• Introducing XDA Computing: Discussion zones for Hardware, Software, and more!    Check it out!

General Initial Video results - Needs tweaking

Search This thread

Ultimoose

Senior Member
Jan 17, 2011
69
35
Note this isn't a bash the pixel 6 or google post. It's just my initial impressions with only a couple of shortish video samples.

I took the P6P out yesterday and put it in a mount next to a GP9. I went out to the trail and recorded sections of footage with runs and walk sections.

It was full sun for the most part, no clouds to speak of, at 1:00 p.m. (ish).

The results were... lets say interesting. I can't share the footage unfortunately, it was BF and gorgeous weather here so no one was working or in school and there were families all over the place and I don't post videos with minors in them. I'll have to go back out on Monday during the day when there won't be any one around.

Pixel 6 Pro settings - 4k/30 Active mode stabilization, exposure and color set to auto adjust (defaults)

GoPro 9 - 4k/30, flat color profile, white balance 5000, ISO 100/1600, Sharpness Low, Shutter speed Auto, bitrate High (100mbps), EV -0.5, Hypersmooth Boost+Horizon Lock.

Both were left to record out to HEVC format.

The GP9 settings are my default trail running settings. For me 1gb of space was used on the P6P in roughly 7:30 so to make for easy comparisons I checked the same 7.5 minutes storage burn rate on the GP9.

7.5 minutes of 4k30 on the P6P consumed 1.07gb of space.
7.5 minutes of 4K30 on the GP9 consumed 5.50gb of space.

That puts the P6P with an effective about 20mbps bitrate, at least for this one sample.

Pulling the footage into DaVinci the Pixel footage and putting them side by side, initial impressions -

P6P was obviously sharper since the AI is doing that on the fly. I'd like to see an option to turn this off as I prefer to handle it myself in post but I find it acceptable. Adding 0.44 sharpness in DaVinci to the GP9 brought the two by eye pretty close.

The color space between the two was visibly reasonably close to each other which I liked. In at least this footage there's room to color grade the P6P footage, it's not blown out or over saturated like I get with the GP's native color profile. I could probably use the same grading on both footages with only minor tweaks to merge them somewhat transparently in the same comp.

I noticed a bit of exposure and color wobble at times on the P6P footage. I think auto exposure and color needs to be turned off on the P6P if you're at all going to do any color grading or post work on the footage.

The lens flares on the P6P were noticeably worse than the GP9.

A major complaint I have right now is the P6P footage seems like it would just randomly pick something to focus on and shift the video off to the side. There are a couple of spots it was like I had the two devices on different mounts and was pointing the P6P off to the side of the trail. It was bad enough at first I thought, "did the phone mount loosen up on me and I didn't notice it?" But then it would correct itself and 'aim forward' again.

There were also what I can only assume are frame drops or weird focus choices as there are a few places that look like jump cuts were done on the P6P footage or the AI jumped around the sensor to focus on something else.

There's also signs of the jello'ing in the P6P from time to time.

Overall, without updates/tweaks or opening up some values for user control, I don't think the P6P is going to become my primary recording device on ultra runs unfortunately. Which is a shame as that's why I bought the 512.

For less motion heavy recording like walking/running on technical trail this may not be an issue.

I'll have to see if there are alternate camera apps or putting the P6P on a gimbal and turning stabilization on the phone off. If DJI would ever get the OM5 working 100% with newer androids (P4 is the last official supported Pixel) then that might make for a solid combination. Or wait for updates.

And there's also the incredibly annoying issue of "No you can't turn off the screen while recording because perverts." problem with mobile devices which also adds to the power burn problem. There needs to be some quick way to drop the screen brightness down to 0 while recording IMO.

Other points, 4K/30 video burns through the power as well, more than I like. I started around 68%, finished with 34% but to be fair while I only recorded about 12-15 minutes total footage, I took a crap ton of pictures out on the trails so I don't have a solid idea yet of exactly how bad the burn is going to be. Nor what the impact of setting the display to it's lowest possible brightness will do to help with that.

Also to be fair my Garmin live track was running for the entire 3 hours of the run and there was crappy cell service in that area. But that's the normal for what I wanted to use it for.

Once I have footage I'm okay with posting publicly I'll throw up a side by side view in case someone finds it interesting or helpful.
 

MacGuy2006

Senior Member
Sep 15, 2006
886
83
I don't usually take video, but on a couple of occasions I tried to, it was unusable. Granted, I tried to zoom in at 4x on both tries, and the results were so pixelated and overprocessed, that the footage was unwatchable on anything larger than a phone.
 
  • Like
Reactions: gie62001

86rickard

Member
Oct 29, 2018
32
16
Thats because Google stupidly don't use the 4x telephoto on video. It's a crop.
I don't usually take video, but on a couple of occasions I tried to, it was unusable. Granted, I tried to zoom in at 4x on both tries, and the results were so pixelated and overprocessed, that the footage was unwatchable on anything larger than a phone.
 

Izy

Senior Member
I don't usually take video, but on a couple of occasions I tried to, it was unusable. Granted, I tried to zoom in at 4x on both tries, and the results were so pixelated and overprocessed, that the footage was unwatchable on anything larger than a phone.

Thats because Google stupidly don't use the 4x telephoto on video. It's a crop.
That's where your wrong
It uses the telephoto but only if your using 4K30fps


@Ultimoose the P6P uses 43mbs for 4k30 and 62/63mbs for 4K60

Already tested it before and checked mediainfo for bitrates
 
  • Like
Reactions: spr33

Az Biker

Sr. Mod / Mod & RC-RT Committees / Shred'r of MTBs
Staff member
I think comparing GoPro 9 video results to a smartphone video results is setting the P6P up for failure.

The GoProv9 (I have the Hero * Black) is solely built to be an active sport recoding device, nothing else.

The P6P is a smartphone that offers the ability to capture video, which I'm pretty sure wasn't designed around mountain biking, trail running, or active outdoor sporting.

Even if the comparison was sitting at a table filled with friends using these two devices; one is specifically designed to capture video, and one has a video capturing feature.
 

Golf c

Senior Member
Jan 4, 2015
1,697
880
Google Pixel 6 Pro
I think comparing GoPro 9 video results to a smartphone video results is setting the P6P up for failure.

The GoProv9 (I have the Hero * Black) is solely built to be an active sport recoding device, nothing else.

The P6P is a smartphone that offers the ability to capture video, which I'm pretty sure wasn't designed around mountain biking, trail running, or active outdoor sporting.

Even if the comparison was sitting at a table filled with friends using these two devices; one is specifically designed to capture video, and one has a video capturing feature.
I just tried shooting a horse show on video. Experimenting with different settings. Didn't turn out that great. Lol. I told girlfriend I need pro equipment
 
  • Like
Reactions: Az Biker

Az Biker

Sr. Mod / Mod & RC-RT Committees / Shred'r of MTBs
Staff member
I just tried shooting a horse show on video. Experimenting with different settings. Didn't turn out that great. Lol. I told girlfriend I need pro equipment
I'm a pretty aggressive mountain biker and the GoPro hero 8 black is amazing at video stabilization.

Biggest issue with the GoPro imho is the inaccurate depth perception. Not sure if you meant video horse shows while on a horse or on a static mount.
 

Golf c

Senior Member
Jan 4, 2015
1,697
880
Google Pixel 6 Pro
I'm a pretty aggressive mountain biker and the GoPro hero 8 black is amazing at video stabilization.

Biggest issue with the GoPro imho is the inaccurate depth perception. Not sure if you meant video horse shows while on a horse or on a static mount.
I was sitting in chair watching. No mount. Auto focus and the horse's motion were glitchy. Still learning those video settings. I had people's heads in front of me and horses in background. The focus on people's heads were perfect. Lol. Maybe turn off auto focus and do manual?
 
  • Like
Reactions: Az Biker

Izy

Senior Member
I was sitting in chair watching. No mount. Auto focus and the horse's motion were glitchy. Still learning those video settings. I had people's heads in front of me and horses in background. The focus on people's heads were perfect. Lol. Maybe turn off auto focus and do manual?
if you look in bottom right corner and select the hand there are 4 different stabilisation options

(some affect resolution)
 
  • Like
Reactions: old_fart and Golf c

Ultimoose

Senior Member
Jan 17, 2011
69
35
Found the issue with only seeing 20mbps bit rate. With Active mode stabilization you lose the ability to shoot in 4k/30, it drops to 1080P/30 (technically 28.7 and 28.6 in two different clips so it's not quite the normal). This is a personally painful limitation for me.

So if you want active motion video you either settle for 1080P, less stabilization or use a gimbal as of right now. Except DJI doesn't fully support the P6P / Android 12. Just mostly works.

Side note, the DJI Fly app doesn't work at all on the P6 (or android 12 to be fair). DJI's current official recommendation is to find a phone that their app works on.

The joys of early adopter.
 

Ultimoose

Senior Member
Jan 17, 2011
69
35
Quick question from a noobie: why not using 4k/60fps but only 4k/30fps?
Depends on the recording device. On a GoPro 4k/60 and 4k/30 both use 100mbps (with high bitrate selected) to record the video/audio. In order to fit 60 frames per second into the same storage space as 30frames per second the GP uses higher compression which results in lower amount of data per frame. i.e. 4k/60 is lower visual quality than 4k/30. And 4k/24 would be slightly higher quality than 4k/30 but the Pixel doesn't record in 24.

I only use 4k/60 for clips I specifically intend to slow down in post personally.
 
  • Like
Reactions: Utini

Ultimoose

Senior Member
Jan 17, 2011
69
35
I captured some more footage and rendering it out now. I'm going to have to say the Pixel 6 suffers in comparison at 1080/30 with Active stabilization. To anything that records video in some respects, not just against a GoPro.

For example: There are frequent freezes where the Pixel's video records the same frame over several frames, I've counted as high as 10 frames of a static image being recorded. This results in what looks like a jump cut transition when it catches back up. This happened several times in the first few minutes of the recording and the outside temp was around 45F which should rule out an overheating issue. Notably the phone recorded the entire 32 ish minute run without shutting down.

The focal point (not focus) drifts pretty badly as well at times. There are spots where it literally looks like the pixel is aimed off the side of the trail while the GP is aimed straight forward with both on the same mount. The camera appears to be shifting which portion of the sensor it's recording from not in a good way. I'm familiar with active stabilization artifacting from this kind of movement, I've owned or own every GP except the 1 so I've seen how EIS has grown and matured over time but the P6's drift and yo yo'ing is not pretty at times.

The jello effect is also noticeable as is the exposure shift although not OMG this sucks kind of way, it's more a ugh, that's ugly kind of way.

The above may be issues with the Active mode stabilization. I wasn't expecting it to be this janky or I'd of recorded other segments with EIS set to 'light' and 'cinematic'. The 'locked' mode which I assume means no stabilization would only be of any use mounted to a stationary tripod or possibly a gimbal.

Once the render finishes, uploads and the full resolution is available I'll post a link. It'll be a few hours at best as YT takes forever to provide the 4k format for me.
 
  • Like
Reactions: Utini

Ultimoose

Senior Member
Jan 17, 2011
69
35
Side note, the Active stabilization when it's working seems solid, I'll need to see the rendered side by side but in my editor it's making a solid showing going up against the GP9's Linear+HL+Boost combo.
 
  • Like
Reactions: Utini

Top Liked Posts

  • There are no posts matching your filters.
  • 4
    Note this isn't a bash the pixel 6 or google post. It's just my initial impressions with only a couple of shortish video samples.

    I took the P6P out yesterday and put it in a mount next to a GP9. I went out to the trail and recorded sections of footage with runs and walk sections.

    It was full sun for the most part, no clouds to speak of, at 1:00 p.m. (ish).

    The results were... lets say interesting. I can't share the footage unfortunately, it was BF and gorgeous weather here so no one was working or in school and there were families all over the place and I don't post videos with minors in them. I'll have to go back out on Monday during the day when there won't be any one around.

    Pixel 6 Pro settings - 4k/30 Active mode stabilization, exposure and color set to auto adjust (defaults)

    GoPro 9 - 4k/30, flat color profile, white balance 5000, ISO 100/1600, Sharpness Low, Shutter speed Auto, bitrate High (100mbps), EV -0.5, Hypersmooth Boost+Horizon Lock.

    Both were left to record out to HEVC format.

    The GP9 settings are my default trail running settings. For me 1gb of space was used on the P6P in roughly 7:30 so to make for easy comparisons I checked the same 7.5 minutes storage burn rate on the GP9.

    7.5 minutes of 4k30 on the P6P consumed 1.07gb of space.
    7.5 minutes of 4K30 on the GP9 consumed 5.50gb of space.

    That puts the P6P with an effective about 20mbps bitrate, at least for this one sample.

    Pulling the footage into DaVinci the Pixel footage and putting them side by side, initial impressions -

    P6P was obviously sharper since the AI is doing that on the fly. I'd like to see an option to turn this off as I prefer to handle it myself in post but I find it acceptable. Adding 0.44 sharpness in DaVinci to the GP9 brought the two by eye pretty close.

    The color space between the two was visibly reasonably close to each other which I liked. In at least this footage there's room to color grade the P6P footage, it's not blown out or over saturated like I get with the GP's native color profile. I could probably use the same grading on both footages with only minor tweaks to merge them somewhat transparently in the same comp.

    I noticed a bit of exposure and color wobble at times on the P6P footage. I think auto exposure and color needs to be turned off on the P6P if you're at all going to do any color grading or post work on the footage.

    The lens flares on the P6P were noticeably worse than the GP9.

    A major complaint I have right now is the P6P footage seems like it would just randomly pick something to focus on and shift the video off to the side. There are a couple of spots it was like I had the two devices on different mounts and was pointing the P6P off to the side of the trail. It was bad enough at first I thought, "did the phone mount loosen up on me and I didn't notice it?" But then it would correct itself and 'aim forward' again.

    There were also what I can only assume are frame drops or weird focus choices as there are a few places that look like jump cuts were done on the P6P footage or the AI jumped around the sensor to focus on something else.

    There's also signs of the jello'ing in the P6P from time to time.

    Overall, without updates/tweaks or opening up some values for user control, I don't think the P6P is going to become my primary recording device on ultra runs unfortunately. Which is a shame as that's why I bought the 512.

    For less motion heavy recording like walking/running on technical trail this may not be an issue.

    I'll have to see if there are alternate camera apps or putting the P6P on a gimbal and turning stabilization on the phone off. If DJI would ever get the OM5 working 100% with newer androids (P4 is the last official supported Pixel) then that might make for a solid combination. Or wait for updates.

    And there's also the incredibly annoying issue of "No you can't turn off the screen while recording because perverts." problem with mobile devices which also adds to the power burn problem. There needs to be some quick way to drop the screen brightness down to 0 while recording IMO.

    Other points, 4K/30 video burns through the power as well, more than I like. I started around 68%, finished with 34% but to be fair while I only recorded about 12-15 minutes total footage, I took a crap ton of pictures out on the trails so I don't have a solid idea yet of exactly how bad the burn is going to be. Nor what the impact of setting the display to it's lowest possible brightness will do to help with that.

    Also to be fair my Garmin live track was running for the entire 3 hours of the run and there was crappy cell service in that area. But that's the normal for what I wanted to use it for.

    Once I have footage I'm okay with posting publicly I'll throw up a side by side view in case someone finds it interesting or helpful.
    2
    I think comparing GoPro 9 video results to a smartphone video results is setting the P6P up for failure.

    The GoProv9 (I have the Hero * Black) is solely built to be an active sport recoding device, nothing else.

    The P6P is a smartphone that offers the ability to capture video, which I'm pretty sure wasn't designed around mountain biking, trail running, or active outdoor sporting.

    Even if the comparison was sitting at a table filled with friends using these two devices; one is specifically designed to capture video, and one has a video capturing feature.
    2
    I was sitting in chair watching. No mount. Auto focus and the horse's motion were glitchy. Still learning those video settings. I had people's heads in front of me and horses in background. The focus on people's heads were perfect. Lol. Maybe turn off auto focus and do manual?
    if you look in bottom right corner and select the hand there are 4 different stabilisation options

    (some affect resolution)
    1
    I don't usually take video, but on a couple of occasions I tried to, it was unusable. Granted, I tried to zoom in at 4x on both tries, and the results were so pixelated and overprocessed, that the footage was unwatchable on anything larger than a phone.
    1
    Side note, the Active stabilization when it's working seems solid, I'll need to see the rendered side by side but in my editor it's making a solid showing going up against the GP9's Linear+HL+Boost combo.