Press Release: HMC150, "the new DVX"
- Thread starter Barry_Green
- Start date
Awwwww. I sort of guessed this but wasn't truly "in the know" on it
)Relative to this diagonal positioning, is there any advantage or disadvantage of it next to the HVX200, 150 and 170? If there is/isn't? What are the differences?
When we look at images captured from the above mentioned styles of cameras (other than the 150 and 170 which we haven't seen yet) it's hard for us to find the differences that are supposed to be there and are wondering if these differences are apparent or technically aesthetic in value?
Aroon_Narayanan
Well-known member
Well, all of Sony's ClearVID sensors are CMOS, so they've have a rolling shutter, as opposed to Panasonic's global shutter. Also, the Sony cams have to interpolate the video from the sensors, seeing as the pixels are rotated 45 degrees. I'm guessing that there isn't that much difference in terms of sharpness. The EX1, with its native chips, probably beats all aforementioned cameras.
ullanta
Veteran
I'm sorry if I've missed the answer to this, but... have we had any confirmation on the issue of whether the camera is efficient in use of different frame rates... that is, does 720/24p use only 40% of the data space used by 720/60p; or better yet if it has a lower compression ratio for 24p than 60p?
Most digitals have a global shutter as opposed to a rolling? Or, is it a mix? Or, the other way?
The image that we've seen off this CMOS ClearVID looks good and actually cleaner than that which we seen off the Canon A1. Moreover, the image on the LCD screen more closely looks like the video as opposed to what we have seen on our HVX200. In fact, the DVX100B more closely represents the actual image than the 200. The real trigger to us on this from the start was that both the DVX100B and the HVX200 out of the box LCD settings were slightly bright relative to the actual video. That is, shooting the image look on the LCD screen resulted in video that was slightly under exposed by at least one and almost two F stops. However, it was easily resolved by just turning the LCD brightness down one or two notches and then we had properly exposed images. We actually used a Gossen meter to compare the respective F values on the meter and the camera and they were not relative.
We'll be very interested in seeing how all these HD images compare when the 150 and 170 are side by side. In fact, it is this side by side comparision that was mentioned throughout DVXuser that we had not done and having read how excellent the technique was in making an image comparison, we intentionally produced our own comparions under low, medium, and over exposed lighting to see how the images stood up against each other.
The image that we've seen off this CMOS ClearVID looks good and actually cleaner than that which we seen off the Canon A1. Moreover, the image on the LCD screen more closely looks like the video as opposed to what we have seen on our HVX200. In fact, the DVX100B more closely represents the actual image than the 200. The real trigger to us on this from the start was that both the DVX100B and the HVX200 out of the box LCD settings were slightly bright relative to the actual video. That is, shooting the image look on the LCD screen resulted in video that was slightly under exposed by at least one and almost two F stops. However, it was easily resolved by just turning the LCD brightness down one or two notches and then we had properly exposed images. We actually used a Gossen meter to compare the respective F values on the meter and the camera and they were not relative.
We'll be very interested in seeing how all these HD images compare when the 150 and 170 are side by side. In fact, it is this side by side comparision that was mentioned throughout DVXuser that we had not done and having read how excellent the technique was in making an image comparison, we intentionally produced our own comparions under low, medium, and over exposed lighting to see how the images stood up against each other.
BobDiaz
The Professor
We do know that on the HVX-200, the image generated is 1080p. From that point it is down converted to the resolution needed. Unlike CMOS, CCD must read all the pixels in order. It seems likely that the HMC-150 generate the 720p image in the exact same way the HVX-200 does.
I'm guessing your question is not how the image comes off the CCD, but do we still have the 21/24Mbps data rate at 720/24p (or 30p) OR does the data rate drop because there's less data? I wish I had the answer to this question.
It will be a long 2 months before we find out...
RE: CMOS vs. CCD:
The debate over which technology is "better" is never ending, but some general things can be said:
CCD: All pixels must be read in sequential order. Can show a vertical column is pointed at a very bright light source. Uses global shutter.
CMOS: All pixels can be read in any order. Can suffer from a rolling shutter effect. Most CMOS sensors have a rolling shutter.
Rolling shutter effects:
Sony's CMOS Sensor Layout:
Bob Diaz
Wow, thanks Bob Diaz for some really nice information. Might it be safe to assume then that this "read in any order" and/or the size/layout configuration of these CMOS is in part a reason for the excellent low light performance?
On our one CMOS camera we haven't seen this rolling shutter although we only have about 6 hours of log time thus far. Is this issue more apparent on low light? Flash/strobe? Or other ??
In addition, I don't know that we've seen a lot of vertical smear (that's what we call it) on our 200's other then when panning in contrasting light conditions (bright on one side lower light on the other) and herein since we haven't used the HVX's for weddings because they just have never cut well with our DVX100B's that do reasonably well in low light atlhough when we ran the Canon A1 recently about three weeks ago, it did quite well in low light and even when shooting high def which was rather amazing. However, we only had a total of about 600 hours across 4 cameras so we're not really that "up" on all there is about the HVX other than the "A" model is definately a lot better on image quality in low light than the older 200 non "A" model.
On our one CMOS camera we haven't seen this rolling shutter although we only have about 6 hours of log time thus far. Is this issue more apparent on low light? Flash/strobe? Or other ??
In addition, I don't know that we've seen a lot of vertical smear (that's what we call it) on our 200's other then when panning in contrasting light conditions (bright on one side lower light on the other) and herein since we haven't used the HVX's for weddings because they just have never cut well with our DVX100B's that do reasonably well in low light atlhough when we ran the Canon A1 recently about three weeks ago, it did quite well in low light and even when shooting high def which was rather amazing. However, we only had a total of about 600 hours across 4 cameras so we're not really that "up" on all there is about the HVX other than the "A" model is definately a lot better on image quality in low light than the older 200 non "A" model.
ullanta
Veteran
Right... DVCPro HD, e.g., has a constant compression ratio which yields an effective data rate of 40mbps in 720p24, 50mbps in 720p30, and 100mbps in 720p60 (and 80mbps in 1080p24, etc.) I'm wondering if the AVCCAMs will be able to change the effective compression ratio so as to maximize image quality for the different frame rates (i.e., have twice as much effective data per frame when shooting 720p30 than when shooting 720p60).
Just wondering if any of the people who've gotten their hands on the prototypes have been able to test this...
Barry_Green
Moderator
Categorically not. The diagonal layout is used on the FX7/V1U, which has hideous low light performance.
Sony's 1/2" CMOS gets about the same low light performance as Panasonic's 1/3" CCD in the 200A/150/170. Unless you're shooting 1080i, in which case the Sony is better.
Barry_Green
Moderator
It is my understanding that yes, that is the way it works. The bitrate is fixed as 6 or 13 or 17 or 21 or 24mbps, so if you are shooting a slower frame rate, it just ends up that you have less compression per frame.
AVC-Intra does it the way DVCPRO-HD does it (i.e., going to 1080/24pN in AVC-Intra does result in an 80mbps data rate). But AVC-HD, as I understand it, doesn't vary the bitrate based on the frame rate, so you could have 24mbps of 720/60p, or 24mbps of 720/24p -- with obviously 2.5x as much bandwidth applied to each frame.
Aroon_Narayanan
Well-known member
Also, the FX7 and V1U had 1/4" sensors, as opposed to most other cams. If I'm not mistaken, the Z7U has 1/3" sensors and the diagonal pixel layout. Anyone know how sharp it is or how it performs in low light?
BobDiaz
The Professor
If correct, that would put the HMC-150 at a higher bit rate (21/24Mbps) than the JVC's 720p HDV (19Mbps). Given the higher bit rate AND the better CODEC, one is going to have to work VERY hard to generate an image at 720/30p or 720/24p that has compression noise due to excessive movement.
CMOS Rolling Shutter:
In order to see the rolling shutter effect on any CMOS Video Camera, try the following:
(1) Do a very fast pan. The objects that touch one edge must reach the other edge in 1 second or less. Using the highest speed shutter will help show the effect. It is possible to do this with a speed of 3 seconds, but the slant (or tilt) is small. Some have done a left - right - left - right pan back and forth to make this stand out. The worst I've even seen was when someone put a CMOS camera on the back of a motor scooter and videoed their driving. Great driving, but the rapid up/down movement of the camera made the world stretch and shrink.
(2) Take flash pictures. In most cases, only part of any frame is exposed to the flash.
CMOS technology has improved. Most of the fixed pattern noise is removed through digital processing, resulting is a less noisy image compared to the older analog solution.
CCD Vertical Line:
Due to improvements in CCD technology, it's getting harder to produce this effect. The light has to be VERY bright compared to the rest of the image. This might be the reflection of the sun off a shiny car (think chrome reflection) or a shooting directly into a spot light pointing at the camera but at one side of the frame.
Bob Diaz
Bob, wonder if this was fixed with all the updates in the Z7U? On the Sony website there was a discussion about tweaks that had been added inside the Z7U to fix issues that were obvious in the Z1U . . PLUS . . there is a whole new video analog to digital processing technology in the Z7U according to the technical bulletin and description. I tried what you said and even shook it mildy and even held it horizontal and fanned it back and forth and then replayed the tape and nuttin'. Could this issue have been fixed in this A/D conversion processor update that was added with the Z7 generation?
I have not tried the flash/strobe tests yet but two people who have Z7's here in Pittsburgh said that they use them at weddings and have not had this problem stated with the CMOS in the Z1 which apparently has serious issues with flashes.
Just curious? I will try some flashes in a few days. Also, it doesn't show on the tape or on the card when played off the card. Is this issue only after importing? Or visible in the raw?
EDIT: Bob, here's the link describing this A/D dual noise cancelling "rolling shutter" issue by an independent source: http://digitalcontentproducer.com/hdhdv/depth/sony_hvr_series_0101/index1.html
Last edited:
ullanta
Veteran
...good news indeed! It'll be interesting to see how all this plays out!
Richard J. Johnson
Joystick Member
Why can't these things be easy. Like good, gooder, goodest. I don't even see how you guys can hold of that information in your brain. Barry you must have a firewire port in the back of your neck.
Excellent analogy "filthrich" and also amusing )
It is this reason that we changed our whole way of looking at this whole acquisition and editing process with regard to selecting cameras and came up with this analogy:
"Who cares about the technical stuff as long as: A) the picture looks great; B) the media is easy to get into a machine, edit, and output; and, C) the material is easily archivable/reusable." Dr Dimento)
Here we termed it this way:
Astetics = excellent picture
Accessible = easy to get media in, work with, and get it out
Archivable = easy to store and reuse
Took me a while to get here and although I'm still going to try to stay sort of up on what's going on and how the things tick, a camera is a camera is a camera and if it takes excellent images and we can work with them without encumberance while storage is simple and doesn't eat up a fortune or space then that's good enough which is all of those things mentioned, "good, better, best" in one space which is "synergy" or easier said 1 + 1 = 3 . . or . . the result is greater than the sum of the constituents . . it would be like voting for and getting a politician that actually did something useful for the people that voted him or her in.)
)It is this reason that we changed our whole way of looking at this whole acquisition and editing process with regard to selecting cameras and came up with this analogy:
"Who cares about the technical stuff as long as: A) the picture looks great; B) the media is easy to get into a machine, edit, and output; and, C) the material is easily archivable/reusable." Dr Dimento
)Here we termed it this way:
Astetics = excellent picture
Accessible = easy to get media in, work with, and get it out
Archivable = easy to store and reuse
Took me a while to get here and although I'm still going to try to stay sort of up on what's going on and how the things tick, a camera is a camera is a camera and if it takes excellent images and we can work with them without encumberance while storage is simple and doesn't eat up a fortune or space then that's good enough which is all of those things mentioned, "good, better, best" in one space which is "synergy" or easier said 1 + 1 = 3 . . or . . the result is greater than the sum of the constituents . . it would be like voting for and getting a politician that actually did something useful for the people that voted him or her in.
)
Richard J. Johnson
Joystick Member
Thank you! you read my mind. I am just going to get the HPX170 learn it and shoot. end of story.
ilauzirika
Boring Senior Member
I have a question that I think I asked before in this thread, but I can't find it......
[Question]
The camera films (in PAL) 720p 50, so if you take that 50 frames progressive stream into a 25 or 24 frames progressive timeline the logic says you would achieve slowmo.
So even if it isn't framerate swatchable could you get slowmo?
IS this right, or it doesn't work like thaat.
Thanks.
I'm starting to feel this camera in my hands........
[Question]
The camera films (in PAL) 720p 50, so if you take that 50 frames progressive stream into a 25 or 24 frames progressive timeline the logic says you would achieve slowmo.
So even if it isn't framerate swatchable could you get slowmo?
IS this right, or it doesn't work like thaat.
Thanks.
I'm starting to feel this camera in my hands........
BobDiaz
The Professor
The Sony Z1 used a CCD image sensor, so the CMOS rolling shutter problems don't apply to the Z1.
I understand that the Z7 has an improved CMOS image sensor that helps reduce the effect of the rolling shutter. It's unclear exactly what they did, but my guess is that the speed of the rolling shutter has increased. A look at the layout of the image sensor (see below) shows that it's possible to read 2 rows at the same time. This would take a 1/60 of a second rolling shutter and increase it to 1/120 of a second. While I have no proof, it may be a bit faster, like 1/180 of a second.
Anyone in the LA area with a Z7 who wants to play around a bit, I have a way to test the speed of the rolling shutter using some of our lab equipment at El Camino College....
The faster the shutter, the longer the time the whole shutter is open at once and the greater the chance the flash will fine during this time. One would have to take around 10 to 15 flash shots and view the footage frame by frame to spot the rolling shutter. Increase the shutter speed in the camera to 1/500 of a second and the chances increase of seeing the effect.
The faster the speed of the rolling shutter the less the impact of a whip pan on the image. If my guess is correct of 1/180 of a second for the rolling shutter to go from top to bottom, that would reduce the lean caused by a rapid horizontal pan by 1/3.
Go outdoors, set the camera's shutter speed as high as possible, put the camera on a tripod, zoom in to maximum telephoto, and spin the camera around quickly. If you compare single frames, the vertical edges to images where the camera was not spinning, you should see some lean in the image when the camera is moving.
It's a fact that with an increased shutter speed, the impact of the rolling shutter is reduced. The Sony EX-1 has a rolling shutter speed of 1/60 of a second. The images of the rolling shutter effect (several message above this one) show a flash image (partial exposure), lean caused by rail cars moving quickly through the frame, lean of a test chart caused by a rapid pan.
"Who cares about the technical stuff as long as: A) the picture looks great; B) the media is easy to get into a machine, edit, and output; and, C) the material is easily archivable/reusable." Dr Dimento
You're right, most of the technical details can be ignored. The only time these become an issue is it impacts the image, causes editing problems, or some other problem. No joke intended, because it's only when we have problems that the technical details suddenly become very important.
I have nothing against CMOS technology; like any technology it has it's limits too. However, over the years it's been improving and will continue to improve.
Bob Diaz
I C that B&H has the HMC150 in their store now at $3495 but not available yet: Click Here