After posting this, I’m wondering if the subject is to complex to keep to one thread.
There are currently two posts going with discussions of the future of video, its pros and cons, the AI, techniques, testing, etc. AZ_2021, you started both. This one and Testing AI Video Preferences in an Arizona Park I’m wondering if it may be time to start a new post that combines the things we have been discussing. I know there will be many more to come, but since this is so much on our minds, I thought it might be simpler not to have to decide which thread to post something under.
I haven’t reached out to any moderators about this and I don’t think they would need to move new posts around.
I told Tony_b that I’m on a short trip and inside of businesses I am only shooting video. I am watching some of my recent contributions where I added both video and photos to what happens. I strongly believe that what the thumbnail of a video looks like is going to be as important as how our storefront photos crop down,
I am also working on blurring a face that moves through a video. It can be done, but looks like it will be a pain in the neck.
Jumping in here @Rednewt74 and @MasaWithMaps , Alfred your question/suggestion appears to be related to the Connect forum and as such should be submitted as an suggestion or feedback in the How-tos forum for consideration.
Hi @Rednewt74 ! Sorry for the delay as I’ve been away on hols and only got back a few days ago; I am still (!) trying to sort out my videos/photos from my trip abroad! The lack of space in the cloud is forcing me to move things to a local hard disk and … well… it’s just a pain pushing files around!
Thanks for your comments/questions/observations on my videos. I do try to make them as visually interesting as possible (hence the use of the escalator to add that extra movement axis). Reading all the various comments on this post has made me realize that, perhaps, there’s no ‘correct’ way to take a video since there are too many factors that contributes to it’s view count. For example:
user’s preference for short/long videos
user’s goal in watching the videos (to see the vibe of the place, to see what the food is like, or perhaps just simply to browse); their viewing time will vary based on how well the videos match what they are looking for.
Google’s algorithm to surface your video to the user
There will be some ‘best practices’ that we could employ e.g.:
quality of video; good lighting, not blurry, good video resolution
framing of subject; are we showing the subject of our videos within the first, say, 2 seconds to catch the attention of the user correctly? Framing to me also means that whatever we show in our video is also visually intriguing so things like action vs static shots, bland vs vibrant shots etc. Panning and dynamic movements can be attractive here.
choosing a good subject. Some of the dishes that I’ve seen are so bland looking that it’s really hard for me to ‘sell’ it to anyone! In this instance, maybe a photo will suffice.
vertical/portrait orientation. For certain scenarios, the vertical shot is not helpful, but the user interface is, normally, in this orientation so when a landscape video comes on and it doesn’t fill the screen, I’m afraid that this will affect how long (or even if) the user will view the video. Plus, the popularity of Instagram, Tik Tok etc that are in this vertical orientation really does make it hard for me to recommend recording in landscape.
Anyway, the above is not to answer your questions… just my rambling mind trying to give context to how I approach video contribution on Maps. So to, finally answer your questions:
Videos over 30 seconds long. I believe that these are anomalies where the user has accidentally found a loop hole that allows sharing of long videos. Some of these loop holes have been highlighted to the Maps team in the past and they have eventually closed them. I’m not familiar with the details though so hopefully those videos that you’re seeing are older ones. If it’s recent (like in the past year or so) then, perhaps some of those loop holes still exist.
The Concourse panning direction. Unfortunately, I didn’t notice this at the time but it would certainly be a better video if I panned from Left to Right so that I can show the signage as early in the video as possible. Because of that need to capture the users’ attention as quickly as possible, I think it would certainly be helpful to have any signages (or important details/moments) captured earlier in the video.
Suria KLCC zooming. No zooming mechanism used and it’s purely a perspective effect due to panning from nearby objects to objects further away. I have tried to do some zoom in/out in other videos but it’s very janky because I can’t get a smooth zoom. However, I believe that if you’re on iOS and using a smart gimbal, then you should be able to pan and zoom easily. I am looking for a gimbal for my Pixel 7 Pro but can’t find anything suitable. Most in the market have been designed primarily for iOS and has that tight integration allowing for better/advance controls. I think the Insta360 Flow is the ‘best’ option so far but even then, Android users don’t have access to all the features.
I don’t do any trimming unless it’s super bad. I just do lots of takes to get it right! My tips are:
Plan out your panning path; where’s a good start and end point and be familiar with them so that you know when to start/end recording
Practice your panning path; my first take almost always go wrong and I end up feeling like a contortionist! Repositioning your stance usually helps to give you max comfortable twist from start to finish. Just try out the movements before recording to make sure you’ve got it right.
Figure out which joints to give the smoothest panning; I personally find joints closest to the phone (i.e. wrists) provide the most jitter so I usually pivot from my shoulders and waist. Sometimes I use my knees but they are generally quite janky as well. It all depends on your own body I suppose.
You can also start moving first before you start recording. This just takes away that transition from stationary position. Likewise for the end of the recording.
The Cinematic Pan mode really, really, really hides a lot of the jitter so it’s really the best way to do panning shots.
I am shooting in 4K. It really kills my cloud space though as I keep all my videos/photos. I know some Local Guides that clear their photos/videos after they upload to Maps but I don’t do that. I am not sure whether you can see the differences though on Maps since the video is usually very compressed and streamed at a lower resolution (well, it looks that way to me). Even when I’m at home on WiFi, videos on Maps are still at a poorer quality; the resolution doesn’t scale with your internet’s bandwidth.
Cinematic Pan mode does indeed mute the audio. The distorted audio comment was to highlight the effect on audio should the audio data be captured as well. Apologies for the lack of clarity there. I also agree with you that for most of the videos that I’ve seen with audio, the audio is actually pretty bad. I find that the Pixel 7 Pro captures way too much background audio (it seems to amplify them quite a bit more than it is in real life too). I think it’s got a 360 cone whilst I’d prefer a more directional (front) cone so that the audio matches what the user sees.
Thank you for taking the time to share your thoughts and answering my questions. I too am returning from being away, so I understand about trying to get caught up.
It seems to me we are in agreement on most every point.
I don’t see the “rotating your phone issue” as significant.
I too shoot in 4k and cinematic pan and yes the files eat up a lot of storage.
Recently been shooting mostly vertical video inside of POI’s and landscape photos of the storefront. But occasionally the interior doesn’t seem to lend itself to a video and then I go back to photos.
I am watching the success of my videos, but so far I haven’t been able to reach any conclusions about what works consistently. As you point out , there may just be too many variables.
I have been watching how my videos vs photos (before I started doing only interior videos) have been performing and over all the videos are doing better than the photos.
I have 40 plus videos up and only 1 is doing very well. It is from May and has 64,771 views or 1506/vides per day.
I want to note that this video was made from landscape photos put together in Capcut.
@Mikeinthefalls mentions in his recent post that he has yet to see a video as the first image in a POI gallery; 2nd place is the best they get. I have observed the same thing as well.
By the way it is a landscape photo montage that I made.
I am continuing to focus on adding vertical videos (using cinematic pan) and I am finding that many of them are showing up in the #2 spot in the galleries.
I may do a few more montage videos to see how they perform in comparison.
@Rednewt74 I was updating my spreadsheet of view counts (something that I’ve been doing for a while but not very regularly recently) and came across some interesting changes.
Old photos getting a jump in view counts - this is nothing new to me as I’ve noticed the behaviour before so I always advice folks not to delete underperforming photos (unless they are pretty poor and don’t match best practices) in case they get bumped up by Google’s AI.
My top 3 videos are completely different from when I last checked (from when I was commenting on this post):
Two are old and pre-dates the introduction of Cinematic Pan mode, which I believe was made available on the Pixel 5 phones since late 2020. The more recent video was captured in standard mode so all three videos have audio to them. Anyway, I’ve no idea how they made such a big jump in views or whether I somehow missed them at my last check.
My spreadsheet doesn’t have a focus on video view counts. Basically, I only care about views above 1M but I start keeping a watch on anything with at least 750k views just so I can see if it’s a slow or fast rise to the 1M mark. This time, I was just curious and started scrolling down to see what was in the 500k region and that’s when I spotted the videos above.
I do have a mini list where I separate out food shots. I was curious to see how well food photos were performing against exterior/interior shots. Anyway, I’ve included screenshots of my spreadsheet but only showed the most recent columns that I could fit on my screen (I did have a gap between April and today… so that’s why the changes are decent with a number of new additions (yellow highlight). Items in red are no longer on Maps (they still show in my list but when I click on them, it goes to a blank page or to an error page saying that the photo has been removed). There is an exception with the HOME Manchester entry where it’s completely gone from my list.
Oh, for clarity, the blank screenshot is when I click on WOOD Manchester and the error message was when I click on Maenaam Thai Restaurant. In the latter shot I don’t really see anything that is ‘wrong’ since that’s the kind of photos that I normally take anyway (no blurring of faces).
I think it’s really good that you and I are having this discussion, because I believe we bookend the range of guides that are considering this topic of "what does the new emphasis on videos vs photos mean for us". You are interested in a photo once it hits a million views. I get excited when one hits 50,000 views. Of my top 34 photos only 6 have more than a million views, and only 2 of these are less than a year old, but one is no longer the featured photo. I have 5 others in the top 34 that are less than a year old. Why top 34 you ask? Because a video I posted on May 20 is now #33 and has over 190,000 views. It is in landscape format and is a photo montage that I made in Capcut.
Below is a table summarizing what has happened in the last 72 days. I posted 184 images and about 25% (46) of these were videos. The videos received 9% of the views I received in this time period. I should note that my star photo has only received about 300 views in this time period. Also, only 2 of the videos are landscape format, but these two have received almost 58% of my total video views. I’m not prepared to come to any conclusions yet. I intend to start shooting and uploading both landscape and portrait mode videos for the same POI. I also am going to do a few more photo montages to see how they performed.
I look forward to our continued discussion. Thank you @AZ_2021 for starting this post. And @tony_b I’d love to hear what you have to say about this, as we are in the same sized boat (even though I’m landlocked).
Some great ideas here @AZ_2021 . Taking pictures is so ingrained in me (and maybe a lot of us), that remembering to take video is so hard. But I am working on it and your tips are truly appreciated!
And to you @TerryPG & @EvaBar , see what you’ve been missing. I’m having trouble remembering to shoot photos after I do my videos. See here for some stats re my recent efforts.
I’m having trouble remembering to shoot photos after I do my videos. 
See 
