Announcement

Collapse
No announcement yet.

First impressions of the HVX

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    Originally posted by harddrive
    Regarding all the comments about gain ("I never use more than 0db" etc) it's worth mentioning that "0db" is not absolute - it doesn't refer to a precise signal to noise ratio but is rather an arbitrary manufacturer chosen figure. One camera may actually be less noisy with 3db than another with 0db, and similarly two cameras each set to 0db and the same aperture etc may give widely differing exposures.
    Good point,
    This is the reason I would like to have a -3dB gain option. If you had enough light, you could give up some sensitivity for lower noise.


    Steve

    Comment


      Originally posted by The Machinist
      It's not just memory.

      This point has been brought up before in a previous thread and i think pookie and Barry had a long discussion (which was alot of fun for the board spectators by the way).

      In the end certain people will rationalize the cost:stock ratio and if you are familiar with shooting on film stock it is comparable (and i am using the term comparable loosely here). I am not saying its the same thing but for some of us it is just a way to justify the costs of shooting on P2 "stock". But as time passes there should be much much cheaper and better options however for now the people who don't want to wait will attempt to intrepidly adopt this new disruptive tech now. Time will tell ultimately.

      To return to the point at hand i don't think it's completely unfair to compare film stock and P2. There are similarities, however I don't think anyone is saying they are the same thing.

      I believe to compare digital still photography with digital video in terms of storage or stock capacity is to err. The mediums aren't as comparable. To this guy, at least, it's apples and oranges.

      edit*** I see Haakon has responded also. I'm with him on this one. I think it's fair to compare the capacity of film and P2.
      It is just memory, and you can not compare it with film stock. I takes a little time to explain. P2 is 4 SD memory cards combined, with raid software and software tools to manage the clips. On a regular 1 Gig SD card in your photo camera you can also shoot video but with lower data rates.

      Film stock has its own characteristic, it changes by chemical processing, it is not at-random, there are within the same length a lot af flavors. There is intermediate negative, there is answer print, there is Fuji there is Kodak... and what else do I have to explain about what film is? A P2 card gives you a data rate and it can hold data. ( 4x1 SD card =4Gig 4x2 SD card = 8 Gig) , that´s it. No hocus-pocus frame rate ´quality filmlook´ comes out the P2 card, it comes out software/ computer.

      Sure you can say film 35mm/24bps holds 5 minutes = 138 meters and for 16mm/24bps holds 5 minutes = 55 meters. And digital data with frame rate x will give you x minutes on 4 Gig. But is that all you find as similarities?

      I can say film weights more, has to be processed, uses chemicals, has is own characteristic, is not reusable, is not at random, is not a digital domain, is a also a distribution format etc.
      Look outside...Do you see noise?

      Comment


        Originally posted by Haakon
        I compared P2 capacity with film stock. In that sense, they are comparable.

        The other thing that can be compared is the resulting workflow, which is also not tangible.

        As far as quality, I was comparing 16 and 35 to SD and HD, not P2. That's clearly stated.
        Similarities are almost not there. Workflow is completely different, but then I mean completely. Quality by film is the right stock choice, laboratory, combined with all that is named cinematography. What is SD? What is 16mm? You can not compare that. Which 16mm stock you want to compare with standard resolution?
        Look outside...Do you see noise?

        Comment


          Originally posted by harddrive
          Not true, in 720p mode 50Hz models record 1280x720 (luminance)
          Where did you find this info? It's been hard to come by in the US, and as I understand the VariCam used in Europe is the exact same model as used in the US, so the codec is the same 960x720...

          so if 24fps is an absolute must, the 1080p mode must be better.
          That's what I say. Data rate is the only advantage to 720/24p.

          On the original point, (comparing 1080 50 and 60Hz models) I disagree that it wouldn't tell much. Your own measurements indicate that the front end of the camera resolves more than 960 horizontally. Repeated with a 50Hz model they would tell whether the lens/chip combination is capable of better than 1280. Surely a yes or no to that is quite significant?
          I'm talking about seeing a difference between 1280 wide and 1440 wide. I doubt that there's going to be a noticeable difference.

          If there's truth to the 960 vs. 1280 wide, yes that would result in a noticeable difference. I just don't know about that though. I mean, even in the Avid DV100 codec options there are two choices for 1080 (60i or 50i) but in 720 there's only one choice. I thought 720 was worldwide standardized...
          ..
          The AU-EVA1 Book - The DVX200 Book - The UX180 & UX90 Book - Lighting For Film & TV - Sound For Film & TV

          Comment


            Originally posted by Stevet
            This is the reason I would like to have a -3dB gain option. If you had enough light, you could give up some sensitivity for lower noise.
            It may not be as simple as that - at the very least you may be giving up some highlight handling ability as well as sensitivity, and you may find the minimum noise threshold is set by other components elsewhere in the chain anyway.

            The chips will only give increased output for more light up to a certain limit. For nominal peak white the chip output should be well below this, and the headroom is usable for making a knee charecteristic. Decrease the gain too far, and it may be found that the reduced headroom becomes a disadvantage.

            My camera does enable -3db, but I can't say I find the noise difference a reason to use it (0db is for all practical purposes, "clean"). Then again, the difference due to 2/3" chips is such that there is much more to play with, the S/N at 0db can be better than is really necessary for simply not noticing noise. Hence even 9db can be put in and the result still be very acceptable for most use.

            Comment


              In a future HVX wouldn´t it teoretically be possible to clock down the chip to 48p to get 108024p with the better 50Hz-resolution? And wouldn´t it be interesting to have 49p to get 24,5p-images - just for the case yo want to produce something for the 50 AND 60 Hz-world? You just had to speed it up or slow it down in post a mere 2% to make it fit for both standards.

              Comment


                Originally posted by Barry_Green
                I'm talking about seeing a difference between 1280 wide and 1440 wide. I doubt that there's going to be a noticeable difference.

                If there's truth to the 960 vs. 1280 wide, yes that would result in a noticeable difference. I just don't know about that though. I mean, even in the Avid DV100 codec options there are two choices for 1080 (60i or 50i) but in 720 there's only one choice. I thought 720 was worldwide standardized...
                Regarding the first paragraph, it may not make a lot of difference viewing wise, but I'm really referring to it being educational - giving more of an insight into how good the chip/lens combination is.

                I've hunted for a while now for a reference about DVCPRO HD horizontal resolution at 720p/50 and failed to find anything one way or the other. (Lots about DVCPRO HD 720 for 60Hz, very little for 50Hz.) Regarding 720 being worldwide standardised, then if there exists a standard it's 1280x720 - 16:9 square pixels. I'm sure the rationale went along the lines that 1280 for 50Hz and 960 for 60Hz gave similar clocking rates for DVCPRO HD.

                Up until quite recently I don't think there was an official 720p/50 standard - 720p meant either 60p or 24p, period.

                Comment


                  Regarding DVCPRO-HD: the 4:2:2 color sampling comes from the notion that there are two codecs working together (hence DVCPRO50 and its 50mb data rate); the high-def frame uses four codecs, and the frame is exactly twice as many pixels, so it all works out and yields the data rate you would expect. Four codecs, each at 25mbps, = 100mbps data rate. The frame size for DV is 720x480, the frame size for HD is 720x960, exactly double.

                  In 1080, it's the same thing -- the frame is 4x the size of the SD frame. 1280x1080 = 1,382,400 pixels; an SD frame = 345,600 pixels; 1382400 / 345600 = 4. It's exactly 4x the # of pixels.

                  So how 50hz DVCPRO100 could be 1280x720 doesn't seem to follow the pattern. With 1080, they went to 1440 pixels, presumably taking the frame rate savings and applying those to more pixels in the frame. Haven't figured out a formula that works for that yet though; if you took 50/60 * 1440 you get 1200, not 1280... if you take 60/50 * 1280 you get 1536, not 1440... hmmm... and furthermore, the PAL DV codec being 720x576 doesn't help matters either. If you quad 720x576 you get 1,658,880; if you multiply out 1440x1080 you get 1,555,200. So the reality of 1440x1080 seems to be less than what you'd get from just running quad DV codecs anyway, without even taking into account the savings you should be able to get from the slower frame rate.

                  But with 720p it doesn't work out to go to 1280 no matter how you slice it; if it was using PAL codecs at exactly double for 720p like the US version, that'd still be 1152x720. And, again, the VariCam codec is worldwide; the VariCam in Europe uses the same 960x720 that the US version does. So I really don't expect there to be any difference in the European HVX as far as luma pixels in 720. Don't know what they do with the frame rate savings though; hopefully it results in a little lower compression ratio.
                  ..
                  The AU-EVA1 Book - The DVX200 Book - The UX180 & UX90 Book - Lighting For Film & TV - Sound For Film & TV

                  Comment


                    Originally posted by Barry_Green
                    Regarding DVCPRO-HD: the 4:2:2 color sampling comes from the notion that there are two codecs working together (hence DVCPRO50 and its 50mb data rate); the high-def frame uses four codecs, and the frame is exactly twice as many pixels, so it all works out and yields the data rate you would expect. Four codecs, each at 25mbps, = 100mbps data rate. The frame size for DV is 720x480, the frame size for HD is 720x960, exactly double.

                    In 1080, it's the same thing -- the frame is 4x the size of the SD frame. 1280x1080 = 1,382,400 pixels; an SD frame = 345,600 pixels; 1382400 / 345600 = 4. It's exactly 4x the # of pixels..
                    Whilst I knew the numbers I must confess I hadn't put them together quite like that before - but then I instinctively think of an SD frame as 720x576, so.........

                    I'm not quite so sure where DVCPRO 50 comes in though - whilst the doubling of the data rate does indeed give 4:2:2 v 4:2:0 (4:1:1 for the States?) I'd understood that the main benefit was down to most of the extra data being used to give much less general compression within the frame?

                    Looking at your figures slightly differently, I can follow it better by thinking that the first doubling turns 30i into 60p, and 4:1:1 to 4:2:2 by default, the second doubling then making 720x480 into 960x720. Very logical.


                    Originally posted by Barry_Green
                    So how 50hz DVCPRO100 could be 1280x720 doesn't seem to follow the pattern. With 1080, they went to 1440 pixels, presumably taking the frame rate savings and applying those to more pixels in the frame. Haven't figured out a formula that works for that yet though; ..........

                    And, again, the VariCam codec is worldwide; the VariCam in Europe uses the same 960x720 that the US version does. So I really don't expect there to be any difference in the European HVX as far as luma pixels in 720.
                    Regarding the last paragraph, then at the risk of getting somewhat out of my depth, I believe that at least until fairly recently a Varicam in Europe was considered more like an electronic cinema camera than a TV camera. Hence operating it at 24fps would be fine, and therefore yes, I believe the US version was the worldwide one.

                    But it's only fairly recently that 720p/50 (unlike 60) has been seen as a standard at all. The expectation was (and in most countries still is) that Europes progression to HD would be via 1080i/25 (for sport, news etc) and 1080p/25 (drama etc) with 1080p/50 as the ultimate goal. Hence any standards to do with 720p/50 must be very recent. The logic you outline above would indeed lead to a frame size of 1152 (2x576)x720. Since that would lead one to think that clock rates have to tweaked somewhere, well, 1280 isn't that far away, and has the rationale of being square pixel, the 720p "norm".

                    But could somebody from Panasonic definitively put us out of our speculative misery!?!

                    Comment


                      Spinflight,

                      If you're in the seattle area, I think it would be cool if we could put the HVX and my DSR 370 side by side, that at least would alleviate the pains of NBC Shooter. What do you think?

                      Comment

                      Working...
                      X