Google has finally been able to approach Openai with the numerous updates of its generative artificial intelligence models and certain functions of Gemini, as happened days ago with two exclusive ones that expand Gemini Live’s experience. Now he can overcome himself again and the video analysis function has been seen exclusively when the user goes one to Gemini.
Openai updated chatgpt at the end of last year with the possibility of upload videos and thus analyze them When taking certain photograms of it to give a summary or describe what the user wants. It is a very different experience from the one that can be given when you upload an image, since Chatgpt will process the images with your computer vision and can take minutes to understand what is in a video.
Google has been developing this experience since March we can witness how it works thanks to Android Authority that has ended it in version 16.15.38.sa.arm64 beta of Gemini for Android. And from what it seems, it works quite well when making the consultation of the uploaded video.
The Gemini Advanced 2.0 Flash model is used and in a video you can see how the consultation is made as if it were an image. In fact, the Try with three different clips so that in the first Gemini is able to make a correct description of what he sees. It is even able to give the correct location of the place by identifying the architecture of the building that appears in the video.
The other two video clips were analyzed without failures. There are certain limits in this experience of video analysis with Gemini. One of them is the maximum duration of the video: 5 minutes. And the second is that only one consultation on the uploaded video is allowed. That is, it seems that to make another consultation, it would be necessary to upload it again.
Uploading a video to Gemini
The free android
Already last month several lines appeared in the code that indicated that Google would allow to go up and Analyze several videos both. Its description ability is quite high as you can see in shared catches to give all the details of what appears in the video.
About His possible deployment would not be very farsince normally when these functions can be activated in an unofficial way it means that the date would be close. The Google I/or will be in the middle of next month, so it would be a special date to announce this Gemini capacity and so we are able to upload videos to quickly analyze its content.
The free android
The possibilities would be expanded, and when downloading the video of an online meeting or simply to the Capture a video of what appears on the screen of the mobile so that it helps us with a certain adjustment that we have to modify. An important advance for Gemini and we will have to see how Google offers it to follow the steps to ChatgPT that has this function since December last year.
