Jump to content

7D resolution vs sony EX 3


Bobby Shore

Recommended Posts

  • Premium Member
The numbers between HD and SD don't correlate that way. 4:2:2 and 4:2:0 don't mean the same thing between SD and HD because HD is sampling (at least) twice as much information, and recording a higher data rate. ...

We're discussing DSLR "HD" video quality in comparison to a more "traditional" HD video cam, such as the Sony EX3. (And for fun, also compared to the live SDI out of a DSR-450WSL properly up-converted to HD-SDI.)

 

DSLRs record their best "HD" video to SD/CF memory cards at relatively low data rates (their _live_ video HDMI video output is problematic for quality external recording*), a tiny fraction of the data rate of uncompressed SDI video (let alone HD-SDI). DSLRs record "HD" video to SD/CF cards using the 4:2:0 8-bit MPEG-4 AVC h.264 codec, with no B-frames, @ <50 megabits/sec.

 

... Plus there are a lot of questions about the authors methodology for shooting the resolution chart. He didn't give any information about his method or what lens he used. He very likely could have used a better lens that could have resolved more detail.

In the article and related threads Barry says he used quality DSLR lenses in his tests. The lenses were also used to shoot the res chart as a digital still (in addition to the video), and the example still image in the article displays at least 2,000 lines. Typical DSLR lenses don't seem to be the limiting factor here -- instead these cam's poor "HD" video resolution performance primarily results from how they read their image sensor (binning) and how they compress frames to create motion video. Again, see:

http://www.dvxuser.com/V6/showthread.php?t=187503

 

I don't believe we've discussed yet the dynamic range of the "HD" video recorded by these DSLRs. It's also not stellar performance -- no surprise given the relative lost cost of these cams:

http://www.dvxuser.com/V6/showthread.php?t=187606

http://www.dvxuser.com/V6/showthread.php?t=187624

 

DVXuser.com isn't the sole source of this information; it been confirmed by others & reported elsewhere online.

 

Again, I believe these cams can be of terrific value for certain productions, and it's helpful to understand their weaknesses so as to get the "best bang for your buck" out of them. If (very) carefully handled, in some cases the video they create can be intercut with HD from many other cams, including the EX3.

 

Myself, I'm _this close_ to buying a GH1 (really)! :)

 

- Peter

 

* For example, a 7D's live HDMI output connected to a NanoFlash recorder, with frustrating results:

http://www.dvxuser.com/V6/showthread.php?t=185934

Link to comment
Share on other sites

  • Replies 80
  • Created
  • Last Reply

Top Posters In This Topic

The Canon vDSLRs are not 1080p cameras, period. Maybe 720p, but even then you are subject to horrendous aliasing if you point it at the wrong subject.

 

Have you used one of them? Or, alternatively, have you looked at some footage on a proper HD monitor (NOT as clips on the Internet as seen on a computer monitor)?

Link to comment
Share on other sites

  • Premium Member
Are they less 1080P cameras than previous HD cams that shot 1440P with 2/3 inch sensors?

 

Yes they are, very much so.

 

These cameras use huge shortcuts and workarounds. The stills sensors quite simply do not go fast enough to do video, so they end up leaving out enormous amounts of the chip - just by dividing the vertical resolution of the 5D's sensor (about 3800 lines) by the output lines (1080) we can see that it doesn't actually have to do much more than every third or fourth line, and possibly less than that. This means the OLPF is now wrong and the debayer has scanty information to work with. The other alternative would be to clock everything much faster and run it much harder and then you get noise (see: Red).

 

There's no doubt that these things would look stunning if they took the entire chip, debayered as for stills, and gave us that as video, but they don't even nearly do that and nor are they even nearly capable of doing it. Remember that the 1DS needed two DIGIC processors to do 10fps, and that's just the processing load, never mind downloading it off the sensor.

 

The other problem with the 5D is that the default picture setup is awful, just incompetent beyond belief; select whichever the softest picture preset is and wind the contrast all the way down (under C.Fn II: Image, I think but I'm not sure) and you get an extra real stop of latitude at the expense of some odd but easily fixed colour banding around highlights.

 

But to answer an earlier query: the 5616 by 3744 sensor of a 5D2, sampled in let's say 16 bit and stored as the raw bayer matrix: forty megs a frame, or just under a gigabyte a second at 24fps. But it'd look very nice.

 

P

Link to comment
Share on other sites

DSLRs record their best "HD" video to SD/CF memory cards at relatively low data rates (their _live_ video HDMI video output is problematic for quality external recording*), a tiny fraction of the data rate of uncompressed SDI video (let alone HD-SDI). DSLRs record "HD" video to SD/CF cards using the 4:2:0 8-bit MPEG-4 AVC h.264 codec, with no B-frames, @ <50 megabits/sec.

 

What you are describing is the end point. The camera is working with higher data rates in the luminance/chroma sampling and the A/D image processing before it is compressed and laid to flash card.

 

 

In the article and related threads Barry says he used quality DSLR lenses in his tests. The lenses were also used to shoot the res chart as a digital still (in addition to the video), and the example still image in the article displays at least 2,000 lines. Typical DSLR lenses don't seem to be the limiting factor here -- instead these cam's poor "HD" video resolution performance primarily results from how they read their image sensor (binning) and how they compress frames to create motion video.

 

That's my point quality DSLR lens is vague. There are a lot of good quality DSLR lens, and some are of higher quality than others. Since the lens is where the image is started it does play a limiting factor is the image you end up with. Yes you are right its not the only limiting factor.

 

There is some experimentation with creating PL to EOS mounts, where it will be possible to mount cinema 35mm lens on DSLR. A Master Prime, Primo, or Cooke S4 may make a significant difference in the amount of detail resolution from the 7D. We really don't know until it is tested.

 

Again, I believe these cams can be of terrific value for certain productions, and it's helpful to understand their weaknesses so as to get the "best bang for your buck" out of them. If (very) carefully handled, in some cases the video they create can be intercut with HD from many other cams, including the EX3.

 

I agree its a $2000 camera. You're only going to get so much for that much money. The 5D or 7D are not going to compete directly with full HD cams.

 

My overall point is that the test from this article can only account for the variables and methodology of that particular test. This one test does not cover all circumstances and possibilities of using these cameras. There are several other tests that can be run that may yield different results.

Link to comment
Share on other sites

Its true they do use work around because there is no way they can capture a full 21MP image at the $2000 price point. You don't need 21MP to have a great looking 1080P image. Most 3CCD block HD cams are only using around 2MP themselves.

 

I'm willing to cut Canon some slack. Every camera is a compromise between quality, size, and price. I don't think Canon has any intention for the 5D/7D to compete directly with full digital cinema cams.

 

 

 

Yes they are, very much so.

 

These cameras use huge shortcuts and workarounds. The stills sensors quite simply do not go fast enough to do video, so they end up leaving out enormous amounts of the chip - just by dividing the vertical resolution of the 5D's sensor (about 3800 lines) by the output lines (1080) we can see that it doesn't actually have to do much more than every third or fourth line, and possibly less than that. This means the OLPF is now wrong and the debayer has scanty information to work with. The other alternative would be to clock everything much faster and run it much harder and then you get noise (see: Red).

 

There's no doubt that these things would look stunning if they took the entire chip, debayered as for stills, and gave us that as video, but they don't even nearly do that and nor are they even nearly capable of doing it. Remember that the 1DS needed two DIGIC processors to do 10fps, and that's just the processing load, never mind downloading it off the sensor.

 

The other problem with the 5D is that the default picture setup is awful, just incompetent beyond belief; select whichever the softest picture preset is and wind the contrast all the way down (under C.Fn II: Image, I think but I'm not sure) and you get an extra real stop of latitude at the expense of some odd but easily fixed colour banding around highlights.

 

But to answer an earlier query: the 5616 by 3744 sensor of a 5D2, sampled in let's say 16 bit and stored as the raw bayer matrix: forty megs a frame, or just under a gigabyte a second at 24fps. But it'd look very nice.

 

P

Link to comment
Share on other sites

  • Premium Member
Have you used one of them? Or, alternatively, have you looked at some footage on a proper HD monitor (NOT as clips on the Internet as seen on a computer monitor)?

I'm pretty sure Tom's a 5D owner, as he's been doing a lot of his timelapse work with it (but in RAW stills format, not video).

Link to comment
Share on other sites

But to answer an earlier query: the 5616 by 3744 sensor of a 5D2, sampled in let's say 16 bit and stored as the raw bayer matrix: forty megs a frame, or just under a gigabyte a second at 24fps. But it'd look very nice.

 

But doing that is going to cost a lot more than the the price even most professional stills photographers are willing to pay for their cameras. I suspect most of the non newsgathering market demand for the HD video DSLR cameras will be satisfied with something that gives better quality than a HDV camera with a 35mm adapter fitted.

 

I must say to date, the miost impressive stuff from these DSLRs has been low light, with shallow DOF, with everything else I can't say I'm that impressed.

Link to comment
Share on other sites

Are they less 1080P cameras than previous HD cams that shot 1440P with 2/3 inch sensors?

 

Point them at a chart and find out. The only thing I know for sure is that these Canon DSLRs simply are not 1080p.

Link to comment
Share on other sites

If I understand you correctly you are saying these camera's cannot record 1920x1080 even though they do fill the 1920x1080 pixel grid.

 

My point with the 1440 was, with so many cameras that claim to shoot 1080P on so many HD formats. What do you feel draws the line of delineation between true 1080P and false 1080P?

 

Point them at a chart and find out. The only thing I know for sure is that these Canon DSLRs simply are not 1080p.
Link to comment
Share on other sites

If I understand you correctly you are saying these camera's cannot shoot 1920x1080 even though they do fill the 1920x1080 pixel grid.

 

My point with the 1440 was, with so many cameras that claim to shoot 1080P on so many HD formats. What do you feel draws the line of delineation between true 1080P and false 1080P?

 

Well, I can upres a DVD and fill up 1080p; that doesn't mean it's 1080p.

 

If it holds up pixel for pixel and looks pristine at 1080p, then it's 1080p. It also has to pass muster on the charts. That is my definition anyway.

 

The real test for me is whether I can seamlessly intercut "1080p" video with with my 1080p timelapse renders, which are being downsampled from 5.6K RAW still images.

Link to comment
Share on other sites

  • Premium Member

Just an FYI if you haven't already seen this.

 

These are direct links to Quicktime videos of Tim Smith of Canon making a presentation at the Digital Cinema Society in CA on 7/23/09, discussing "Digital Still Cameras for Motion Picture Use":

http://www.digitalcinemasociety.org/media/..._Tim_Smith1.mov

http://www.digitalcinemasociety.org/media/..._Tim_Smith2.mov

 

These links are somewhat buried on the DCS site, but I got them from here:

http://www.digitalcinemasociety.org/conten...ting%20Workshop

 

Among other things, in his presentation Tim states that Canon DSLRs construct video frames by binning pixels.

 

I believe Barry Green accurately explains the disadvantage of using pixel binning here:

http://www.dvxuser.com/V6/showpost.php?p=1...mp;postcount=23

 

Again, just an FYI for those who hadn't seen this info previously.

Link to comment
Share on other sites

The real test for me is whether I can seamlessly intercut "1080p" video with with my 1080p timelapse renders, which are being downsampled from 5.6K RAW still images.

 

Given the level of the technology, the methods being used to generate moving video on these devices, and, yes, the price, I think it would be rather shocking if this were the case. The real issue isn't whether a $1700 SLR can compete with a $20,000 video camera, it's whether it can generate moving images that are useful and in many ways more than acceptable under many conditions. Those conditions could include things like a need for interchangeable lenses, selective focus, and performance at extremely low light levels. The Canon cameras fill all of these needs. That doesn't necessarily make them suitable to be an "A" camera on a feature length drama, but it does make them uniquely suited for capturing shockingly nice moving images in many cases. There may be aliasing issues (of course there will be, all of the low pass filtering is tuned for the full size still mode), and charts may look really lousy, but real images are surprisingly robust, especially for many situations in which a typical video camera - especially 2/3" or less - would fall apart. Not to mention their ability to bring to the masses - the real masses, not just the ones who can afford a Red One - a device that can allow them to compose images they just couldn't achieve with a camcorder in the same or even remotely similar price range.

 

Everything has its place. There are lots of tools for lots of different things.

Link to comment
Share on other sites

Given the level of the technology, the methods being used to generate moving video on these devices, and, yes, the price, I think it would be rather shocking if this were the case. The real issue isn't whether a $1700 SLR can compete with a $20,000 video camera, it's whether it can generate moving images that are useful and in many ways more than acceptable under many conditions. Those conditions could include things like a need for interchangeable lenses, selective focus, and performance at extremely low light levels. The Canon cameras fill all of these needs. That doesn't necessarily make them suitable to be an "A" camera on a feature length drama, but it does make them uniquely suited for capturing shockingly nice moving images in many cases. There may be aliasing issues (of course there will be, all of the low pass filtering is tuned for the full size still mode), and charts may look really lousy, but real images are surprisingly robust, especially for many situations in which a typical video camera - especially 2/3" or less - would fall apart. Not to mention their ability to bring to the masses - the real masses, not just the ones who can afford a Red One - a device that can allow them to compose images they just couldn't achieve with a camcorder in the same or even remotely similar price range.

 

Everything has its place. There are lots of tools for lots of different things.

 

Well said. I, like all here, still wish it could look better (less aliasing, more dynamic range, less artifacts, etc.), but considering the money these cameras cost, it seems unlikely at this point that they would. The future will tell how they resolve these issues, and I am sure they will. For right now, gotta take the good with the bad . . . ;)

Link to comment
Share on other sites

  • Premium Member

Hi, enthusiasts of the moving image,

here´s some stuff i have experienced from working with the 5D2 and 7D...

First

As the camera skips lines awful moire patterns, like i.e. cloth, appear from cranking up the sharpness in the menue. Just keep it at its lowest level. Resharpen if necessary in post. Also better for H264 compression.

Second

Display on standard monitors is RGB, right? When your *.mov shows compression artefacts like blocking noise, it may be more visible because for display RGB clips the super white and super black in each channel. I have seen huge improvements when switching to sRGB or other color profiles. As you need to convert your footage for editing just don't use QuickTime Pro.

Use Streamclip or some application that can properly handle YUV (YCbCr/YPbPr) video material. Maintain for the entire intermediate process.

http://lists.apple.com/archives/quicktime-...p/msg00142.html

I deliver mostly in SD QuickTime PhotoJpeg for our Avid machines. I have seen improvements when switching AfterEffects to sRGB and 16/32 bit color depth (per channel). All 8 bit had stairs in gradients and the upsampled composites had not - even if the rendered movs were 8 bits (per channel).

Third

Don't use 720p mode when recording. The line skipping (reads every fourth line then, or so) causes the aliasing problems. If you need 720p as the final size, use appropriate software that does proper downscaling with interpolation. The 7D offers slomo and that's the only 720p i'd use.

Fourth

The statement "The BBC or any other broadcaster does not allow 5D2 material" sounds just like the following:

"You can't shoot a feature film in S16" or "MP3 music is bad quality" or or or. It is just sooo hard to believe...

Link to comment
Share on other sites

Given the level of the technology, the methods being used to generate moving video on these devices, and, yes, the price, I think it would be rather shocking if this were the case. The real issue isn't whether a $1700 SLR can compete with a $20,000 video camera, it's whether it can generate moving images that are useful and in many ways more than acceptable under many conditions. Those conditions could include things like a need for interchangeable lenses, selective focus, and performance at extremely low light levels. The Canon cameras fill all of these needs. That doesn't necessarily make them suitable to be an "A" camera on a feature length drama, but it does make them uniquely suited for capturing shockingly nice moving images in many cases. There may be aliasing issues (of course there will be, all of the low pass filtering is tuned for the full size still mode), and charts may look really lousy, but real images are surprisingly robust, especially for many situations in which a typical video camera - especially 2/3" or less - would fall apart. Not to mention their ability to bring to the masses - the real masses, not just the ones who can afford a Red One - a device that can allow them to compose images they just couldn't achieve with a camcorder in the same or even remotely similar price range.

 

Everything has its place. There are lots of tools for lots of different things.

 

Michael Most, I'm just curious if you bother to even read my posts with an open mind, or if your entire goal in interacting with me is to quickly make a contrarian point? As far as I can recall, both here and at Reduser, you have never, ever agreed with me on anything. From my point of view, it grows tiresome.

 

I said that the camera is not 1080p, and I stand by that. Shoot some charts and prove me wrong. I'm not sure why your other points are aimed at me.

 

Anyhow, here is a "shockingly nice" and "surprisingly robust" image I was able to acquire on my 5D2 in video mode in the early days when the camera came out. I only had to drive 300 miles and walk through half a mile of slick, ankle deep mud for the pleasure of having these shots ruined by this camera.

 

1zx744.jpg

Link to comment
Share on other sites

  • Premium Member
Anyhow, here is a "shockingly nice" and "surprisingly robust" image I was able to acquire on my 5D2 in video mode in the early days when the camera came out. I only had to drive 300 miles and walk through half a mile of slick, ankle deep mud for the pleasure of having these shots ruined by this camera.

 

I agree it is a very pretty photograph...but...if the color fringing apparent in the waves in the HQ version of the shot posted here is also in the original frame capture then your very pretty picture is also a very dramatic demonstration of 5DII moire.

Link to comment
Share on other sites

I agree it is a very pretty photograph...but...if the color fringing apparent in the waves in the HQ version of the shot posted here is also in the original frame capture then your very pretty picture is also a very dramatic demonstration of 5DII moire.

 

Hal, that was EXACTLY my point. To me, that shot I posted is useless because of the ridiculous chroma aliasing. You might have missed my sarcasm.

 

This was not Canon's finest moment.

Link to comment
Share on other sites

I agree entirely. I'm surprised I'm defending this camera, I'm usually the one saying don't believe the hype.

 

What I'm objecting most too is the notion that this particular type of camera is so inferior to the others. When its problems are to one degree or another shared by all the others.

 

HDSLR may not be the best choice for shooting on a bright sunny Southern California beach. It may not be the best choice for shoot large vista of a mountain range. But if you are doing a small intimate show where you control most of the set design and lighting, its perfectly adequate.

 

Given the level of the technology, the methods being used to generate moving video on these devices, and, yes, the price, I think it would be rather shocking if this were the case. The real issue isn't whether a $1700 SLR can compete with a $20,000 video camera, it's whether it can generate moving images that are useful and in many ways more than acceptable under many conditions. Those conditions could include things like a need for interchangeable lenses, selective focus, and performance at extremely low light levels. The Canon cameras fill all of these needs. That doesn't necessarily make them suitable to be an "A" camera on a feature length drama, but it does make them uniquely suited for capturing shockingly nice moving images in many cases. There may be aliasing issues (of course there will be, all of the low pass filtering is tuned for the full size still mode), and charts may look really lousy, but real images are surprisingly robust, especially for many situations in which a typical video camera - especially 2/3" or less - would fall apart. Not to mention their ability to bring to the masses - the real masses, not just the ones who can afford a Red One - a device that can allow them to compose images they just couldn't achieve with a camcorder in the same or even remotely similar price range.

 

Everything has its place. There are lots of tools for lots of different things.

Link to comment
Share on other sites

  • Premium Member

I think Mike is a fashion follower.

 

The 5D is a Hot Toy, and he works for people who see no further than that, so he's more or less required to like it. It's genuinely amazing how some people do this, managing to turn commercial imperative into what appears to be a strongly held and genuine view. I read one interview with the guy who shot one of the Star Trek spinoffs (I think) decrying how awful video was soon after a decision had been made not to shoot the next season on video. The next but one season went digital, and there's the same guy waxing poetic about how fantastic the new technology is.

 

I have too much regard for logic and critical thinking to be able to do this. Unfortunately it appears to be a prerequisite to success.

 

P

Link to comment
Share on other sites

Again another interesting article. In this one Barry Green is using an educated guess to construct a story of how the 5D bins frames. He doesn't fully know if its true or not.

 

But for the sake of argument lets say he right. If the 5D is creating 1080P from 750,000 pixels, this is still nothing new in practices of digital video. 1/3 inch HD cams shooting 1080P, are commonly using less than a million pixels. The Red One is creating 4K from a 12MP sensor, full 4K is a 36MP image.

 

 

Among other things, in his presentation Tim states that Canon DSLRs construct video frames by binning pixels.

 

I believe Barry Green accurately explains the disadvantage of using pixel binning here:

http://www.dvxuser.com/V6/showpost.php?p=1...mp;postcount=23

Link to comment
Share on other sites

Every camera has some type of weakness, no camera is the best at everything. That does not automatically prove that particular camera incapable of taking nice picture under any circumstance.

 

Its just like when digitally shot films shoot the most dynamically challenging scenes in 35mm.

 

Anyhow, here is a "shockingly nice" and "surprisingly robust" image I was able to acquire on my 5D2 in video mode in the early days when the camera came out. I only had to drive 300 miles and walk through half a mile of slick, ankle deep mud for the pleasure of having these shots ruined by this camera.
Link to comment
Share on other sites

Every camera has some type of weakness, no camera is the best at everything. That does not automatically prove that particular camera incapable of taking nice picture under any circumstance.

 

I understand this. Honestly, if some indie band came to me right now and said, "Let's go shoot a music video at Joshua Tree in low light" I would probably just do it on a 5D2. Keep the DOF shallow, avoid moire bait, etc.

 

I was one of the first people out there trying to shoot good cinema stuff on these cameras, so it's not like I have some bias against them. My only bias is caused by the poor results I have gotten -- ie, the framegrab I posted above.

 

The main point I made on this thread is that the 5D2, 7D, etc, simply are not 1080p cameras. Maybe 720p at best.

Link to comment
Share on other sites

  • Premium Member
Again another interesting article. In this one Barry Green is using an educated guess to construct a story of how the 5D bins frames. He doesn't fully know if its true or not. ...

Hi Tenolian: I'm sorry, but I don't know what you're trying to say here. Are you implying Barry is making stuff up, or that his educated guess is groundless, or what?

 

More importantly, if _you_ know how these DSLRs create video frames, are you at liberty to explain it to us? Or are you under NDA? Barry says he's spoken with the product managers at Canon & Panasonic. If you have first-hand information about the tech in these cams, can you share it with us?

 

... But for the sake of argument lets say he right. If the 5D is creating 1080P from 750,000 pixels, this is still nothing new in practices of digital video. 1/3 inch HD cams shooting 1080P, are commonly using less than a million pixels. The Red One is creating 4K from a 12MP sensor, full 4K is a 36MP image.

May I respectfully request we not drag RED into this discussion, please? I'm certain no good will come of it. :lol:

 

As for "1/3-inch HD cameras shooting 1080P", in Barry's posts (linked above) he shows res chart frame grabs from a popular & relatively inexpensive _1/4"_ HD video camera which records 1080p HD video -- its HD video resolution beats the pants off the "HD" video recorded by DSLRs. Sure, there are 1/4" & 1/3" HD cams which produce truly sucky HD video, but not all do.

 

Isn't also true that optical low pass filters play a big role in all this? It makes sense to me that DSLRs must include OLPF which strongly favors high amounts of detail for best results in digital still photography. But this works at cross-purposes in an HD video application -- unless some future DSLR camera constructs its video frames very differently and uses a faster sensor and faster/more capable processing circuitry, both of which currently cost more, use more power, generate more heat, and are larger than what's inside current DSLRs.

 

"Traditional" HD video cameras*, even some relatively inexpensive ones, don't have this "problem" because they have a OLPF designed for HD video, not >10MP digital stills, and they also include varying amounts of the faster/better/bigger circuitry a DSRL currently does not.**

 

However, if I've gotten any part of all this wrong or backwards, please advise.

 

- Peter

 

*What a weird concept "traditional HD video cameras" is! :-)

 

** It's unfortunate the megapixel wars in the digital still market have resulted in DSLRs which "must" have image sensors with far more than 8MP, because wouldn't approx. 8MP be about perfect for a hybrid digital still camera that also shot 1080p HD video? (1920 x 1080 x 3.5) With today's tech, wouldn't it be easy to make a fast, APS-C or S35-size CMOS sensor with minimal rolling shutter jello artifacts, and with ~8MP relatively large, light-sensitive pixels, and the circuitry to process & record the data? IMHO, 8MP is plenty good enough for most non-pro & some pro digital still photography applications. Unfortunately, it may be too late for such a "low MP" DSLR design. But maybe soon a 1080p "video camera" will be designed like this, and sell for under $2K? I mean from Canon/Panasonic/Sony/Nikon etc., not RED (Scarlet)? Or is something like this not possible/likely for under $2K for technical or profit reasons? For example, maybe it's not profitable to make a APS-C or S35-size, fast CMOS sensor with ~8 mil. _big_ pixels for use in a <$2K 1080p video & digital stills camera?

Link to comment
Share on other sites

Hi Tenolian: I'm sorry, but I don't know what you're trying to say here. Are you implying Barry is making stuff up, or that his educated guess is groundless, or what?

 

More importantly, if _you_ know how these DSLRs create video frames, are you at liberty to explain it to us? Or are you under NDA? Barry says he's spoken with the product managers at Canon & Panasonic. If you have first-hand information about the tech in these cams, can you share it with us?

 

I'm saying he's taking bits of information and using an educated guess to fill in the blanks. I'm saying he may be right, he may not be. These companies don't get very specific exactly about how their sensors work. A lot of that stuff is kept secret.

 

 

May I respectfully request we not drag RED into this discussion, please? I'm certain no good will come of it.

 

I know its a hot button topic, but its difficult to just completely ignore it exists.

 

As for "1/3-inch HD cameras shooting 1080P", in Barry's posts (linked above) he shows res chart frame grabs from a popular & relatively inexpensive _1/4"_ HD video camera which records 1080p HD video -- its HD video resolution beats the pants off the "HD" video recorded by DSLRs. Sure, there are 1/4" & 1/3" HD cams which produce truly sucky HD video, but not all do.

 

"Traditional" HD video cameras*, even some relatively inexpensive ones, don't have this "problem" because they have a OLPF designed for HD video, not >10MP digital stills, and they also include varying amounts of the faster/better/bigger circuitry a DSRL currently does not.**

 

 

The best I would say is that the results of his test are the circumstances of his particular test. From the information given I would not consider his test all encompassing covering every possibility. I would give it more time and more testing. I think HDSLR will be found to have advantages and disadvantages. You weigh that against what is needed to tell the story, the same as any camera system.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Forum Sponsors

Broadcast Solutions Inc

CINELEASE

CineLab

Metropolis Post

New Pro Video - New and Used Equipment

Gamma Ray Digital Inc

Film Gears

Visual Products

BOKEH RENTALS

Cinematography Books and Gear



×
×
  • Create New...