|
|||||||||
|
Thread Tools | Search this Thread |
November 1st, 2007, 08:18 AM | #16 |
Trustee
Join Date: Nov 2005
Location: Sydney Australia
Posts: 1,570
|
Werner,
from our tests and the comments of others 1080i produces much better slomo than 1080p or 720p for that matter and that makes perfect sense scientifically. 1080i has double the temporal resolution to start with. |
November 1st, 2007, 08:24 AM | #17 | |
Major Player
Join Date: Aug 2005
Location: Belgium
Posts: 497
|
Quote:
We'll also have to see if the CCD's are progressive 1080 CCD's - that hasn't always been the case. I believe the verical resolution of the Z1 was 1080 as well, but those CCD's weren't able of true progressive readout (ergo the 'frame' mode). Actually I'm not saying anything on the quality of the SONY, I am merely reacting to some people that have never gotten the right info on p and i. the sony and the JVC aren't very comparible if you ask me. They are both designed for a segment in the market. The 16mm and 35mm adapters for the JVC go well e.g. and there's no substitute for that with this sony. Furthermore - JVC offers no handheld option. I personally think this sony model, just as the HVX200, is way to big and bulky for a handheld type camera, but nevertheless, sony and panasonic offer the option and some people like it.
__________________
High-Definition Video Consultant - CEO of Delimex NV - http://www.delimex.be gear of choice : http://www.wespgear.com |
|
November 1st, 2007, 08:28 AM | #18 | |
Major Player
Join Date: Aug 2005
Location: Belgium
Posts: 497
|
Quote:
Don't believe what you hear - just really test it if you don't believe the numbers: 1/2 speed: 720p50 > 720p25 1080i50 > 540p12,5 (advanced deinterlacing techniques can make this a bit better, perhaps +/- 600p12,5 so to speak) if you test it: take 1080p50 fotage, render it to 1080i50 and to 720p50, then slow down both files, compare
__________________
High-Definition Video Consultant - CEO of Delimex NV - http://www.delimex.be gear of choice : http://www.wespgear.com |
|
November 1st, 2007, 09:08 AM | #19 | |
Inner Circle
Join Date: Feb 2004
Location: Brooklyn, NY, USA
Posts: 3,841
|
Shooting in 24p or 25p requires a different (improved?) set of skills due to the motion. Too many people think it's "flip a switch" and you can move the camera as before.
The EX1 is CMOS, not CCD BTW. The chips and the processing are Progressive coming from 1920x1080 (but there are various shooting modes). For me, I need to shoot hand held frequently so I like the form factor. I've used shoulder mount and have found that awkward in some circumstances. I don't think either is inherently "better." I think it's a matter of shooting style and technique. It looks like the EX1 rotating control grip should help wrist fatigue a bit. Quote:
|
|
November 1st, 2007, 10:19 AM | #20 | |
Major Player
Join Date: Mar 2005
Location: Northampton, England
Posts: 500
|
Quote:
I've been working with 25p for years with barely any special consideration. It would be just like going from 60i to 30p... It practically looks the same (only a little better IMHO)!
__________________
Alex |
|
November 1st, 2007, 11:01 AM | #21 | |
Inner Circle
Join Date: Feb 2004
Location: Brooklyn, NY, USA
Posts: 3,841
|
Not all 24p involves pulldown. Both the HVX200 and EX1 have modes that record 24p (23.98) natively to cards. It's one of the advantages of a card (solid state) based workflow.
One can certainly work from acquisition to master without using 29.97. Of course if you're heading for NTSC broadast the pulldown is going to happen at some point though. Quote:
|
|
November 1st, 2007, 02:10 PM | #22 | |
Trustee
Join Date: Mar 2004
Location: Milwaukee, WI
Posts: 1,719
|
Quote:
VBR just means it can adjust the bitrate depending on how complex the scene is. VBR 35 mbits will look just as good if you had CBR 35 mbits. VBR was more of a space saver then a quality booster. The only time the quality gets better is when you have to lower the bitrate. For example a DVD at 4.5 VBR will mostly still look as good as 4.5 CBR. It is just when it needs it it will bump the bitrate up to help. With XDCAM the 35 VBR is the max so most of the time you might end up with bitrates lower then 35 mbits such as 25 or 30. So yes it looks better but not just because of VBR it looks better mainly for the reason it can go up to 35. 35 at CBR would look just as good but it would have been a waste of space because not all scenes would need 35 mbits. That is the key to VBR. Easy scenes can use lower bitrates. HDV isn't bad because it uses CBR. It is bad because sometimes 25 just isn't enough and it needs a little extra umphf. With that said another thing I would like to point out is that with 35bmits on the EX1 it has to encode 1920x1080 instead of 1440x1080. It still uses a better quality encoder chip (I hope) but I think the 35mbit mode is going to look compression wise more like the 25mbit mode on the higher end XDCAM HD cameras. 1920x1080 has 1.33x more data to deal with so that almost cancels out the extra 1.4x worth of bits. Of course it isn't an exact science so it will still look better then 25 mbit HDV because most of the time a 1920x1080 image isn't going to have 100% unique pixels. My whole point is that 35mbits on the EX1 does have to work harder then 35mbits on other XDCAM HD cameras so do not expect a 1:1 quality compression between those cameras. Sure the EX1 is nicer because it is 1920x1080 but it will be slightly more compressed. And that is why I may prefer shooting 720p with the EX1. Some of you may have seen my thread on 720p 24p and how awesome it is going to look on the EX1. Well even 60p or 50p is going to look great. Progressive is much easier to compress and 35mbits is about double of that of normal 720p broadcasts or what the JVC cameras use for their bitrates. |
|
November 1st, 2007, 02:24 PM | #23 | |
Trustee
Join Date: Mar 2004
Location: Milwaukee, WI
Posts: 1,719
|
Quote:
One thing people forget about interlaced video is that it has to low pass filter the video so there isn't any interlace flickering. So yes you may have 1080 lines but those 1080 lines are slightly softened to the point where they only have a tiny tiny bit of extra detail in still scenes. Like you pointed out most 1080 video even from HDCAM tape is 1440 which isn't all that much higher then 1280. Although to be fair until there was HDV 720p tape only used 960x720 pixels which is why some people may notice a lack of detail in some 720p shows. True 720p however is just as good if not better then normal 1080i video. Sure 1920x1080 may have a little bit of an edge over 1440x1080 but it is still interlaced and it still has a low pass filter. Even shooting progressive with a 1080i camera you have to be carefull. Sure the video may be progressive but a lot of HDTV's and a lot of HDV formats will still play this back as 1080i. 1080p material still has to have low pass filters or as soon as you watch it as 1080i you will get flickering. Again why 1080i or 1080p can give a lot of inconsistant results. I like to think of HD in terms of a sliding scale. On one end you have dirty detail and the other end you have clean softness. It kind of depends on what you like but to me quality equals clean video with as few of artifacts as possible. I didn't hear a lot of people complaining about the World Series which was on FOX which uses 720p. In fact it looked pretty darn good. |
|
November 1st, 2007, 02:27 PM | #24 | |
Major Player
Join Date: Feb 2006
Location: Philadelphia
Posts: 795
|
Quote:
What your numbers describe is de-interlacing first to produce a progressive frame and then slowing it down - but that's simply a bad workflow, not a limitation of interlaced video. The deinterlacing should occur naturally as part of the process of slowing it down. Drop 50i/60i footage into AE, tell it to separate fields and conform to 50/60fps in the interpret footage dialogue and each field will be scaled to a progressive frame, retaining your full temporal resolution at the expense of up to half of your spatial resolution (depending on motion levels). Now you're back to comparing 50x1280x720 to 50x1440x540, which is a 15% difference - certainly significant but close enough that the I'd argue the deciding factor in which looks better comes down more to the lens and imaging system of the cameras than the format.
__________________
My latest short documentary: "Four Pauls: Bring the Hat Back!" |
|
November 1st, 2007, 02:43 PM | #25 |
Regular Crew
Join Date: Mar 2007
Location: Weatherford, Texas
Posts: 79
|
35MBS vbr vs. 25MBS cbr Question
Would someone exlpain the definition and diferrence between 25mbx CBR and 35mbs VBR terminology. Thanks.
Joel |
November 1st, 2007, 03:23 PM | #26 | |
Regular Crew
Join Date: Jun 2006
Location: NJ / NY Tri-State Area
Posts: 31
|
Quote:
Keep in mind that the EX uses only 2 channels of audio whereas the XDCAM HD uses 4. The 2 channel difference may help even the score you are talking about above. Cheers, G |
|
November 1st, 2007, 03:34 PM | #27 |
Major Player
Join Date: Mar 2005
Location: Northampton, England
Posts: 500
|
And if XDCAM HD on the EX1 behaves anything like XDCAM on the F330/350/355, etc. then (according to a Sony Engineer who talked to Greg Boston) that the data rate can actually peak slightly above 35mbps.
__________________
Alex |
November 1st, 2007, 04:24 PM | #28 | |
Major Player
Join Date: Feb 2006
Location: Philadelphia
Posts: 795
|
Quote:
25mbs CBR means every second of video consumes 25 megabits - no more, no less. 35mbs VBR means every second of video may consume up to 35 megabits, but no more. VBR primarily affects recording capacity - a static talking head shot may not need the full 35mbs and therefore your recording time may increase. The picture quality difference comes mostly from the increased total data - 35mbs is 40% more data per second than 25mbs. However, as others have noted, the EX's full raster (1920x1080) mode has 33% more pixels than 25mbs HDV - so the difference might seem like it's not that big. However modern compression doesn't scale linearly with resolution - a big part of the compression is eliminating redundancy between frames. So your I frame (first frame in your group of pictures, GOP) may be 33% larger, but the differences in the remaining frames are not necessarily that much larger than the corresponding frames in an HDV-resolution file, so the net quality improvement may be higher than the numbers would indicate. That gets hard to estimate though because so much of it depends on variables in the image itself like image detail, detail movement, and camera movement - so we can guess all we like but until we have a lot of sample footage under a variety of shooting conditions we won't really know. If what Alex mentioned is true about peaking it may be that 35mbs is really an average data rate, not the maximum, and that situations where a scene moves between static and motion shots may be able to 'bank' data not needed during the simpler portions and apply it to go above 35mbs where needed as long as the average data rate over a given period (probably a few seconds) doesn't exceed 35mbs.
__________________
My latest short documentary: "Four Pauls: Bring the Hat Back!" |
|
November 1st, 2007, 04:32 PM | #29 |
Trustee
Join Date: Mar 2004
Location: Milwaukee, WI
Posts: 1,719
|
The audio plays no part in that bitrate. In fact the audio is uncompressed and sits apart from the video bitrate. You could have 16 channels of audio and the video bitrate would still be the same.
|
November 1st, 2007, 04:34 PM | #30 |
Trustee
Join Date: Mar 2004
Location: Milwaukee, WI
Posts: 1,719
|
yes peak slightly but not by a whole lot. Maybe 1 or 2 mbits/s is what I heard.
|
| ||||||
|
|