Video Performance

Along the same lines as still image performance, video performance is evaluated by using relative comparisons with various smartphones. However, unlike still image testing we attempt to more strongly emphasize the need for effective image stabilization, whether through digital or optical means. As we lack a standardized testing rig for various types of motion, these tests are purely relative and should only be used to compare between two different devices that are recording at the same time.

There are also a number of factors that affect image quality here, as there is both a need to meet real time latency targets and keep spatial resolution as high as possible. As a result, there is a stronger emphasis on ISP performance as digital image stabilization, sensor correction, lens correction, demosaicing, and other compute tasks must be done at rates as high as 480 FPS depending upon the device, with resolutions reaching as high as 2160p for output.

Looking at the stats for 1080p30 video, we see that the LG G4 uses 17Mbps high profile H.264 and 96Kbps AAC at 48KHz stereo audio, which is pretty much par for the course for high-end smartphones as far as I can tell. Comparing it to the Galaxy S6 shows that contrast is a bit lower, but dynamic range appears to be higher as well. Detail is also pretty much comparable, although the high contrast of the Galaxy S6 seems to help with improving edge contrast in certain scenarios.

The big differences are in audio quality and stabilization, where the G4 appears to reset often and is less able to dampen significant motion while the GS6 manages to avoid resetting the OIS as often and appears to have a greater maximum angle before the OIS reaches a travel limit. The G4 also has sharper wind noise, although the difference isn't enormous there, and in general noises seem to be a bit louder on the G4's video here. The color balance here is also slightly too cold for some reason.

In the interest of trying to add some more comparison points I also added the iPhone 6, which has noticeably more natural stabilization than either the G4 or the GS6. The tighter crop of the iPhone 6 also helps with improving detail, as I suspect all three are pretty much limited by the 1080p resolution rather than encode settings. The G4 also ends up noticeably louder relative to the iPhone 6, which is closer to the GS6 in terms of sound profile.

Unfortunately, there's no 1080p60 sample because for some reason I couldn't find the option to enable it in the G4's camera application, which feels like a pretty significant omission as any sort of movement would benefit greatly from greater temporal resolution. However, LG does have 4K30 recording, which has the same settings as 1080p30 but with a 30Mbps bit rate for video. Once again, the Galaxy S6 and G4 are incredibly close here in terms of detail, but the G4 avoids excessively high contrast and seems to expose to allow for much more shadow detail. OIS continues to reset more often, and noise tends to not be suppressed as well as it is on the Galaxy S6.

Moving on to slow motion, the G4 supports a maximum of 720p120 played back at 30 FPS, with 24Mbps baseline H.264, with identical audio to the other two video modes. Comparing this mode to the Galaxy S6, it's clear that the Galaxy S6 has much better detail and less obvious pixelation/blockiness throughout the video. As a result, it's also not as good as the iPhone 6 at slow motion video at 720p120. Neither the Galaxy S6 nor the G4 support 720p240, so the iPhone 6 definitely stands alone there as well.

Overall, video quality is acceptable, but slightly trails the Galaxy S6 due to slightly worse noise cancellation, worse slow motion video capture, and the lack of 1080p60 video mode. The slightly more obvious OIS might also be a negative, but this is somewhat subjective as it's basically up to the end user to decide whether they want to see large resets of the OIS somewhat rarely or smaller resets that happen more often. I personally would rather see simple dampening of motion rather than completely compensating for it, as in cases where it's almost impossible to completely defeat the motion with OIS video will appear unnatural and strange, to say the least. If the G4 has a downfall in camera, it's in video quality.

Still Image Performance Snapdragon 808 & System Performance
Comments Locked

84 Comments

View All Comments

  • Fidelator - Wednesday, August 5, 2015 - link

    The 6+ doesn't even come close to either the S6 or the G4, that is some bs right there
  • Mattzerminater - Friday, August 7, 2015 - link

    Dude! Aspect ratio for the camera can be changed between 16:9, 4:3 and 1:1 in the settings. Turning off the quickshot feature when double tapping the volume down button can also be turned off in the phone settings.
  • cdm283813 - Sunday, August 9, 2015 - link

    After owning a Note 2, 3 and 4 I really wanted to try something different (and smaller) for 2015. So I waited for the G4 especially after seeing the S6 but I ultimately went to the S6. That initial pick up and feel moment of the G4 had me thinking about the Note 2 which for me was a step backwards. It felt too light and cheap; things that I defended Note 2 and 3 for having. Plus 5.5" is still a phablet in my book no matter how much they strink the body. And I'm just not a fan of LG's UI look. So far the theme engine in the latest touchwiz is freaking awesome. It really helps when you just need a fresh coat of paint to liven up the phone again.
    It's only been about 4 weeks since owning the S6 and I'm still happy. If I really had to change anything it would be a slightly larger battery. But at the same time this phone charges very quick. I have 3 quick chargers (1 car and 2 wall) and so far I haven't run out of battery. So far the S6 has been my best phone to date.
    In saying that I can't wait for the 2015 Moto pure review. Like to see how much compromise when into a phone costing $200 to $300 cheaper than a big boy flagship. Ater reading this review I have a new respect for the SD 808. But with the SD 820 coming out it may be short lived
  • EagleX - Tuesday, October 13, 2015 - link

    How did you get 6.6 hours in pcmark battery work test? I got 5.51, exactly as the average appearing in Pcmark's website. Is it just me?

Log in

Don't have an account? Sign up now