Originally Posted by Bob Pariseau
If you don't mind doing an additional experiment, could you check some of your key scenes using the HDMI 2 output of the 103? The HDMI 2 output does not use the Marvell video processor, but it has the same calibration levels as the HDMI 1 output (I.e., when using the default (0) Picture Adjustment values, it too matches the pixel values coming off the disc). If you still see some or all of the subtle differences (such as the brighter highlights) that would eliminate the Marvell as the cause for those.
Hi Bob (et al),
Just spent some time in a dark room, plugging each player into the same input, and watching the same scenes (and freezing frames) on the 983, and the 103 through the QDEO on HDMI 1 and the Mediatek on HDMI 2. I made sure all settings were matched between the machines. I tried as best as I could to let go of my first impressions and just see what I saw.
Everything I'm about to report is subjective and unscientific. I didn't have meters measuring brightness, I wasn't running a test disc, just carefully watching scenes from a couple of films I directed and of which I oversaw the video-mastering.
And to really do this right I could have spent many more hours, watching, re-watching, going over more material, having someone else do the switching so I couldn't be influenced by pre-conceived notions, etc.
What I saw;
Each chip had slight differences. It's own 'flavor' if you will. Different strengths and weaknesses.
That said, the differences seemed smaller and less dramatic this time. Probably because they were all going through the same input (?). Or maybe my eyes were just getting more used to seeing the scenes over and over?
The QDEO (HDMI 1 on the 103) still had the most contrasty image. Some minor image detail seemed to be lost in dark areas because they'd fall off more completely to black. Also, the color seemed the warmest, with skin tones edging to red a tiny bit more. It still felt like the bright highlights were brighter, but that could be an illusion caused by the extra contrast; if the darker areas are more dark, the brights might just seem brighter. It also seemed to edge towards artificially sharp in certain shots. None the less, I liked the image more this go-round, it just looked a tiny bit less 'cinematic' and a little bit less like the film as I know it as compared to the 983. Please understand, I am talking VERY subtle differences. If I didn't know the films so wel I probably wouldn't have noticed.
The Mediatek (HDMI 2 on the 103) actually really took me off guard with how good it was. It was the least contrasty of the 3, and to my eye produced a very slightly more smooth 'film-like' image than the QDEO. It was in many ways closer to the 983 in 'feel'. That said, it had a little less detail overall and the color rendering didn't seem quite as dead on to my eye as the 983, but it was very close. In some ways I found it more 'cinematic' and was less aware of the chip 'doing something' to the image than the QDEO. Again, we're talking small, subtle differences.
The 983 was still my favorite balance of all the elements (contrast, color, sharpness vs. natural film quality). But it was closer this time. And there were definitely individual shots that were better on the QDEO or the Mediatek. But that was sometimes because their innate tendencies would 'fix' things I wasn't crazy about in the original transfer (e.g. there was a shot I've always felt was just a tiny bit too contrasty in the final master, at least as I've gotten used to seeing it on the 983. That shot actually looked better on the Mediatek. Similarly, the slightly more contrasty QDEO gave a taste of dramatic depth to scenes that were more flatly shot that was certainly appealing).
What does it all mean? I don't know. Just reporting what my very subjective eyes are saying. If the 983 was an "A" to my eye, the QDEO and Mediatek were both "A-".