KyleProhaska
Veteran
This is NOT an AF100 bash thread...so please keep it civil.
OK so I just got my AF100 today and took some shots. It was a choice between this camera or an XF300 with SGBlade. I took a stab at the AF100 because I figured if I was wrong and didn't like it, it would be A LOT easier to sell then another 35mm Combo, even a new one. There were several reasons why the 35mm Adapter was still appealing to me. I shot my last feature on the A1 with Letus and found the image it gave to be amazing and organic. Here comes my theory, and if I'm wrong that's fine but it seems to make sense.
My AF100 (or even my 5D) puts an image together that is pretty clean overall, and suffers from some banding in a lot of situations. This is based on my tests under lighting today. When I look at my A1 + Letus footage one thing I always knew helped was the Letus ground glass giving the overall image a wash of grain that made the image pretty organic looking and filmic (in its texture). Granted, I couldn't stop down past F4 with that.
As I got my 5D I tested it out and did noise tests where I would do complicated noise addition in After Effects that took horrendously long to render, in order to help offset some of the nasty artifacts and some of the banding. Issue being is although it helped, I'm still working with footage with banding in its source footage. You can only hide things so much.
Why am I saying this? When I was using my A1 with Letus, although 420 and HDV at a low bitrate...the overall wash of grain across the image you'd think would choke the codec to DEATH, but in fact since the grain was apart of the original image before the camera compressed it to tape, it actually handled it very well and gave me something adding grain in post could never do. This was my reason for thinking about the XF300 and SGBlade again. It seemed to me that what I needed to do was simply update the camera at hand and a newer adapter that allowed me to stop down, and I would have a similar image but cleaner and have more functionality. But I opted for the AF100.
The AF100 is like a normal video camera in a lot of ways, and a DSLRish feel to it. I found the DOF to be pleasing, the robustness of the body satisfying and comforting, lens mount options awesome, etc., but the initial footage I took was pretty shocking. I was checking exposure, had my ISO in a healthy range (400 or so) and my lens at about 2.8. It was just ducks in a pond outside my apartment, and the codec got SLAUGHTERED. Yes, I setup a good scene file as suggested by others here, and YES I was sure I checked to shoot in the highest quality format. Yes, this camera was absolutely dying on this footage, with nasty blocks all over the place. A real disappointment for my first shots in broad daylight! NOTE: I SHOT SOME CLOSEUPS OF WELL LIGHT FLOWER FIXTURE AND STUFF INDOORS AND IT LOOKED PRETTY GREAT...SO I KNOW IT CAN DO THE JOB IF GIVEN THE RIGHT CIRCUMSTANCES. My A1 with Letus would'nt have handled the light I had very well, but if properly exposed the old HDV codec would've handled this extremely well, I'm sure of it as I've shot over 100 hours on that setup at least. I think I know why...
There are other factors at play here, and I could be wrong about this, but isn't it possible if you somehow attached a 35mm Adapter to the AF100, and put the same lenses on it...isn't it possible that it would've handled it differently?...with the adapter giving it a wash of slight grain BEFORE the image hit the sensor and then the compressor? I think so...I'm certain of it. It would've helped vary each pixel it had to compress together of the water I was shooting, and because of that not group them so easily giving me less artifacts and larger blocks. It's the same reason adding noise in post can work, because it's making every pixel slightly different, helping smooth out gradients and stuff. In this case the adapter on my A1 is doing that BEFORE it compresses...not after.
Anybody else understand what it is I'm saying here? I want to get the KiPro Mini to see just how much it'll effect the banding especially of the picture, and I know the bit-rate will help significantly...but I wonder just how good the footage will turn out? Since it's an uncompressed signal to a virtually lossless codec and 422, I hope for some really good results compared to what I saw today and compared to what I see my really old HDV camera with 35mm Adapter pulling off.
Please anyone with an AF100 I'm NOT bashing the camera at all so don't take this that way. You're right I'm not seasoned on it and I could've messed something up, but try to comment regarding the idea I had, and not "you're such an idiot" or "you haven't used the camera enough" type of posts.
The ultimate question is: Is adding a slight grain pattern (spinning or vibrating) over the image BEFORE it hits the compression (whether it's 420 or 422) create less or no banding and better comopression? According to my tests and the footage from my old film, I would say yes.
OK so I just got my AF100 today and took some shots. It was a choice between this camera or an XF300 with SGBlade. I took a stab at the AF100 because I figured if I was wrong and didn't like it, it would be A LOT easier to sell then another 35mm Combo, even a new one. There were several reasons why the 35mm Adapter was still appealing to me. I shot my last feature on the A1 with Letus and found the image it gave to be amazing and organic. Here comes my theory, and if I'm wrong that's fine but it seems to make sense.
My AF100 (or even my 5D) puts an image together that is pretty clean overall, and suffers from some banding in a lot of situations. This is based on my tests under lighting today. When I look at my A1 + Letus footage one thing I always knew helped was the Letus ground glass giving the overall image a wash of grain that made the image pretty organic looking and filmic (in its texture). Granted, I couldn't stop down past F4 with that.
As I got my 5D I tested it out and did noise tests where I would do complicated noise addition in After Effects that took horrendously long to render, in order to help offset some of the nasty artifacts and some of the banding. Issue being is although it helped, I'm still working with footage with banding in its source footage. You can only hide things so much.
Why am I saying this? When I was using my A1 with Letus, although 420 and HDV at a low bitrate...the overall wash of grain across the image you'd think would choke the codec to DEATH, but in fact since the grain was apart of the original image before the camera compressed it to tape, it actually handled it very well and gave me something adding grain in post could never do. This was my reason for thinking about the XF300 and SGBlade again. It seemed to me that what I needed to do was simply update the camera at hand and a newer adapter that allowed me to stop down, and I would have a similar image but cleaner and have more functionality. But I opted for the AF100.
The AF100 is like a normal video camera in a lot of ways, and a DSLRish feel to it. I found the DOF to be pleasing, the robustness of the body satisfying and comforting, lens mount options awesome, etc., but the initial footage I took was pretty shocking. I was checking exposure, had my ISO in a healthy range (400 or so) and my lens at about 2.8. It was just ducks in a pond outside my apartment, and the codec got SLAUGHTERED. Yes, I setup a good scene file as suggested by others here, and YES I was sure I checked to shoot in the highest quality format. Yes, this camera was absolutely dying on this footage, with nasty blocks all over the place. A real disappointment for my first shots in broad daylight! NOTE: I SHOT SOME CLOSEUPS OF WELL LIGHT FLOWER FIXTURE AND STUFF INDOORS AND IT LOOKED PRETTY GREAT...SO I KNOW IT CAN DO THE JOB IF GIVEN THE RIGHT CIRCUMSTANCES. My A1 with Letus would'nt have handled the light I had very well, but if properly exposed the old HDV codec would've handled this extremely well, I'm sure of it as I've shot over 100 hours on that setup at least. I think I know why...
There are other factors at play here, and I could be wrong about this, but isn't it possible if you somehow attached a 35mm Adapter to the AF100, and put the same lenses on it...isn't it possible that it would've handled it differently?...with the adapter giving it a wash of slight grain BEFORE the image hit the sensor and then the compressor? I think so...I'm certain of it. It would've helped vary each pixel it had to compress together of the water I was shooting, and because of that not group them so easily giving me less artifacts and larger blocks. It's the same reason adding noise in post can work, because it's making every pixel slightly different, helping smooth out gradients and stuff. In this case the adapter on my A1 is doing that BEFORE it compresses...not after.
Anybody else understand what it is I'm saying here? I want to get the KiPro Mini to see just how much it'll effect the banding especially of the picture, and I know the bit-rate will help significantly...but I wonder just how good the footage will turn out? Since it's an uncompressed signal to a virtually lossless codec and 422, I hope for some really good results compared to what I saw today and compared to what I see my really old HDV camera with 35mm Adapter pulling off.
Please anyone with an AF100 I'm NOT bashing the camera at all so don't take this that way. You're right I'm not seasoned on it and I could've messed something up, but try to comment regarding the idea I had, and not "you're such an idiot" or "you haven't used the camera enough" type of posts.
The ultimate question is: Is adding a slight grain pattern (spinning or vibrating) over the image BEFORE it hits the compression (whether it's 420 or 422) create less or no banding and better comopression? According to my tests and the footage from my old film, I would say yes.
Last edited: