The 1080P mode vs the 720P mode!

[Barry Green]

I'd vote for "A" as being the 1080 shot. The live E-to-E image shows about the same res in 1080 or 720, but the recorded image in 1080 retains more resolution than the recorded image in 720.

Why, since the camcorder doesn't even deliver true 720p resolution -- why would Panasonic add insult to injury by offering 1080p?

Oh, come on. This is getting silly.

If we want to talk numbers, and we want to talk resolved definition, let's talk about it all the way. Steve, are we still arguing pixel count? Because if we are, I propose that you have to decide one of two choices: either pixel count is important, or it is not. Only one of those two can be true, correct?

If it matters (as Steve seems to think by exhorting the JVC's native pixel count) then obviously Steve must abandon the JVC immediately, since the Canon offers 69% more pixels than the JVC does. The Canon offers 1,555,200 pixels vs. the JVC's 921,600 -- that's a difference of 68.75%. If pixel count is so important, stay true to your convictions and go with the Canon and forget the JVC, since the Canon has indisputably the highest pixel count of all the low-cost HD cameras.

So if you choose to stick with the JVC, that pretty much means admitting that pixel count isn't really all that important, right? So if pixel count does not matter that much, then why continue to bring it up in an attempt to discredit the Panasonic?

Here's a great, great, great example. Compare these two res charts. One is from a 378,840-pixel DVX100 using the Andromeda, and one is from the 921,600-pixel JVC. So if pixel count is so important, one of these shots should have roughly 2.43x as much definition as the other one, right?

http://www.fiftv.com/HVX200/Andromeda-vs-HD100.JPG

Ummm... Hmmm. Son of a gun. Maybe overall pixel count isn't such a big thing after all? Maybe it's actual resolved image definition that's a bit more important perhaps?

The Andromeda is showing equivalent horizontal resolution from its 770 pixels (vs. the JVC's 1280) and it's showing substantially higher vertical resolution from its 492 (vs. the JVC's 720). The Andromeda-ized DVX with 378,840 pixels is showing a sharper image than the HD100 with its 921,600 pixels.

I for one don't really care what the pixel count is, as long as the resolved image definition is competitive. And Juan Pertierr of reel-stream has shown us how you can get equivalent definition (and arguably BETTER definition) from as few as 770x492 pixels. So as long as the HVX has at least 770x492 pixels it should be able to deliver an image comparably sharp to the JVC, right?

And, surprise surprise, it does. Here's an extraction of the HVX vs. the JVC:
http://www.fiftv.com/HVX200/HVX-vs-HD100.JPG

In that shot the JVC is showing higher horizontal res, the HVX is showing higher vertical res, they're showing about equivalent amounts of edge enhancement, and the overall image of the HVX is a tad sharper than the JVC.

So again, I don't care what the pixel count is. It could be 770x492 for all I care, what I care about is resolved definition in the image. And the HVX is every bit as sharp as the JVC with its "native pixels".

(res chart extractions taken from the originals posted here: http://forum.reel-stream.com/viewtopic.php?t=363)

[Aaron Koolen]

Originally Posted by Steev Dinkins

Here are two screen grabs:

http://www.holyzoo.com/content/xl-h1/1920x1080_a.jpg
http://www.holyzoo.com/content/xl-h1/1920x1080_b.jpg

Which one is an upres from 1280x720?!?!

Can you really say there's a tremendous difference?? I think not.

It looks pretty different to me. The 'a' shot looks higher res but was is sharpened in post? Both those images are identical.

[Alister Chapman]

Yes, but neither the JVC or Andromeda are claiming to be native 1080 cameras. The argument is very much like the difference between a small engine with a turbo and a large engine, they may claim to produce similar horse power, but the way the power is delivered is very, very different, torque, lag, revs etc. It is the same with a camera that uses clever sharpening or enhancement to give a percieved increase in resolution. To my knowledge no one has yet created a way to create picture information from picture information that simply wasn't captured in the first place. Maybe Panasonic has, perhaps all manufacturers will follow Panasonics lead and fit low res CCD's into thier new HD camcorders (saving us lots of money) and then use trickery to magically restore all that lost detail.... somehow I think not. Any camera can be set up to give good test chart results, but it is real world images that count and these results need to be viewed and compared on decent sized monitors truly capable of displaying the whole story, I belive then people will realise why the difference IS important.

What is the point of buying into HD if you don't care about resolution? Yes I would also hope that you are looking at all the other aspects of the image, but the one thing that is supposed to differentiate HD and SD is resolution. In theory there sould be no difference in lattitude, dynamic range or contrast ratios, the one and only difference is RESOLUTION and if I was being sold a 1080 camera I would expect 1080 resolution. Hell my old JVC HD10 could output 1080i over the component output but JVC never claimed it was a 1080 camcorder as this was simply an up sampled output from 720p.

For me it has to be 1080, preferably 1080P which the HVX200 just dosn't seem to deliver. I await future developments.

[Steve Mullen]

Originally Posted by Barry Green

The official info from Panasonic was that it samples at 1080/60p (or whatever frame rate you're using). It takes two 60p frames and slices them into fields and creates the 60i stream from that.

Since the very definition of HDTV is about resolution -- it IS "a" critical specification. The very fact you continue to talk about "1080p" means you continue to think it is an important issue. However, it's time for you to come to terms with your past and current claims about HVX200 RESOLUTION. So you have a choice:

1) Continue to support the "offical" 1080p info and explain how 50% of the lines get lost OR decide there is something invalid about the official info. A 1080p chip would yield nearly 1000 TVL, not 540 TVL. You understand video well enough to know this.

Or, do you continue to worry Adam screwed-up his tests OR the HVX200 was flawed? In that case, you need to get Adam to correct his DV story or run tests with a new HVX200.

2) If you believe interlace is generated by outputting two 540-line fields without row-pair summation -- then the interlace V. rez. should be nearly 1000 TVL, not 540 TVL. Look at the Canon's resolution data. (Such video would also be unwatchable because of line flicker/twitter and wouldn't really be "interlaced.")

The CCDs yield 550x540 RESOLUTION. That number is lower than 720p which yields 700x700. And, the only way 550x540 can fill a 1440x1080 frame is by scaling. And, after scaling, the resolution remains the same -- which indicates scaling is indeed used.

These published data clearly indicate the HVX200 provides a measured resolution -- TVL/ph and TVL -- not "pixels" or "pixel count" -- that does not equal either that measured from 720p or a 1080i/p camcorders.

Plainly put, the HVX200 records a slightly better than PAL SD resolution image within either of two HD "wrappers."

------------

If one believes that 50% of the lines are not simply "lost" then IMHO the only reasonable belief is that the CCDs are 960x576. And, the sensitivity data support this belief. But, you don't have to believe this. It's the test measures that are important -- not my belief about the CCDs.

[Mike Marriage]

Originally Posted by Steev Dinkins

Here are two screen grabs:

http://www.holyzoo.com/content/xl-h1/1920x1080_a.jpg
http://www.holyzoo.com/content/xl-h1/1920x1080_b.jpg

Which one is an upres from 1280x720?!?!

Can you really say there's a tremendous difference?? I think not.

Is this a trick question? I can't see how you can record the identical frame in 720p and 1080i at the same time...? Picture A looks "sharper" but seems to be digital sharpening rather than actual resolution. In fact, sharpening "B" in photoshop results in a better image than "A" because "A" has more compression artifacts.

So go on, tell us the answer :)

[Aaron Koolen]

You guys are way more knowledgeable about this than me, but from my perspective, the fact that we have heard nothing from Panasonic (Particularly Jan Crittenden Livingston who was on DVXUser quite consistently during the lead up to the HVX) does make me wonder. Now I know it the final image that matters, and franky, I can't afford one anyway, but my curiosity wouldlike to be satisified, and I'm sure so would the curiosity of all those who ARE thinking of buying.

Just answer our questions Panasonic. Can it be that hard?

[Mike Marriage]

I was just thinking about this.

Right, we all agree that the HVX horizontal resolution isn't great? From what I have seen it is worse than the vertical.

So why is HR lower than VR? The lens shouldn't favour one axis and the codec has more pixel horizontally than vertically.

That leaves the DSP and the CCDs. Maybe the pixel shift favours vertical res?

BUT, maybe the CCDs have higher vertical res than horizontal. That would also explain why Panasonic won't release the CCD info! It would look pretty odd if they admitted it used 576x720 CCDs wouldn't it.

Hey, I'm just throughing a new spanner into an old debate :)

[Stephen L. Noe]

Originally Posted by Barry Green

If it matters (as Steve seems to think by exhorting the JVC's native pixel count) then obviously Steve must abandon the JVC immediately, since the Canon offers 69% more pixels than the JVC does. The Canon offers 1,555,200 pixels vs. the JVC's 921,600 -- that's a difference of 68.75%. If pixel count is so important, stay true to your convictions and go with the Canon and forget the JVC, since the Canon has indisputably the highest pixel count of all the low-cost HD cameras.

Not quite so fast. The JVC is true to it's format (1280x720) and lives within it's format resolution parameters. The Canon's count is higher, but the camera is intended for a higher resolution to live within it's format.

Pixel count is important once the camera goes in motion.

[Greg Boston]

Originally Posted by Aaron Koolen

... the fact that we have heard nothing from Panasonic (Particularly Jan Crittenden Livingston who was on DVXUser quite consistently during the lead up to the HVX) does make me wonder.

Just answer our questions Panasonic. Can it be that hard?

Keep in mind Aaron, that Jan was involved in an automobile accident at Sundance film festival. This is about the same time that the HVX began shipping to customers. I don't know that she 'feels' like posting at this time. She may be on pain medication and doesn't want to post something while under the influence. See the 'Get Well Soon' thread elsewhere on DV-INFO.

-gb-

[Steev Dinkins]

Originally Posted by Steve Mullen

Plainly put, the HVX200 records a slightly better than PAL SD resolution image within either of two HD "wrappers."

So seeing how the HVX200 has more resolution than the JVC HD100, I guess the HD100 is a little less than a little PAL CCD camera. ;)

Okay, I finally figured out what bothers me the most about this whole thread. This reminds me of the 96khz 24bit audio crowd (or SACD) that says they hear such a huge magnificent difference compared to 44.1khz 16bit or 24bit audio. Trained "professional" ears. Sure. Now... Go ask the general public! Now go ask people who actually produce award winning material!! How many people can hear above 16khz? How many people can see the difference between 600 and 900 vertical lines on a res chart? How many people sit down and watch SD broadcast recorded to Tivo in glorious triple compressed and halved sliced diced hashed to hell Mpeg-2 and thoroughly enjoy themselves? Or even the ubiquitous and cherished SD commercial DVDs Ha hah ha ha. This is great!! These cameras are simply a big cut above the SD-DV crop, and calling the HVX200 a little better than PAL resolution is ridiculous.

Originally Posted by Mike Marriage

Is this a trick question? I can't see how you can record the identical frame in 720p and 1080i at the same time...? Picture A looks "sharper" but seems to be digital sharpening rather than actual resolution. In fact, sharpening "B" in photoshop results in a better image than "A" because "A" has more compression artifacts.

So go on, tell us the answer :)

So, although not scientific, it makes a point. I took a screen grab from Canon XL-H1 footage at 1920x1080. This is image "A". Then I downconverted it to 1280x720 in Photoshop, throwing away the pixels by saving the file out. Then I closed the file, re-opened it and scaled it up to 1920x1080. This was done to see if there's really much of a difference. And there is not. There *is* a difference, but it's slight. And I would say the difference looks similar to the difference betweeen 1080 and 720 on the HVX200.

I know this may not get addressed here, but is this thread done yet? ;)

[Mathieu Ghekiere]

Steev, I don't think it botters people as much that the resolution of the HVX is a bit low, I think they are more dissapointed or concerned about the fact that the camera doesn't output the resolution Panasonic says it does.
If you promote your camera as a 1080p camera, you should deliver those lines of resolution. Otherwise you are lying to the customer. (I'm not saying Panasonic does this and I ain't saying they don't, but there seems to be doubt here...)
Well, I shouldn't be speaking for other people, but that's what I think anyway.
I don't care about all that resolution that much, but I AM interested in this thread, because I do think it's a relevant issue. People who buy this cam have the right to know if the camera delivers the true lines it says it does. Even if most of the customers don't care - as I do -, the ones that DO care have the right to know the RIGHT information.
I know what matters is the FINAL image, but you can't always throw that in as an excuse to lie as a company about the CCD's (this is NOT an attack to Barry Green, NOR an attack to Panasonic, just a general statement)

PS: I am not saying the camera does or doesn't deliver the lines, I don't have the camera and I wouldn't know how to make a proper test, I am only interested in the conclusion of this thread if what Panasonic says is right or wrong. Not because I don't like Panasonic - I do - but I just think that's relevant. That's all.

Now I'm done with my preaching ;-)

[Stephen L. Noe]

Originally Posted by Steev Dinkins

How many people can see the difference between 600 and 900 vertical lines on a res chart?

Everybody, when it's on a silver screen. You are right about SD but when you're shooting for the big screen, resolution and absolute critical focus makes a big difference. Everything is amplified...

[Steev Dinkins]

Originally Posted by Stephen L. Noe

Everybody, when it's on a silver screen. You are right about SD but when you're shooting for the big screen, resolution and absolute critical focus makes a big difference. Everything is amplified...

Will the audience know that you could have shot it with better res? Will it matter? It probably matters more to the content creator, yes. And I think it was said once before that if you are doing a resolution battle, the Canon XL-H1 would be the winner.

I know much was said and specifically not said, before the camera was released regarding CCD specs and such. Well this drama continues here without real specs on the CCDs, and I suppose some will go the grave always wondering, "man, I still wonder what the HVX200s CCDs were?"

On their site, right now it just states "1/3" 16:9 native high-sensitivity progressive 3-CCD with 1080/60p scanning". Is anything pointing to this being patently false? I think Jan had said this isn't some numbers war, and/or certainly she didn't want it to be. And with that being said, I suppose there's no way to avoid the numbers war with this. But I wouldn't count on Panasonic coming back to make you feel better about what you perceive as being a lie, or numbers not being real, or as they should, etc.

Do you like the images the camera produces or not? No? Isn't good enough? You can purchased something else today for sure. Or you can wait for another rev after all of this painstaking waiting already.

I for one made my decision without seeing any footage at all and not knowing any numbers. Ha ha. If I had received the camera and thought it was all hype and sorely lacking, I would have returned it. Plain and simple.

On the other hand, I'm blown away by the image quality and feature set. If you're going for absolute best the camera will do, shoot 1080. It does look better. To me, 720p is the sweet spot, in my opinion, as I've stated before.

Weeeeeeeeeeeeeeeeee!!

[Thomas Smet]

One area people seem to never look at when starting with a higher resolution is the compression artifacts.

On a DVD the compression artifacts look great on a standard sized SD TV. Once you get into a HDTV or very large SDTV you can really start to see the image break up.

When I take DV images and scale them up it isn't the softness that kills me it is the artifacts and the low color space.

This is why is some ways it is better to start with 1080p/i even if there isn't an increase in resolution. At least you know at that size all of the compression is done and chances are it either will not be scaled up or at least has more pixels if it does get blown up to film. Even if there is no more detail there are more pixels to make up that image and it will scale up or down better.

Even 720p may start to have problems when watching on a newer true 1080i/p display or on a film transfer because there are less pixels to blow up. The compression artifacts from both 720p HDV and 720p DVCPROHD (yes there are compression artifacts equal to around DV artifacts) will get enlarged more than they would if you were starting with a 1080p source.

It is always better to scale up a raw image (either in a computer or dsp) before compression.

[Joel Aaron]

Originally Posted by Steev Dinkins

This reminds me of the 96khz 24bit audio crowd (or SACD) that says they hear such a huge magnificent difference compared to 44.1khz 16bit or 24bit audio.

I agree with the one caveat that Panasonic has strung everyone along for 10 months to believe we were going to get 1080p that was going to be better than the competition. It isn't and now people are arguing the HVX looks about as good as the competition, which I would agree with. But it's pretty comical to watch from the sidelines. Panasonic should be squirming and taking some hits for that one.

Now in my personal opinion there's PLENTY of resolution on the HVX200 from all the stuff I've seen... and I've seen enough of it on big monitors etc.

The reverse of this resolution argument is the HDV argument. It's just as theoretical!!!

I've never seen anyone post ANY footage or audio that the HDV format has ruined... much less JVC's 6 frame GOP. That's 1/4 second sampling. I've seen a BUNCH of action stuff that looked GREAT. So far all the smoke, fog and haze I've seen looks great. A friend of mine who has much more of a critical eye for this stuff than most has 40 hours of footage on his Sony that looks great. No audio problems ever. He just wishes it was 24p and was more light sensitive.

If I was shooting nature stuff for HDTV I'd want the Canon.

For 35mm adapter stuff (other than P&S) and Final Cut users (right now) I think the HVX gets the edge. This is me...

For everything else that JVC looks really good to me. The form factor is amazing and the focus assist makes focusing on the LCD possible which is HUGE in my mind though I want to test more. It's the lowest cost solution and delivers at least as good quality.

For overall picture quality I now think I could dial any of them in to look pretty good to me.