Nice C300 peice by Jonathan Yi

[Simon Denny]

I have watched this many times today and like the image. I think Canon have a great camera here, the first of many to come and the small form factor would be great for me. Although, the current price I feel is set way to high and needs to come down.
My 2 cents worth!

[Don Miller]

Reading the comments on Vimeo, this video catches a bit of the buzz that Laforet did for the 5DII with Reverie. It's interesting how far we've come from Reverie looking technically good (to me at least).

If the 1DX is closer to the C300 than the 5D that could make some of us happy.

[David Knaggs]

A really nicely done review. I loved the ISO demonstration (and I thought they did a good rolling shutter comparison, moire comparison, etc.). It was good to see so many ungraded shots, especially those showing the different gamma settings. And they did some nice grading of the Canon Log footage, which showed it off to good advantage, I thought. I agree with the other comments on the price point and it will be interesting to see what the release price will actually be in January.

[Sean Seah]

Originally Posted by Sareesh Sudhakaran

It makes me wonder how the 1DX will perform in low light...

I had a chance to mess with the 1Dx 2 weeks ago, and I would say it outperforms the 5D2 in terms of low light. ISO6400 is like the old ISO3200 in 5D2. However the noise pattern is still ugly unlike the C300 which is supposed to be film like.

[Alister Chapman]

It's a shame they didn't normalise the exposure in the gamma tests. The differences in range make it hard to judge whether the Canon log brings any real advantage over the other gammas and the DSLR was considerably over exposed. The overexposed yellow sweater becomes not over exposed when they go to Canon Log, but we also see a drop in overall range with shadow detail getting lost, mid grey shifts up and down greatly between each gamma. It makes no sense to expose for one gamma and then just expose everything else the same when your exploring the different dynamic ranges of different curves. Much more telling would have been to have set mid grey at the same point for each so you can see how much under/over range you have. So the question is how would the standard gammas have looked at the same grey card exposure as the Log?

I also think the jello tests were a bit of a cheat as on the DSLR the bases of the vertical columns go completely out of frame and the shake frequency is higher while the C300 column bases stay in frame and the shaking is less violent, so two very different amounts of shake. On the driving shot the DSLR is completely overexposed and it looks to me to be a longer focal length so shake and jello would be amplified. Would have been nicer to have had both cameras on the same mount at the same time and both correctly exposed for a fair test. Don't get me wrong, I would expect the DSLR to perform significantly worse that the C300, but for the tests to be really meaningful they need to be done as fairly as possible.

[Brett Sherman]

Great looking footage - no doubt. However, if they are really thinking this should be used for documentaries why didn't they include the 3X sensor crop capabilities that the T3i has. We know the chip has the resolution for it. This helps avoid changing lenses to capture things as they are happening. Depending on what you are covering, many times it is simply impossible to take the time to change lenses. It seems like it would have been an easy feature to add. Makes me wonder if documentary film-making was truly a target audience when they developed this camera. Or if they actually talked to any documentary filmmakers.

One great thing this camera addresses is getting deep depth of field. While I love the shallow DOF for a lot of shooting, however there are times when I want deep too. Clean high ISOs allow that and aren't only for low-light situations. Sensor crop also helps this as it gives you an effectively smaller chip.

Anyone else notice how soft the image is when they show how the C300 doesn't have moire? It would be more convincing if it had been as sharp as the 5D footage. We all know one way to deal with moire is to put subject slightly out of focus and sharpen in post.

[Brian Drysdale]

Traditionally cinematographers and broadcast camera people don't crop their images, even for documentaries. There's a resolution drop and the effect is very different to changing your focal length.

There are a number of zoom lens options around, one of which is to use a 2/3" video lens. Alistair Chapman has been working on a cheaper adapter which seems to be going into production. http://www.dvinfo.net/forum/sony-xdc...w-adapter.html

Another is the parfocal 3:1 zooms which are compact and have a f2.8 stop.

[Brett Sherman]

I don't think you are understanding how the sensor crop works. An oversampled sensor like the one in the C300 can generate a cropped image without losing resolution just like the T3i does. Remember that the output is only 1080P, so if you have a 6 megapixel imager you have a lot of latitude to take only part of the image before writing as a 1080P file.. It works quite well on the T3i, with minimal, if any, loss of resolution.

Yes the imager becomes smaller when you do this. However, there are times when shooting you'd rather be quick than worry about getting the shallowest depth of field possible. Sensor crop is a powerful tool to have in our toolbelt. I find I use it A LOT with my T3i. If you don't have it or have never used it, you might not know what you're missing.

[Brian Drysdale]

Yes I know how the crop works, but the C300 doesn't operate in the same manner as the T3i. when creating the 1080p in the camera. The T3i has a larger pixel count sensor to begin with and it line skips when creating the 1080p, the C300 doesn't do that..

David Heath explains his take on the C300 sensor here: http://www.dvinfo.net/forum/new-cano...arlet-x-8.html

The RED allows windowing and it's noticeably softer at 2k compared to 3k or 4k, objectionably so for quite a few people.

[Tim Le]

Originally Posted by Brett Sherman

I don't think you are understanding how the sensor crop works. An oversampled sensor like the one in the C300 can generate a cropped image without losing resolution just like the T3i does. Remember that the output is only 1080P, so if you have a 6 megapixel imager you have a lot of latitude to take only part of the image before writing as a 1080P file.. It works quite well on the T3i, with minimal, if any, loss of resolution.

Yes the imager becomes smaller when you do this. However, there are times when shooting you'd rather be quick than worry about getting the shallowest depth of field possible. Sensor crop is a powerful tool to have in our toolbelt. I find I use it A LOT with my T3i. If you don't have it or have never used it, you might not know what you're missing.

Unfortunately, the C300 does not work this way. The sensor really only has effectively 1920 x 1080 "pixels." There just happens to be four photosites (1 red, 1 blue and two greens) for each recorded pixel. It reads these out directly and does not de-Bayer, so there is no possibility of a larger resolution from which to crop.

[Brett Sherman]

My mistake for taking the Canon published pixel count at face value - Effective Pixels
Approx. 8.29 megapixels (3840 x 2160).

[Tim Le]

Originally Posted by Brett Sherman

My mistake for taking the Canon published pixel count at face value - Effective Pixels
Approx. 8.29 megapixels (3840 x 2160).

I can understand the confusion because the C300 works definitely than most other large single sensor cameras. The 8.29 megapixels is correct, although a better term is 8.29 mega-photosites. It's just people are not used to "photosites" so they just used the term pixels.

If you add up all the photosites:

1920 x 1080 Red
1920 x 1080 Blue
1920 x 1080 Green1
1920 x 1080 Green2

You end up with 8,294,400 photosites or 8.29 "megapixels"

If Canon de-Bayers this sensor, they theoretically could record a 4K image. But the C300 is not doing this. It's just reading out the 1920 x 1080 directly. So because of the way the C300 constructs its image, there is no possibility for a higher resolution from which to crop.

[Evan Donn]

If Canon de-Bayers this sensor, they theoretically could record a 4K image.

I suspect this is exactly what is planned for the 4k DSLR prototype they've been talking about - take the same sensor and add an option to debayer to a full 4k image. The fact that they've described it as using MJPEG compression for 4k makes me think they're using the simpler compression (and likely higher data rates) to free up processing power to handle the debayering process in real time. It also makes me think it's likely to show up as a firmware upgrade for the C300 at around the same time the DLSR actually ships, as it seems unlikely at this point that they'd ship a (presumably) less expensive DSLR with higher performance capabilities than their flagship cinema camera.

[Jon Fairhurst]

I agree, Evan,

In fact, my guess is that it will be similar to the 1D X in that it will use two Digic 5 chips. It doesn't need the DV3, since it won't likely have XLRs, genlock, HD SDI, etc. However, I expect it to use the C300 sensor.

This is a nice differentiation strategy. Video cameras will have the pro features and a higher price. DSLRs can have similar video quality, but fewer features and lower prices (along with higher volumes.)

Time will tell...

[Tim Le]

Originally Posted by Evan Donn

I suspect this is exactly what is planned for the 4k DSLR prototype they've been talking about - take the same sensor and add an option to debayer to a full 4k image. The fact that they've described it as using MJPEG compression for 4k makes me think they're using the simpler compression (and likely higher data rates) to free up processing power to handle the debayering process in real time. It also makes me think it's likely to show up as a firmware upgrade for the C300 at around the same time the DLSR actually ships, as it seems unlikely at this point that they'd ship a (presumably) less expensive DSLR with higher performance capabilities than their flagship cinema camera.

This seems logical, but unfortunately the C300's sensor is only S35 sized and Canon has said the C-DSLR will have a full frame 35mm sensor. The 4K video will be cropped to APS-H sized, but the 1920x1080 HD video will use the full width of the sensor (according to Erik Allen at Canon).

I also don't think the C300 will get any sort firmware update to enable de-Bayering and record to 4K. This would require a whole new codec because the XF codec doesn't record beyond HD. Also, it's unknown if the DIGIC DV III can physically handle a 4K de-Bayer in real time.

However, it is intriguing to think that all Canon needs is a new DSP and codec and they could be recording 4K in no time flat. But then again, for true 4K resolution, you have to oversample like what the F65 does at 8K. I would think Canon would prefer that method, but it would require a new sensor.