, and while I found a lot to love about it, the rubbery 1080/24p was a problem. (You could avoid it by dropping down to 720/24p to greatly minimize the skew, so it's not like it was a complete showstopper, but it was still annoying to have a 1080/24p native camcorder that had to have restrictions placed on when you could use that 1080/24p mode).

What a difference a year makes. Panasonic has indeed taken all that was good about the HPX300, fixed the shortcomings, and even refined it a bit more for news usage. They consider the updates significant enough that they've christened this camera with a new model number, the HPX370.

So what's different? Mainly a new chip block, which provides a huge reduction in skew in the 1080/24p mode, better sensitivity, and even lower noise. These chips use what Panasonic refers to as U.L.T., Ultra Luminance Technology.

Operationally, there are a few things I noticed as well, aimed at making the HPX370 more usable for its intended news market. One feature is inherited from the larger cameras, which is the ability to extract or trim portions of a clip without having to use any other hardware devices or a computer. The other new function is the ability to record multiple snippets all into one continuous clip, instead of each start/stop session creating its own clip. This is a surprise and a nice one; I don't think any of the other P2 cameras offer this function, but I can attest that it's one that news editors have wanted for years. It lets the editor import one big clip into the editor and slice it up however they'd like, rather than dealing with dozens of individual clips. And, the big one, the HPX370 is a "world" model, offering both NTSC and PAL switchability (this was previously offered via an $800 factory upgrade to the HPX300; it's now standard on the HPX370).

But back to those chips since it's introduction, the HPX300 has been compared to the EX3, and the major advantage on the EX3 has been that its imaging chips are bigger (1/2" vs. 1/3"). And in all the internet chatter about these two, that one factor keeps being brought up again and again. Can you just imagine what Panasonic's focus groups were like, when coming up with ideas for the HPX370? I bet it went something like this:


Seats are packed with video professionals, notepads in hand. Several blue-suited Panasonic executives stand proudly around their new HPX300. An EX3 on a tripod stands in the corner.


We are interested in learning your feelings about our HPX300, especially in comparison to the Sony EX3.


(in unison)

The EX3 has 1/2" chips.

The Panasonic exec smiles knowingly, as he begins to explain the advantages of the HPX300.


Yes, ours has smaller chips, but ours uses the master-quality AVC-Intra recording format, which is far superior to the older MPEG-2 of the Sony.


Yes, but the Sony has 1/2 chips.

Another Panasonic exec points to the form factor of the two cameras.


The HPX300/370 is a full-size shoulder-mount professional form factor, whereas the Sony EX3 is a bizarre half-shoulder-mount form factor more like an old Canon XL1 DV camera.


True, but the Sony has 1/2 chips.

Frustrated, the Panasonic broadcast manager steps forward.


The P2 recording format has been adopted by over 80% of the world's major broadcasters. It is the new standard in broadcast.


Agreed, but the EX3 has 1/2 chips.

At this point, most of the Panasonic staff commences pulling their hair out in frustration. One steps forward, clearly at the edge of his patience.


The 300/370's recording format is certified Gold by Discovery HD and the BBC. It records true 4:2:2, true 10 bits, in a true intraframe format using the most advanced compression technology available, instead of Sony's creaky and ancient 8-bit 4:2:0 long-GoP MPEG-2, which is a reheated HDV and isn't even Sony's best implementation of MPEG-2. The 300 is a full-size camera, with a wireless mic slot, records four channels of audio, and is the main camera of NBC News!


Fine... but did we mention that the EX3 has 1/2 chips?

A quick huddled discussion ensues... What could Panasonic do? After all, if they changed to 1/2 chips it would require engineering a whole new series of 1/2 lenses that were CAC-compatible, and that would certainly blow the price point. (well, either that, or take away the P2, the AVC-Intra, the wireless mic slot, the full-size body, the proxy recording option, the multiple SDI output ports, and everything else that the HPX300 has... the Sony can afford bigger chips because it doesn't have all those other features.)

Finally, one of the engineering staff steps forward and offers a suggestion.


How about if we make our 1/3" chipset outperform their 1/2 chips?


And now, back to reality -- is that how it went down? Probably not, but the question at the end there is one that needs to be addressed. Because that's what their marketing says it says that their 1/3 chips "rival the image quality and sensitivity of 1/2" imagers." Bold claim. Is it true?

I wanted to find out, but I'm currently traveling and would have to rent some gear to really put it to the test. So, I rented a Sony EX3 to compare against a preproduction HPX370. I contacted a rental house, TexCam, who was able to supply lights and charts and other support equipment, but they were out of stock on EX3's. Fortunately I was able to contact DVXUser member ChrisCCW, who runs Clifton Camera Works in Houston, Texas, who came to the rescue and supplied a complete Sony PMW-EX3 kit.

I then set about exploring the technical comparisons between these two. I didn't bother trying to get "beauty" shots, as I knew who else had been using the HPX370 and I am quite confident that they will be able to show the camera off to its best. Instead, I was more interested in finding out the extent of the improvements, and what the realities are, versus the claims that are made.


This was my first test. After discovering the level of skew (by performing the infamous "skew-off" test between the original HPX300 and EX1), I know that the 1080/24p skew was a major complaint about the HPX300 and I agree; it's probably the main reason I never bought one. So how is the skew in the HPX370? Did they "fix" it?


It's sooooo much better! It's pretty much on par with the EX3 in 1080 mode, and better in 720 mode.

My testing was a little different this time; given limited resources I couldn't really configure the cameras the same way I did during the "skew-off," so I made do with what resources were on hand and in Texas, that means trucks and speeders. So I set the two cameras up side by side pointing at the freeway, zoomed in to full telephoto on the EX3 and matching (as best I could) the framing on the HPX370, and shot speeding semi trailers on the freeway at a 90-degree angle. It was actually a pretty ideal test, at least as indicative as the skew-off. Now, remember that skew is worse the more telephoto you zoom in, so I was trying to get it at its worst. And while it's there, it's drastically improved. In the skew-off, the 1080/24p of the HPX300 showed about twice as much skew as the EX1; the new HPX370 shows almost exactly the same amount of skew as the EX3. In various tests the numbers were within about 3% to 7% of each other, with the HPX370 showing just ever so slightly more than the EX3. The numbers can't be absolutely conclusive because (try as I might) the frame sizes were still very slightly different, but I'm confident in saying that if you're comfortable with the level of skewing on an EX3, you'll find the HPX370 is pretty much identical (in 1080 mode).

1080/24p 1080/30p 1080/60i

In 720p mode, the HPX370 shows consistently less skew than the EX3; across all three modes I measured the HPX370 at about 10-15% less skewing.

720/24p 720/30p 720/60p

The EX1/EX3 have been the standard of CMOS skew, and now the HPX370 goes toe to toe with them. So, I think it's fair to say that yes, the skew issue has been fixed!


One of the main advantages of bigger chips is that, all other things being equal, the bigger chips will have an advantage in sensitivity and noise performance. And it is perhaps these two subjects that have kept audiences doggedly persistent in pointing out that the Sony cameras have larger chips. So how does the HPX370 stand up?

Well, it stands up, all right it stands up and slaps the EX3 right on the cheek. The HPX370's 1/3 chips match the EX3's in most modes, and actually outperform the EX3's for sensitivity in some modes! I was expecting an improvement, but I didn't expect it to actually surpass the larger-chip competitor. But match and surpass it, it does.

I set the cameras to standard Rec 709 gammas (HD NORM on the HPX370, STD3 on the Sony) and comparable knee and black stretch settings, and used the waveform monitor on my BT-LH80 monitor to calibrate them so that they were delivering comparable brightness. And in 1080/24p mode and 1080/30p mode, the HPX370 is about stop more sensitive. The EX3 was consistently on f/3.4 when the HPX370 was on f/4. In 720p mode, they're identical for sensitivity. In 1080/60i mode the EX3 does have a slight advantage, but it's very slight I estimate the EX3 to be about stop faster in 1080/60i mode. Unfortunately I couldn't lay hands on a light meter to do definitive ASA testing, but from prior testing I have rated the EX1 at 400 ISO in 1080p mode, 500 ISO in 720p mode, and 800 ISO in 1080i mode. Based on those numbers, I would estimate the HPX370 at about 600 ISO in 1080p modes, about 500 ISO in 720p modes, and about 700 ISO in 1080i mode. That last bit was a bit of a surprise; typically the progressive-chip Panasonics show the same or even lower sensitivity when recording in 60i interlaced mode, but the HPX370 does indeed gain 1/3 stop sensitivity when going into 60i mode.


I didn't expect much of a difference in resolution, and the results in my testing were the same as I've gotten previously on the HPX300 and EX1. In 1080/24p mode, both exceeded the chart's ability to measure, clearly and easily resolving over 800 lines in each direction. I didn't have access to my 4K chart (which might have shown any difference at the very highest levels) but the Chroma Du Monde has resolution trumpets to 800 lines, and both cameras easily resolve the trumpets all the way. In 720/24p mode, they were also basically identical to each other as well.

So, the resolution summation yep, they're pretty much identical.

1080/24p 720/24p


The EX1/EX3 have been hailed as being quite low-noise. The HPX300 (especially after the firmware update) was competitive, but not quite as clean (and when the black speckles would hit, it would indeed be noisier than an EX1 or EX3). So, how about now?

Now, it's a different story. The HPX370 is cleaner. Period. In 1080/24p mode (which I spent most of my time in) it's noticeably cleaner than the EX3. Not dramatically, but noticeably. It was easy to see the difference on the monitor, and when I pulled in the chart footage and split-screened it and magnified it, yeah, the HPX370 is cleaner than the EX3.


Low Light Performance

You'd think that with more-sensitive chips that are lower-noise, the HPX370 would be a better low-light performer, right? Well, it is. But there's more to it (there's always more to it, isn't there?) At equivalent gain levels, the HPX370 shows less noise and more brightness (as you can see in the sample video above.) But, because the HPX370 is about a half-stop more sensitive (in 1080p mode) you can actually use -3dB of gain on the HPX370, versus 0dB on the EX3, and have the same level of brightness but a substantially lower level of noise. Furthermore, the HPX370's lens is about a half-stop faster than the EX3's (f/1.6 at full wide angle, vs. f/1.9) so that gives the HPX370 even another half-stop of brightness advantage. Assuming you want to stay at 0dB (or even -3dB) of gain, in 1080p mode the HPX370 wins the low-light battle. However, when you start pushing the gain, the Sony appears to have a cleaner gain circuit than the HPX370 does. I found the noise difference to be substantial and in the HPX370's favor at -3dB, but when both were set on +12dB it appeared that the Sony was slightly lower noise than the HPX370. Of course, I should have been trying the HPX370 on 9dB to match the Sony at 12dB; that might have equalized them out, but I didn't make that connection until the cameras were already put away.

As for the texture of the video, the stock settings make the HPX370's image appear to be a little "grittier" than the EX3, which looks a bit "smoother." I suspect this is due to the exact same changes that were made to the HPX170 (prioritizing the appearance of sharpness over the quest for noise suppression). The solution with the HPX170 was to increase the detail coring a couple of notches. The HPX370 behaves similarly; I would recommend anyone wanting that slick/smooth look to turn the detail coring up to +2 or +3 rather than leaving it at the default setting of zero.


Panasonic has indeed done just what they're bragging they've created a 1/3 chipset that matches and even outperforms the competition's 1/2 chips. They've fixed the flaws on an otherwise brilliant camera, and they've added the most-requested news-oriented features. The HPX370 is a solid performer that matches its bigger-chipped competition in imaging, and easily outstrips it in performance in most other categories.

Discuss this in the Forum