PDX10 test

[Shawn Mielke]

This has been one of the most intriguing threads I've read. No doubt because I work with cam in question. What a strange thing for Sony to not publish whatsoever. And for it to be beneficial, ultimately, too. Has someone already guessed why this would be?

[Andre De Clercq]

Small CCD's rapidly start to show diffraction resolution loss at higher f-settings. So the only way to counter this (and if 1/60 is chosen) is to limit the shutdown of the aperture (limit the f number) In order to still cover a large range of light condition, given the small aperture range (underlimit is about f1.6) one needs to use a wide range ND setting, and because this small aperture range 3 steps are needed..

[Steve Nunez]

What makes this all interesting is that us "guinea pig" test buyers are all learning this via our user observations- I wish Sony would make internal working specs like these known- especially since this seems to be a non-standard way of reducing light....they could have touted it as an "intuitive" video method but instead not publish any info on these "internal ND filters"....

...today while shooting under sunny conditions, the PDX exhibited the "streaking" I've read about (the producing of lightened video running vertically throughout the video frame starting and stopping wherever very bright light is encountered- such as where tree tops end and the sky dominates til the next tree- you'd see vertical lightening of the entire frame)...I'm also noticing the PDX has a very tough time with metering and seems to overexpose very easily.....

...for the most part I'm finding the PDX a bit tough to get set the way I like....with a few more remaining days left til B&H reopens- the fate of my personal PDX remains bleek- this model may go back and i may just pickup an Elura 50 and wait a few months tol the new HD cams come out- there's simply no choice for native 16:9 shooting and the DV953 may not be much better (if at all) than the PDX.

I wanna like the PDX- but just can't help feel it's metering is sub-par and it's propensity to blow-out highlights a bit overbearing for a $2000 semi-pro camera...I hope I discover settings i'm not using or else this cam goes back!

(how lame is not being able to use ND filters in the traditional sense??)

Stay tuned!

[Ignacio Rodriguez]

> I'm also noticing the PDX has a very tough time with metering and
> seems to overexpose very easily.....

Steve, did you try using the custom preset menu to change the AE point? In my PDX10 this is always set at least two steps lower than default.

[Steve Nunez]

Ignacio,

Thanks for the tip- I plan on reading the manually thoroughly tonight, end to end!

Here's something I threw together for my son- he's an avid praying mantis collector....shot with the PDX a few hours ago........

http://stevenunez.com/video/pages/chimantis.html

It's pretty interesting, let me know what you guys think.

[Tom Hardwick]

Yes, my PDX will over-expose normal 'auto' footage if left on the factory default settings, and it definitely needs correcting with the custom presets as Ignacio points out.

Steve - I'm sad to say that you've noticed the next big failing of the PDX, and that's its CCD flare. Remember this camera costs just a few dollars less than a VX2k, yet side by side the results couldn't be more different. I've had both cameras locked down on my big L bracket so that their lenses 'converge' at about 3 metres. I've been out filming with this combo and it shows with the cruel slap in the face of the A/B test how much better the VX is in bright light (CCD flare) and low light (three solid stops more sensitive).

What *is* good about the PDX is the native resolution of the 16:9 facility, but only if you steer well clear of dim surroundings or bright lights. Pretty restricting, huh? These situations can be so image degrading. My footage shot on Sunday has some shots that are completely unacceptable due to this CCD flare, and this was caused by simply having the cloud-obscured sun in frame. Looking up you couldn't see the sun, just a 'brighter cloud'. I upped the shutter speed as I wanted to selectively print the frames of kids on a trampoline. The CCD flare has made the frames totally - and I mean totally - useless.

This is indeed a sad state of affairs. OK, in normal, everyday shooting this cam gives great video. But hey - Sony are selling this as a Pro cam at a pro price, and it's simply not good enough.

tom.

[Ignacio Rodriguez]

Yes. Flare is something we are learning to live with. As things are, this was the only 16:9 cam available to me at the time I bought it, as I did not have more cash. Nor did I wish to pay to be a JVC HD-10 beta tester which was the other under $4000 16:9 alternative.

If you take care not to overexpose and keep contrast under control (which is good videography practice anyway) you can get very good images out of this camera. It's just not as good as the PD150 in some ways, and that is why it is cheaper. And it's better in other ways, which is why I like it ;-)

[Steve Nunez]

Tom, absolutely well said!

I'm pretty sure the PDX10 is going back- as much as I WANT to like it- the video issues are a bit too much to overcome for a camcorder of this supposed "quality" and price. The streaking is very drastic and like you said Tom, it will streak even when the offending "bright" element is not even in the video at all but just outside the frame! For a nature videographer/enthusiast who will likely tape under mixed lighting at any given moment- the PDX is simply not up to this type of shooting....speckeld light coming through branches absolutely kills the video- very sad indeed!

So I'll play with it for a few more days until B&H reopens from their holiday schedule and return it and perhaps get a DVC80 and a tele lens......I'm losing 16:9 but I had a DVX and the video quality was excellent (as was the GL2) so if the DVC80 is anything like the DVX sans progressive modes (and other adjustments I'm aware of) I can see the DVC80 in my future......I'll do some more research tonight and look into my alternative cams.

Note: I do think the PDX is capable of beautiful video- but only under perfect lighting (outside the issues as noted) and under controlled lighting- so I'm not bashing the camera- just don't think it's the camera for me- time will tell!

[Ignacio Rodriguez]

> Note: I do think the PDX is capable of beautiful video- but only
> under perfect lighting (outside the issues as noted) and under
> controlled lighting- so I'm not bashing the camera- just don't
> think it's the camera for me- time will tell!

I have to agree. It would be so nice to have a cam with the PD150/PD170's luma response and a higher resolution CCD capable of real 16:9. But it would most likely have to have a much larger CCD array, and thus a higher price. The more pixels are on a CCD of a given size then the smaller each pixel is, thus the less light it gets. This explains why the PDX10 is less sensitive to light than the PD150/PD170.

I guess it's just a question of time before they make better megapixel CCD arrays at small sizes.

Right now, the next step up for 16:9 would be the expensive Sony DSR-500 and JVC DV-700... oh and of course there is the Panasonic AJ-SDX900, it only cost's about $25k. No problem.

And well there is that 1-CCD experiment, the JVC HDV cam, I understand it can also do 16:9, not sure if in SD or 'HD' mode though. Seems to have even more low light probelms than the PDX10 (but less vertical smear).

[Boyd Ostroff]

Sorry it didn't meet your expectations Steve. From your posts it sounds like you were often using the camera in one of the auto exposure modes; is that true? It seems that many of the complaints I've read about the TRV-950 and PDX-10 were centered on how the camera performed in "auto".

I have never used my PDX-10 in anything but manual mode, and have found the vertical smear problem to be pretty limited. The other day while filming one of our stage productions I was surprised to see that an exposed light was not creating a smear. I have plenty of footage of sunsets, skies, clouds, the moon, etc. without vertical smear. My experience is that there's a definite contrast threshold where it will occur. If you dial down the exposure you will reach a point where it becomes less noticeable, then finally disappears completely. Depending on what you're shooting this might cause another area of the frame to be underexposed of course.

Anyway, it has just never been a big deal for me and in fact I like the effect in many cases. I intentionally shot these looking into the sun: http://www.greenmist.com/trovatore/film/20030909/duel. As the swordfighters engaged and blocked off the sun it created a great effect with the "spikes" from the smear flickering on and off.

Camera choice is (and should be) a very personal thing, and the PDX-10 isn't for everyone. But it continues to meet or surpass my own expectations. So I still think it comes down to how important 16:9 is to you and whether you can work around some of the quirks.

Regarding the DVX-100, it looks like a really nice camera but is really in a different league with a cost approaching twice that of the PDX-10.

[Ignacio Rodriguez]

Today I compared a friend's new PD150 with my PDDX10. First I set up both camera's LCD configuration and Custom Presets so they would match as closely as possible. Then I played around with both cams, panning around the room looking for high contrast... and I was surprised to find the PD150 can also be made to smear pretty easily. The problem is not as bad as on the PDX10 for several reasons: (1) because there are less active pixels outside the usable image field (the PDX10's CCD is larger than the image, both in 4:3 and 16:9 mode) and (2) the PDX10 has a tendency to over-expose left to it's default settings. At similar exposure, I observed less smear with the PD150, but not really much less. My suggestion would be: (1) try to use use french flags, so as to keep light falling into the lens restricted to what the video is really using. And (2) set the AE point in the Custom Preset menu at least two steps down from default.

[John Steele]

Ignacio,

If you didn't see much less smear with the 150 then I don't think you had the PDX setup to see the maximum effect of the smear, we all know that it can be reduced by changing camera settings. A better test would have been to maximise the smear on the PDX then change the 150 setting to suit that to see how different it really is and personally speaking I think there is quite a difference.

John.

[Tom Hardwick]

I agree John. With both cameras (the VX and the PDX10) on the same L bracket and both in automatic then yes, you can get conditions where the VX will give you CCD smear. But then look at the footage from the PD. It'll be totally unuseable.

It's such a shame that the PDX10 has lost the progressive scan facility of the PD100. That full res motor drive mode was an absolute boon to any sort of sports evaluation. Upping the shutter speed on the PDX simply makes the smear so bad that any sort of back light - or even simple high contrast - cannot be used.

I've ordered another PDX10. I simply cannot believe that this one I have here for test is representative of the breed, especially when the PD100/TRV900 was so good.

tom.

[Andre De Clercq]

Subjective vertical smear tests are often misleading. It's a matter of defining how much stronger the light from a point source can be versus the scene ilumination for a 100 IRE picture. So it's a matter of ratio's rather than absolute piont source brightness. Pro cams (using smear reduction CCD structures) go up to a 120 and more db reduction factor. Small CCD's with high pixel count and high fill factor (for keeping acceptable sensitivity) are generally more prone to show these effects because of electrical and optical leakages in the vertical tranfert channels

[Steve Nunez]

The PDX10 has the best digital zoom I have ever used- there are no signs of compression blocks or artifacts....I switched it on today and I've got t say- i'm pleasantly surprised at how well it looks onscreen....almost as good as having a 24X optical zoom lens! I'm guessing it's DSP is the answer to it's exceptional d-zoom- anyone else know how they get such non-interoplated looking zoom capabilities?