Panasonic HVX-200 Native Resolution

[BobDiaz]

First of all, let me start by saying that everything I have seen at the December DV Expo West Trade Show about the Panasonic HVX-200 camera, tells me that this is one FANTASTIC camera. Compared to other HD cameras in the under $10,000 range, no other camera can come close to many of the special features it can offer. While it is possible that there is some hidden flaw (or flaws), this seems VERY unlikely. Still there is one issue about the design worth talking about. I do not consider this a design flaw, but is something worth knowing about the camera.

As far as I can find, Panasonic has NOT released any information about the native resolution of the CCDs used in the HVX-200. When I asked during the Question and Answer session about the native resolution of the CCDs, the response never answered my question. Very likely Panasonic didn't want us to get into a numbers mode and fail to consider how the camera performs. Given that JVC must use dual A/D converters per 1 MP (Mega Pixel) CCD at 60 Hz, this suggests that the image resolution of the CCDs in the Panasonic are less than 1 MP.

What follows is speculation on my part and should NOT be taken as a proven fact. All conclusions are based on the assumption that the native resolution of each CCD on the HVX-200 is 960h x 540v or about 1/2 MP. Even if my assumption is correct, Panasonic appears to applied a simple, but brilliant approach to the design, in order to obtain true HD resolution.

If Panasonic had aligned the Red, Green, and Blue CCDs so that they all cover the exact same locations, the total CCD resolution of the camera would be 960h x 540v. This is 1/2 the horizontal and 1/2 the vertical resolution of 1920h x 1080v. However, both the Red and Blue CCDs are offset from the Green CCDs by 1/2 a pixel in both the horizontal and vertical direction. This means that the center of each Red or Blue CCD element lands exactly at where the 4 corners of the Green CCD elements meet. (The offset in the Red and Blue CCDs was covered in the Panasonic HVX-200 presentation.)

With this layout, if you count the number of CCD elements in either the horizontal or vertical direction, the number doubles. That is with this offset, 960h x 540v acts like 1920h x 1980v. In theory, that holds true if the fill factor (or aperture) of each CCD pixel was 25% (or less) of the area set aside for each element. In the past, that might have been true, but today it seems unlikely, because many CCDs are using micro lenses to increase the fill factor to close to 100%. This is done to improve low light performance.

Unfortunately, the fill factor of the HVX-200's CCDs is not known. We will assume it is 100%, which gives us the best low light performance and the worst case on overlap. If so, in the horizontal (or vertical) direction, every Green Pixel is overlapped by a Red and Blue Pixel by 50% of the pixel. The overlap helps to reduce the possibility of moiré patterns, but it also reduces the increase in resolution from 2x to 1.5x. Thus 1.5x increase results in an effective resolution of 1440h x 810v. If the fill factor is less than 100%, the effective resolution is higher than 1440h x 810v. One statement made by Jan in her presentation suggests that the fill factor is less than 100%.

Even with a fill factor of 25%, horizontal (or vertical) line resolution drops from 2x to 1x when the lines or detail are at EXACTLY a 45 degree angle (either way). While true, this seems like a moot point, seeing as how most detail appears to be close to a horizontal or vertical orientation.

Remember these figures assume a worst case conditions of a fill factor of 100% and the native CCD resolution of 960h x 540v. If Panasonic used CCDs with a higher resolution and/or lower fill factor, the effective resolution of 1440h x 810v would be higher.

Even with the lower effective resolution, 1080 HDV has a resolution of 1440h x 1080v interlaced. This means that in the horizontal direction, the HVX-200 offers no better resolution that any other 1080i HDV camera. In the vertical direction the reduced effective resolution may not be significant, because all HD cameras under $10,000 may lack the ability to fully resolve 1080 lines.

As for color resolution, HDV is 4:2:0. That means that the color resolution is 1/2 the grayscale resolution in both the vertical and horizontal directions. This would be what I believe is the native CCD resolution of the HVX-200, 960h x 540v.

If we look at the 1280h x 720v resolution, the worst case effective resolution of 1440h x 810v for the Panasonic is higher than 720p mode. In that mode, effective resolution is a non-issue and the quality of the lens becomes the limiting factor.

To me it appears that the 720p modes are where the HVX-200 shines. All of the camera's VeriCam modes are supported in 720p, the P2 memory requirements are reduced (compared to 1080), and the effective resolution of the CCDs is greater than the 720p resolution.

It can be argued that the Color CCD resolution of the JVC HD-100 is 1280h x 720v. This is true, but the color information is recorded to tape as 4:2:0 or 640h x 360v. The Panasonic records color information to memory at 4:2:2 or 640h x 540v. (Remember the numbers for the Panasonic are based on a worst case guess on my part.)

Even if everything I assumed about the CCDs in the Panasonic is correct, and my guesses could be wrong, the HVX-200 blows away all the other under $10,000 HDV cameras in terms of features and flexibility. My background is electronics, computers, and manufacturing technology. I have some understanding of the design tradeoffs that must be made in creating a product. Based on what I saw at DV Expo West, Panasonic made all the right choices and produced an amazing camera.


Bob Diaz

[stephenlnoe]

Hi,

HDV 720p:__________Frame res 1280x720, 4:2:0, Y sample: 1280x720, Cr,Cb Sample 640x360
DVCProHD 720p:_____Frame res 1280x720, 4:2:2, Y sample: 960x720, Cr,Cb Sample 480x720

JVC (HDV1) wins in Y sample and Panasonic (DVCProHD) wins in Cr,Cb. In the end there's not much difference between the two in 720p

[Jan_Crittenden]

Hi,

HDV 720p:__________Frame res 1280x720, 4:2:0, Y sample: 1280x720, Cr,Cb Sample 640x360
DVCProHD 720p:_____Frame res 1280x720, 4:2:2, Y sample: 960x720, Cr,Cb Sample 480x720

JVC (HDV1) wins in Y sample and Panasonic (DVCProHD) wins in Cr,Cb. In the end there's not much difference between the two in 720p


Stephen,

If you look at this just this part of the compression scheme, you are playing with numbers and you could draw the conclusions you have. However you must keep in mind to look at how much data is there to back up that pixel map. 19Mbs or 100 Mbs. MPEG2 is a very nice transport stream but frankly it offers too much compression to be truely viable. These compression numbers mean simply the sub-sample and when you play them back they are at 720 X 1280 or in 1080i mode 1920 X 1080.

Saying that the JVC wins, doesn't take into account all that needs to be taken into account before you can even make that statement.

Best,

Jan

[Policar]

Bob, you may be right and that would explain the high sensetivity. Smart. Apparently pixel shift can theoretically increase resolution by root(2) pixels in a given dimension, but by simulating bayer interpolation with two colors instead of one, you're right that you can probably increase more. Whether you can get root(2) X root(2)=twice the resolution in both directions (960540 being 1/4 the pixels of 1080p) doesn't seem likely to me, but whatever.

These are the first specs Jan hasn't said are wrong (and she said 960X720) was wrong.

Also, 1080p does look a little fuzzy, though 720p look amazing (although even Canon's frame mode has a resolution drop so whatever.)

I still love the hvx for what it's worth...

[Luis Caffesse]

These are the first specs Jan hasn't said are wrong (and she said 960X720) was wrong.

Interesting point.

Hmm, 960x540?
I suppose that means they could be using the same exact chips that are in the XL2.
Didn't Panasonic's parent company provide those? And if I'm not mistaken they are 960x720.

How funny would that be? (well, I guess most people here wouldn't find it funny)
Would make sense why they haven't released the information.
:thumbsup:

[Flintstone]

Interesting point.

Hmm, 960x540?
I suppose that means they could be using the same exact chips that are in the XL2.
Didn't Panasonic's parent company provide those? And if I'm not mistaken they are 960x720.

How funny would that be? (well, I guess most people here wouldn't find it funny)
Would make sense why they haven't released the information.
:thumbsup:Now that would be ironic! :shocked:

[FatBird19]

960x540? Whoa, that's low!
If that's true, Panasonic had better provide a bunch of h.264 sample videos on their site cause I know that a large amount of consumers and propeller heads are going to see "960x540" and run away screaming. Myself, having seen the footage, know better. :grin:

Panasonic, put lot's of samples on your site. Get some of the talented people in here to make them. :thumbsup:

[JitCam]

people please!

cameras are not just resolutions.. dont get caught up only into the numbers.. there are other things wirth mentioning also.. like lense.. which plays a HUGE role in overall quality and texture. i really dont think that the HVX lacks in any way in resolution. ive gotten really good images from my sony miniDV which is a single CCD cam. i think with proper lighting and good use of codecs u can get pro looking footage from this cam.

[Jan_Crittenden]

Unbelievable. It is not the Canon chipset! and the numbers suggested are not correct.

Best,

Jan

[harlan]

:)


.

[Luis Caffesse]

Unbelievable. It is not the Canon chipset! and the numbers suggested are not correct.


Hey, can't blame us for having a little fun.
:)

I wasn't suggesting that seriously. I figured it was outlandish enough to be seen as a joke, but I suppose you're right, this is how rumors get started.

But, speculation is naturally going to continue.
I suppose that's an unintended but inevitable consequence.

cameras are not just resolutions.. dont get caught up only into the numbers.. there are other things wirth mentioning also.

I don't think anyone is going to disagree with that.
CCD specs are just one of the many factors on the list of things that play a part.
I think people are just wanting to know all the details about the camera, not just some.
Seems perfectly natural to me.

[BobDiaz]

I'll be the first to admit that the numbers were all GUESSES on my part. In no way was I ever intending to say or suggest that the camera was less than fantastic.

I'm a "nuts and bolts" sort of person, so I like to try and figure things out as to how they work and why they work they way they do... In the case of the HVX-200, using worst case numbers, even if they are too pessimistic, tells me that the camera's CCDs can resolve the HD Image.

While I find doing the calculations interesting, the REAL TEST is how the final image looks. In the case of DV Expo, Samy's Camera had an HVX-200 connected to a samll HD CRT Monitor. Even with the less than ideal lighting at the trade show, the image looked very good as far as contrast and colors. (The monitor was too small to judge resolution.) I came the second day of the show and Panasonic connected a large display to show footage from the camera. Both up close and at a distance, the image was great!!!! Very clean and sharp image with good contrast and color.


Bob Diaz

[stephenlnoe]

Stephen,

If you look at this just this part of the compression scheme, you are playing with numbers and you could draw the conclusions you have. However you must keep in mind to look at how much data is there to back up that pixel map. 19Mbs or 100 Mbs. MPEG2 is a very nice transport stream but frankly it offers too much compression to be truely viable. These compression numbers mean simply the sub-sample and when you play them back they are at 720 X 1280 or in 1080i mode 1920 X 1080.

Saying that the JVC wins, doesn't take into account all that needs to be taken into account before you can even make that statement.

Best,

Jan
Jan,

The numbers are straight from the spec. HDV imported can be handled in many different ways. In my NLE setup I have the choice of handling the HDV native (MP@HL-14), RGB (AVI), or 2VUY (uncompressed). Once the HDV is in the system it can be uncompressed and handled as any other uncompressed format irregardless of the compressed transport stream it was captured in.

The numbers above represent the raw resolution/color/lum and you're right there are many variables when you introduce pixel shifting. In the film transfer we are working on now with ProHD, the uncompressed frames are 1280 x 720 @ 3.51MB per frame (TGA) and the Y,Cr,Cb is as expected from the published spec.

In the end, DVCProHD is a transport stream as well. It is compressed and I'm sure anyone doing a film transfer would work in (or at least deliver) uncompressed and not leave the footage in DVCProHD to work with.

[Bruce Morgan]

i am way behind on understanding the archtecture of video albeit
digital cameras ...
So Thanks Bob .
Are you aware of altenatives such as the andromeda up grade ?
And since you are technically inclined Id like to hear your opinion of that add on system .(search andromeda on this site )
Also where "Hd" is concerned many other makers such as sony achieve a reported high numbe of "recorded resolution" through a technology referred to as line double-ing .
Further comments on the andromeda modification (available to dvx users) and the use of line double technology would be appreciated .
Cheers
Bruce

[Luis Caffesse]

Bruce,

You can find more information about the Andromeda on their website:
www.reel-stream.com
The andromeda mod is only done on the DVX, it will not be done to the HVX.

But I think we're getting off topic here...

[Luis Caffesse]

Jan, could you tell us if Panasonic has any plans to release the CCD specs in the future? Or is this a permanent omission?

Just curious.
Thanks.

[stephenvv]

Jan,

The numbers are straight from the spec. HDV imported can be handled in many different ways. In my NLE setup I have the choice of handling the HDV native (MP@HL-14), RGB (AVI), or 2VUY (uncompressed). Once the HDV is in the system it can be uncompressed and handled as any other uncompressed format irregardless of the compressed transport stream it was captured in.


Uncompressing after capture does not magically rid you of the compression artifacts already captured. Jan's point is that taking those pixels and compressing them into 1/4 of the data stream exacts a price. How much, we don' t know. But unless you on capturing uncompressed and NOT using HDV tape, Jan's point it is correct.

Clearly, the footage from the HD100 and XL H1 prove her point. The less pixels to compress to HDV, the better it gets, especially since it appears only 24 discrete frames and it's deinterlaced, softening rez, so it should put the least strain on HDV compression.

DVCPro-HD is obviously also comrpessed but we all agree that is a kinder, gentler compression.

But I despite all the number haggling, I have not seen a clear full workflow analyses of pixel from CCD to final NLE processing between the various cams. That would be interesting.

[Jan_Crittenden]

Jan, could you tell us if Panasonic has any plans to release the CCD specs in the future? Or is this a permanent omission?

Don't know, we may or we may not. I personally would like to see a return to resolution charts, now that really told you what you wanted to know. How much resolution does it have. The number of pixels, the quality of the prism, and the quality of the glass all contribute to the resolution. Shoot you could have 1440 X1080 pixels but lose half of the effectiveness in the prism and the glass. What really is important is resolution, that you measure through the lens and prism.

Anyhow, not sure if or when the spec will be released. I have seen the resolution charts and it is right up there with the Sony Z1.

Best,

Jan

[Jan_Crittenden]

Jan,

The numbers are straight from the spec. HDV imported can be handled in many different ways. In my NLE setup I have the choice of handling the HDV native (MP@HL-14), RGB (AVI), or 2VUY (uncompressed). Once the HDV is in the system it can be uncompressed and handled as any other uncompressed format irregardless of the compressed transport stream it was captured in.

The numbers above represent the raw resolution/color/lum and you're right there are many variables when you introduce pixel shifting. In the film transfer we are working on now with ProHD, the uncompressed frames are 1280 x 720 @ 3.51MB per frame (TGA) and the Y,Cr,Cb is as expected from the published spec.

In the end, DVCProHD is a transport stream as well. It is compressed and I'm sure anyone doing a film transfer would work in (or at least deliver) uncompressed and not leave the footage in DVCProHD to work with.

Stephen you missed my point. The JVC has 19MBs to work with every second. Since it is working in a GOP of 6 , that means that each GOP has approx 6 Mbs to make it happen. Within each GOP the I frame takes 1/3 the bit budget, which means that the other frames in the bit stream or rather the other 5 frames have 4 Mbs to use. Divide 5 into 4 and that is what you have as a per frame opportunity, or approximatel 800 Kb. If we look at DV, there is approximately 833 Kbs per frame. Looking at DVCPRO HD in 720P, each frame has 1.6 Mbs, clearly twice the amount of data and the compression from one frame to the next does not influence the picture. In 1080i it is virtually doubled as it moves from 60P to a 30i. It is the amount of data available that makes HDV stumble. It is virtually bit starved. Once you have thrown away information, you do not get it back. I think it is pretty easy to understand that 100 Mbs is better than 19 Mbs. Certainly 994
Mbs is better than 100, but there isn't anything out there portable enough to record that, yet.

The JVC algorithm is probably the most robust of the HDVs due to the smaller number in the GOP, but as you can see by the math, there is less data available than that for DV on a frame to frame basis short of the I frame, which frankly doesn't save it when it comes to a lot of motion with a lot of detail.

And the word transport stream is defined in the broadcast industry as one for the broadcasting of the final signal, when it is at its final compression level to go out to broadcast.

Best,

Jan

[stephenlnoe]

Stephen you missed my point. The JVC has 19MBs to work with every second. Since it is working in a GOP of 6 , that means that each GOP has approx 6 Mbs to make it happen. Within each GOP the I frame takes 1/3 the bit budget, which means that the other frames in the bit stream or rather the other 5 frames have 4 Mbs to use. Divide 5 into 4 and that is what you have as a per frame opportunity, or approximatel 800 Kb. If we look at DV, there is approximately 833 Kbs per frame. Looking at DVCPRO HD in 720P, each frame has 1.6 Mbs, clearly twice the amount of data and the compression from one frame to the next does not influence the picture. In 1080i it is virtually doubled as it moves from 60P to a 30i. It is the amount of data available that makes HDV stumble. It is virtually bit starved. Once you have thrown away information, you do not get it back. I think it is pretty easy to understand that 100 Mbs is better than 19 Mbs. Certainly 994
Mbs is better than 100, but there isn't anything out there portable enough to record that, yet.

The JVC algorithm is probably the most robust of the HDVs due to the smaller number in the GOP, but as you can see by the math, there is less data available than that for DV on a frame to frame basis short of the I frame, which frankly doesn't save it when it comes to a lot of motion with a lot of detail.

And the word transport stream is defined in the broadcast industry as one for the broadcasting of the final signal, when it is at its final compression level to go out to broadcast.

Best,

Jan
Of course we are comparing apples and oranges with DIF (I frame) and MPEG2 (IBP). DVCProHD is 40Mb when shot at 720p24 and the compression is DCT CBR. This is not the most effecient compression sceme (albeit, tried and true). HDV1's compression sceme (at least what has been hardwired to the HD-100) is vastly more efficient than I frame CBR. It's a completely different capture/aquisition approach but still, the uncompressed frames do follow the spec (as listed on the second post).

I do know this for certain, the ProHD content I work with has been excellent at little (or no) cost to hardware requirements over a DV25 workflow (when working native MP@HL). Anyone would have to give JVC their due for that encoder.

good luck..

[Luis Caffesse]

Don't know, we may or we may not. I personally would like to see a return to resolution charts, now that really told you what you wanted to know. How much resolution does it have. The number of pixels, the quality of the prism, and the quality of the glass all contribute to the resolution. Shoot you could have 1440 X1080 pixels but lose have of the effectiveness in the prism and the glass. What really is important is resolution, that you measure through the lens and prism.

Anyhow, not sure if or when the spec will be released. I have seen the resolution charts and it is right up there with the Sony Z1.

Best,

Jan


Jan,

Thanks for the response.
I completely agree by the way, and plan on shooting a res chart with the HVX as soon as I can get my hands on one.

[stephenvv]

HDV1's comprsssion sceme (at least what has been hardwired to the HD-100) is vastly more efficient than I frame CBR. It's a completely different capture/aquisition approach but still, the uncompressed frames do follow the spec (as listed on the second post).

I do know this for certain, the ProHD content I work with has been excellent at little (or no) cost to hardware requirements over a DV25 workflow (when working native MP@HL). Anyone would have to give JVC their due for that encoder.

good luck..

I'm not sure about several points. One, has anybody torture tested the HD100 compression (I have yet to see native rez footage of anything the HVX200 demo footage).

Two, HD100 is on paper a "more efficient" compressor but again, I have not see actually tests confirming the exact amount of improvement vs. info lost (i.e. what parts of the lossy compression are visually apparent vs. not visually apparent).

Three, is there any evidence that JVC or Canon has a better encoder than Sony? Or it just because 24p/24f is much less data to compress and/or a lower rez image etc.

The bottom line is much of what passes for information is still speculation until some people can really test and compare these cams. It took a few years for DV cams to be thoroughly tested and understood (and yet, myths still prevail).

With all the interest in these cams, I expect it not to take that long, but still, it would be nice to flag stuff as speculation until there is hard, reproducible data to back it up.

[stephenlnoe]

Three, is there any evidence that JVC or Canon has a better encoder than Sony? .

Do you edit HD-100, and Z1 and DVCPro content?

I can't speak for Canon's encoder (yet), however I can for HD-100, Z1/FX1 and DVCPro. The DVCPro is great, but so is the the ProHD. Everyone is eagerly waiting to see what the 13x Fuji lens brings to the HD-100.

[stephenvv]

Do you edit HD-100, and Z1 and DVCPro content?

I can't speak for Canon's encoder (yet), however I can for HD-100, Z1/FX1 and DVCPro. The DVCPro is great, but so is the the ProHD. Everyone is eagerly waiting to see what the 13x Fuji lens brings to the HD-100.
I think you are misunderstanding me. I'm asking for scientific tests not claiming to have done them already. By "evidence" I mean actual tests not anecdotal information.

[stephenvv]

Do you edit HD-100, and Z1 and DVCPro content?

I'm not a editor, I'm a filmmaker. I've not bought any HD camera yet, though my current film I'm posting is 1080p uncompressed (long story). I have downloaded all the Z1/HD100/XL H1 native files available online and run them through Vegas over the past year. I was unable to grab the HVX native clip before it was yanked.

But the test I'm describing require all 4 cameras in the same location shooting the same setups with controlled settings, running them through both HDV, HD and uncompressed workflows and output to both tape, digital HD and film with rez and artifact tests at each stage.

Hopefully, American Cinematographer or Barry or someone else would conduct such a test. It would be very useful information but time-consuming to conduct.

[Barry_Green]

I'm not sure about several points. One, has anybody torture tested the HD100 compression (I have yet to see native rez footage of anything the HVX200 demo footage).
I've put the HD100 and the FX1 in some common but challenging situations. The HD100 definitely held up better than the FX1, as it should -- the FX1 is processing 47 million pixels per second, the HD100 is handling less than half that amount -- about 22 million. So the HD100 has a lot more bits-per-pixel, which gives it more robustness, plus it's using the more efficient progressive encoding vs. the inefficient interlace encoding. The JVC isn't impervious to artifacting, Graeme got massive artifacting by just panning across some leaves, and I've seen odd quilting artifacts even in JVC's own demo footage. But in most situations I'd say the HD100 was significantly more resilient and more reliable than the Sony.

As for the Canon, they showed some challenging scenes on their demo reel, and it looked quite robust as well. In theory the Canon should be about as robust as the JVC -- in 24F mode it's encoding progressive, and the 24F mode is somewhere around 20 to 30% lower resolution than 60i, so there's less high-frequency detail (which codecs hate), and on top of that it encodes in 24 discrete frames, so there's 20% less data per second for it to encode. The result is a significant jump in bits per pixel: in 24F mode the Canon is effectively handling about 28 million pixels per second: about 30% more than the JVC, but the Canon also has about 30% more bandwidth than the JVC. Net result is that 24F looks pretty good and should be comparably resilient.

[stephenvv]

Barry:

I've seen that as well - but that does not address who's encoder is more efficient which I think only side by side testing might though the differences in pixels per frame in the different cams make it tough to judge.

Also, I noted hdforindies.com did not like 24F footage, though the text seem to indicate more complaints about highlight handling.

Still, from what I've seen the HVX200 compressed looks equal to or better than any HDV native to date.

But it's all still speculation until we can see some good side-by-side test with native footage.

[Tzedekh]

The andromeda mod . . . will not be done to the HVX.
Who told you that? I don't think Reel Stream's Juan Pertierra has ever said that. He has said that, considering no one outside Panasonic knows much about the HVX -- particularly the CCDs -- he doesn't know whether it's practical or worthwhile to Andromedize an HVX.

[Barry_Green]

Also, I noted hdforindies.com did not like 24F footage,
He took that back though. Said that Canon blew it when they made their demo footage, so he's taken back his complaints and is waiting to get more hands-on time before rendering a final decision.

though the text seem to indicate more complaints about highlight handling.
Highlight handling is probably the weakest area of the XLH1's image, as far as I could tell from their demo footage.

[yellowdog]

I cant believe you guys actually read that entire first post. I lost interest after I scrolled down to see how long it was. :)

[Luis Caffesse]

Who told you that? I don't think Reel Stream's Juan Pertierra has ever said that. He has said that, considering no one outside Panasonic knows much about the HVX -- particularly the CCDs -- he doesn't know whether it's practical or worthwhile to Andromedize an HVX.

I may have spoken too soon.
It was my understanding that there were no plans for an HVX mod.
But, who knows?

I suppose it's possible.

[BobDiaz]

I cant believe you guys actually read that entire first post. I lost interest after I scrolled down to see how long it was. :)


Sorry, but I teach Electronics and Computer Hardware technology at El Camino College, Torrance CA, so I tend to ramble on and on and on and ... :grin:


Bob Diaz

[Emanuel]

Hi Bob!

You're welcome to this forum!

Don't worry (I'm sure you don't!) about ramble on and on and on and... 'cause there is who reads it like this one dvxuser.com member! Your knowledge is so much appreciated!...and we must thank you, as this community contributor. Just be yourself.

:) Emanuel

[yellowdog]

dont apologize, my attention span is very short. I'm sure you have a lot to offer to this forum.

[BobDiaz]

To Yellowdog: Thanks, I know that my mind is off in another world thinking about bandwidth, 3dB points, and other things that engineers ramble on about. I have a sense of humor about myself, so don't be surprised if I joke about being in another world 1/2 the time; they all know me there anyway. ;-)

One time before a NAB show someone saw me carrying several boxes of sodas to my hotel room. Seeing the weird look, I told the person that I did NOT like paying the rip-off prices for sodas at the trade show. The man answered and said, "That proves you are an engineer." I answered, "That's correct and I'm a cheap one too!" The man answered, "That proves you are a Ham Radio Operator too." He was correct both times.

Bob Diaz

[Emanuel]

:)

[Bruce Morgan]

Bob
Once again
I am interested in YOUR assesment of the Non Panasonic modifications being attempted by real stream as a way to upgrade dxv100's or even (later On ) such things as the hvx200.
I am left without a complete understanding of what "new" cameras are supposed to accomplish in terms of resulting images because of
1)No film out test for the hvx200
2) the gap in explaining the aquisition to recording path of the hvx .
and my concern about compression codecs both known and unknown .
Thanks for any speculations or opinions
Cheers
Bruce

[Bruce Morgan]

Sorry, but I teach Electronics and Computer Hardware technology at El Camino College, Torrance CA, so I tend to ramble on and on and on and ... :grin:


Bob Diaz
:thumbsup: :thumbsup: Bob your post is great and educational .I have some questions still up and Ill reclarify the question as follows .
What if they modified the new hvx200 to take andromeda
would we see images that were better
than
dvx100a with andromeda ..
I believe it is on the topic as this experiment could show and define what the hvx200 is doing at the first stage of image capture .
Thanks
This is directed to Bob Diaz .
Cheers
Bruce

[BobDiaz]

Hi Bruce,

I know very little about Andromeda, but thanks to Luis Caffesse for posting a link, I could see what it is about.

Current SD (Standard Defination) Digital Video in the USA is 720x480x29.97 frames per second. If we capture exery pixel as 3 bytes, one for each color (4:4:4), this comes to about 31MB (MegaBytes) per second. At 8 bits per byte, that comes to roughly 249Mb (Megabits) per second!!! That is one heck of a lot of data. A P2 card can grap data at that rate, but unless you have an interleaved hard drive array (RAID), that's just too much data. Go from 8 bits per pixel to 10 bits per pixel and the numbers increase by 1.25 times!

If we increase the resoltuion to 1280hx720v at 24P, the data rate goes up a little more than 2 times. At that rate we are reaching the upper limit of a P2 card. Go to 1920h x 1080v at 60i or 30P and the data rate is 6 times higher. This is well beyond even a P2 card. So my personal feeling is that with current technology, an Andromeda mod for the 200 in HD is NOT practical; just too much data to deal with.

I've never seen the Andromeda mod for the 100, so I have no idea how much improvement there is or how it compares to the 200 shooting HD. My gut level GUESS is that that the 200 will win, because it was designed for HD rather than SD.

What I do know for a fact is that if you go to:


http://www.salatar.com/hvx200/


and look at the still images from the HDX-200, the images are sharp. Look at HVX720-3.JPG you can see single hairs on the woman's face.


Bob Diaz

[Illya Friedman]

When we shot our test of the HVX200, the post facility had a Z1U handy and we set it up side by side with the HVX and shot Century sharpness charts. There was no question that the HVX in 720p mode produced a sharper image than the Z1U in 1080i.

Hope that helps,

I.

[Stevet]

Thanks Illya
That's good news!


Steve

[Bruce Morgan]

Bob Thanks!
Great info!
it gets me further along the road to knowing when to buy in and why .
i am very interested in seeing all these cameras do thier thing while projected digitally .And equally interested in seeing a PROJECTION PREFORMANCE TABLE !!!
Jan does this sound unreasonable ?
For the dvx100a I was able to see (subjectively speaking) a digital projection
which I term "acceptable " on a 20 foot screen and 30 feet .
Again this projection tab le could be produced by Panasonic(Jan) ,Adam or Barry .
I wish there were specs on possible projected line values that could be published with the rest of the camera specs .
Theaters ,even small ones are still an important venue for work done on the sd and sd (uprezed ) as well as true hd motion picture work
be it feature or documentary .
Thanks again for the response guys .I do appreciate it .
Cheers
Bruce

[toke lahti]

Illya, could you post those chart pics?
Did you compare hvx 720p vs. 1080p?

[Emanuel]

Good questions, toke! Illya are you there? Or someone else with a comparative chart?...

[Illya Friedman]

Hi Toke,

It was easy to see A/B on a 9" BVM HD monitor so we didn't record the Z1U. I know there will be plenty of side-by-sides in the coming months. Comparing the HVX and Z1U is like comparing the DVX and PD150, they're different tools.

The HVX charts we shot are all on data DVDs. Everyone in post is really busy with the holidays right now. However, if someone wants to host the HVX charts I may be able to get some tiffs exported in early January.

I.

[Emanuel]

Thanks Illya! BTW, any news about new 35mm* screening tests?

[(for who doesn't know LINK (http://www.dvxuser.com/V3/showthread.php?t=41095)) *film-out from HVX footage]

[toddeastman]

If Panasonic does not want to give away the res of the chip, how about telling us the actual size of the chip? A horizontal and vertical size is needed to calculate an accurate DoF and object size paired with a lens. Film is easy....ASC manual. Digital seems to be difficult in getting the info. The reply "it's a 1/3 inch chip" just doesn't cut it.

[BobDiaz]

Here's a rough estimate of the size of the chip:

The layout is 16:9

Squareroot( (16x16) + (9x9) ) = 18.3575597506858192984917195187072

The diagonal is roughly 18.358

From here, everything is a ratio of all the numbers.

1/3" x 16 / 18.358 = 0.2963" = 7.526mm horizontal size

1/3" x 9 / 18.358 = 0.1634" = 4.151mm vertical size


The numbers amy be a tad off, because the chip diagonal may not be exactyly 1/3".


Bob Diaz

[thefilmaddict]

A resolution comparison was done and the results were posted on DVInfo Net. I'm not sure how accurate these numbers are so I'm not making any judgements, but I would love to know what you guys think about the following info:

Resolution Comparison:

Canon XL H1 from Shannon's chart http://www.cinemahill.com/hidef/xlh1...s/IMG_0115.jpg
800Hx650V progressive
800Hx800V interlaced from German magazine, not from Shanon

Panasonic HVX-200 from Kaku's test http://home.arcor.de/martin.doppelba...rts/0088YP.png
625Hx600V progressive

JVC HD-100 from Martin's chart http://home.arcor.de/martin.doppelba...e%20Blende.jpg
700Hx525V progressive

Sony FX-1 from Martin's chart http://home.arcor.de/martin.doppelba...P_Off%20f4.png
650Hx775V interlaced

[toddeastman]

Thanks Bob! I wouldn't have know where to start. :)

[Benton]

A resolution comparison was done and

Panasonic HVX-200 from Kaku's test http://home.arcor.de/martin.doppelba...rts/0088YP.png
625Hx600V progressive


Sorry but I'm too young to know...
Is under 700 legally HD?
or will I get carded ...?

[BobDiaz]

A resolution comparison was done and the results were posted on DVInfo Net. I'm not sure how accurate these numbers are so I'm not making any judgements, but I would love to know what you guys think about the following info:

Resolution Comparison:

Canon XL H1 from Shannon's chart http://www.cinemahill.com/hidef/xlh1...s/IMG_0115.jpg
800Hx650V progressive
800Hx800V interlaced from German magazine, not from Shanon

Panasonic HVX-200 from Kaku's test http://home.arcor.de/martin.doppelba...rts/0088YP.png
625Hx600V progressive

JVC HD-100 from Martin's chart http://home.arcor.de/martin.doppelba...e%20Blende.jpg
700Hx525V progressive

Sony FX-1 from Martin's chart http://home.arcor.de/martin.doppelba...P_Off%20f4.png
650Hx775V interlaced



Because the results come from 3 DIFFERENT sources, the numbers are questionable. We have 3 different test setups and the exact same camera would yeald 3 different readings.

Untill Barry can test all 4 cameras under the exact same test setup, we are jsut making wild guesses.

Bob Diaz

[Mr. Blonde]

Indeed we are. You have to remember, Kaku himself was quite confused as to how to perform the test, he admitted it himself, he had no idea what he was doing when it came to the res test and needed assistance.

[Barry_Green]

The numbers amy be a tad off, because the chip diagonal may not be exactyly 1/3".
Numbers are way off, because the way CCDs are measured they're nothing like 1/3". No 2/3" CCD measures 2/3" in any dimension; no 1/2" CCD is anywhere near 1/2" in size, and no 1/3" chip measures 1/3" at all.

You can find various resources on the net to tell you exact dimensions on almost all CCD sizes; panavision new zealand has a calculator for that too. Nobody seems to have a published size for a 1/3" 16:9 chip though. My best guess is that it would measure about 5.25 x 2.95 mm. And yes, that's for the Canon, JVC, Sony and HVX -- they're all going to be the same size.

[Bill__Turner]

Numbers are way off, because the way CCDs are measured they're nothing like 1/3". No 2/3" CCD measures 2/3" in any dimension; no 1/2" CCD is anywhere near 1/2" in size, and no 1/3" chip measures 1/3" at all.

You can find various resources on the net to tell you exact dimensions on almost all CCD sizes; panavision new zealand has a calculator for that too. Nobody seems to have a published size for a 1/3" 16:9 chip though. My best guess is that it would measure about 5.25 x 2.95 mm. And yes, that's for the Canon, JVC, Sony and HVX -- they're all going to be the same size.


Barry, no suprise, you are correct about the chip size. The convention of referring to CCD as 2/3" etc comes from the days of tubes when the outside diameter of the tube was approximately that size (others were 1 1/4" , 1" etc) . the active size of the sensor has always been much less (even in the tubes) with 2/3" having a diagonal of approx 11mm which is less than 1/2" .

Bill Turner
Schneider Optics
Century Division

[Luis Caffesse]

Thought it might be interesting to revive this thread - seeing as the specs are now released. Interesting reading, I think.

[SPZ]

Luis, where are the specs?

[Owen]

http://www.dvxuser.com/V6/showthread.php?t=47570

[Emanuel]

Or more directly -- therefore, see this page, Sérgio:
http://www.dvxuser.com/V6/showthread.php?t=47570&page=11

posts #105 & #110