Hello from hot and steamy Singapore. I’m at broadcast Asia doing some presentations on the PMW-F3, FS100 and 3D. On the Sony booth they have a PMW-F3 with all the options installed including S-Log and what also appears to be the 3D link firmware (mission for tomorrow – investigate 3D link). So I decided to quickly grab a couple of test shots with both S-Log and Cinegamma 1 to see how much of a difference there was. I was only able to record to the internal SxS card, but the results were quite enlightening. This was all done very quickly, so it’s not particularly scientific or accurate, but it does prove beyond any doubt that S-Log brings a significant boost to the dynamic range compared to the cinegammas. I estimate it gives you between 1.5 and 2 more stops to play with. Given that the F3 is already tested and shown to have around 11.5 stops with the cinegammas, this means that the F3 with S-Log is in the 13 to 13.5 stop range, very impressive indeed.
The shot is a pan from the darker side of the Sony booth to the very bright camera set. Exposure was set using the bright end of the shot and setting the iris so that the hotspots in the image were just below clipping, the idea being to look at what was going on in the shadows to get a feel for the dynamic range. Click on the images below a couple of times to get to the full resolution version if you want to see that.
What’s most striking is the difference in the shadow areas in the S-Log footage. Given that the peak white exposures were roughly the same (108ire) there is a huge difference in the darker shadow areas with much more information in the S-Log. It also surprised me just how well the bright part of the S-Log clip graded. Do remember that these clips were taken from recordings on the SxS cards, so they are 8 bit, 35MB/s.
I’ve put together a sequence of these clips for Vimeo. Click Here to go to the video on vimeo. The final test that I did, which you will see in the vimeo clip is to try to grade the shadow areas of the Cinegamma to bring them up to match the S-Log. The result was nowhere near as pleasing as the S-log as lifting the blacks introduces a lot of extra un-wanted noise and it still does not have the contrast range of the S-log.
Alister
There’s been a lot of conflicting reports about S log and recording it on SxS cards, or through dual link in RGB mode. The answers I’ve received is you can only get the full S log curve in RGB mode. You might get a bit in the top and bottom on the SxS or SDI outs, but not the full curve.
I did a quick test when I had the camera and Cinedeck for our perfume shoot. I could see more gradation in the RGB mode on a white roof, and less off the SxS and SDI outs. I thought you would be a good person to answer this technical question.
When we graded the RGB, S log footage, the colourist had a lot of room to manipulate the picture.
The dynamic range will be exactly the same for S-Log recordings on to SxS cards as it will be to an external device, whether it’s 4:4:4 RGB or 4:2:2. However the SxS recordings are only going to be 8 bit 35 Mb/s, so when you grade these back to a sensible looking image you may see banding getting introduced. Given the cost of the upgrade, it really only makes sense to record the S-Log to an external 10 bit recorder, otherwise you are wasting good money.
So you are basicly saying that one will get exact the same dynamic range regardless the recording format? There’s only more color information in 10bit 4:2:2 or 4:4:4 recording, correct? The reason I’m asking this is that I was thinking the s-log with pix240 recorder would be a nice mid point when doing long recordings such as interviews and documentary but still you would get the whole dynamic range the F3 is capable of.
That’s correct. Using the Sound Devices pix240 or any other 10 bit 4:2:2 recorder will give the same dynamic range as 4:4:4. 4:4:4 will have a very small color adavantage, but I very much doubt that most people would ever know the difference.
In what case then would a production need 4444 uncompressed from a product such as the Gemini? It’s not increasing the resolution, just the color space and size of file, when would this really be necessary?
Don’t forget that the Gemini can also record 4:2:2 uncompressed and this is where I see the sweet spot being for the PMW-F3.
It may be that if you are doing something like a commercial or maybe even a feature film where every microscopic ounce of quality is required where you will use 4:4:4 but for the vast majority of productions 4:2:2 will be more than enough.