Can't wait to see comparisons of our pimped up F3's to the Canon C300, especially after EI S-Log enabled. Gonna to be a heated battle and indeed interesting times for all.

Can't wait to see comparisons of our pimped up F3's to the Canon C300, especially after EI S-Log enabled. Gonna to be a heated battle and indeed interesting times for all.

In terms of real world shooting, the extra resolution of the C300 *might* be noticeable (its only ~150 more lines of res); however, the C300 cannot come close to the current 13.5 stops of DR of the F3 + S-Log and 14 stops of DR with EI S-Log. The C300 is limited by its 8 bit codec so squeezing more than 11 stops of DR would cause problems in post. Alister has written extensively about the problem of combining 11+ stops of DR with an 8 bit codec (thanks to Alister - I have learned more from you about signal processing than anyone else :)

Moreover, using a flat gamma, such as Canon's log, with an 8 bit codec can cause even more problems IF it lifts shadows and compresses highlights too much. I haven't seen Canon's log gamma; so, I don't know how it looks. The smart guys over at FXPHD/FXGUIDE.com have taught and discussed the Canon DSLRs extensively, and they were one of the first to show concern over the Technicolor profile because it compressed the image too much for an 8 bit codec.

However, I wonder if Canon will state that their Log gamma is to be used only for recording via SDI. Although, I vaguely recall reading that Canon is touting their log gamma as not needing anything extra like the Sony S-Log; so, I presume they designed the log gamma for internal recording.

UPDATE: I just read that the C300 is *only* 8 bit via SDI. Sorry, but $20k for an 8bit camera is absolute nonsense in 2011/12.

When are Canon and Sony going to realize that Red is the only company providing customers with all of the necessary features we actually WANT such as internal 10-12 bit recording. (although, you must spend another $10k for the SSDs) If the Scarlet can come close to the low noise of the F3 AND the claimed 13.5 stops of DR, the F3 + S-Log + external recorder is not looking so great. Most people don't need RAW or 4k capture (people, STOP SAYING 4k RESOLUTION when it refers ONLY to the # of pixels), but having an internal 10 bit recording would be hugely popular. Sony really needs to make the F3 with S-Log and SR1 built into the body. They could provide the HDCAM-SR Lite for free and unlock the 440 & 880MB/s HDCAM SR via a paid upgrade key. With a 220Mb/s, 10bit AVC codec, the HDCAM Lite is better than Pro Res HQ.

8 bit through SDI kills the C300.

We'll have to see when the Scarlet is in people's hands, but it remains to be seen how many people will accept its compromises.

As you point out 4k acquisition is ~3.6k resolution. What will most people do with that? Down rez it to 2k?

Variable frame rates, but at the expense of sensor cropping, which will affect FOV, surface area of the sensor for loss of light sensitivity, and in effect make your lenses tele's.

Whereas the F3 stays at S35. If you really need 4k I think most people will rent a Red One.

It will be interesting when there is actually use and comparisons going on.

Actually, 4k acquisition equals far less resolution, about 2400-2600 TVL due to being a Bayer sensor.

I have my hands on the latest firmware, EI mode and all. Looks like EI adjusts the LUT applied to the different outputs, but the S-Log output always remains the same. This makes sense when we think of Exposure Index as being just like a post push. I'll test the actual dynamic range tomorrow. Should be interesting.

Andy

I have my hands on the latest firmware, EI mode and all. Looks like EI adjusts the LUT applied to the different outputs, but the S-Log output always remains the same.

Uhg. The Red approach. I was hoping this is what EI mode *wasn't*.

I've been shooting S-log to SxS with great success, at 0, 6, 12, and 18. This is going to compromise that now. Also it's going to mean you need multiple S-Log LUTs in color correction, one for each ASA. But with no metadata to tell the colorist which he should use? The Red way works because the ISO metadata travels in the R3D.

If this is all the case, I hope EI mode I can turn off and go back to the current S-Log method with 1.1?

You can switch EI mode on/off

Ned

And you can still add in camera gain to EI S-log using the gain switch. The base ISO is reduced to 400/800 as opposed to the current 800/1600, but you can still use the gain switch to increase the S-Log ISO even with EI S-Log enabled, so you have the best of both worlds. Add gain direct to the S-Log or just to the LUT output.

Whew.

:-)

Thanks for that very important info.

Alister

That's odd, the firmware I have doesn't seem to work that way. I'll do some more testing. When EI mode is on the gain switch does nothing.

Andy

OK, the only firmware I played with did allow the use of the gain switch. Perhaps it's different on the release version.

Either way you can turn EI mode on and off so you can choose either method.

I'll test the actual dynamic range tomorrow. Should be interesting.


Andy, did you get the S-Log with EI tested for actual DR?

I'm very grateful that you test on the scope so your tests are comparable to one another, unlike some other testers. Keep up the good work.

Quote from Andy "When EI mode is on the gain switch does nothing."

Yes, I can confirm that.

. . . . there is no HDMI output at all when S-LOG is turned on.

Wrong. I have to correct myself on this. You can use HDMI if you don't mind giving up SDI. In other words, when S-LOG is turned on you have to choose SDI or HDMI output, you cannot have both turned on at the same time like you normally can. This adds a wrinkle to the output configuration if you use an HDMI EVF as I do.