Amazon Nova Reel 1.1: With as much as 2 minutes movies of a number of photographs

In Re: Invent 2024, us Amazon Nova introduced fashionsa brand new technology of Base fashions (FMS), together with Amazon Nova ReelA video technology mannequin that creates quick movies from elective textual content descriptions and pictures (joints, “discover”).

Right this moment we current Amazon Nova Reel 1.1which offers high quality and latency enhancements in producing a single 6 -second shot movies, in comparison with Amazon Nova Reel 1.0. This replace lets you generate multi -shooting movies as much as 2 minutes with a continuing type within the photographs. You’ll be able to present a single message of as much as a video of as much as 2 minutes composed of 6 -second photographs, or design every individually take with personalised indications. This affords new methods to create video content material by way of Amazon mom rock.

Amazon Nova Reel improves inventive productiveness, whereas serving to to cut back the time and price of video manufacturing utilizing Generative. You should use Amazon Nova Reel to create convincing movies in your advertising and marketing campaigns, product designs and social community content material with larger effectivity and inventive management. For instance, in promoting campaigns, you’ll be able to produce prime quality video commercials with constant pictures and synchronization utilizing pure language.

To start out with Amazon Nova Reel 1.1
If you’re new to make use of Amazon Nova reel fashionsGo to Roca console on Amazonselect Mannequin entry Within the navigation panel and request entry to Amazon Nova Reel mannequin. When you’ve got entry to Amazon Nova Reel, it applies to each 1.0 and 1.1.

After acquiring entry, you’ll be able to strive Amazon Nova Reel 1.1 immediately from the Roca console on AmazonAws sdk, or AWS command line interface (AWS CLI).

To strive the Amazon Nova Reel 1.1 Mannequin within the console, select Picture/video low Youngsters’s parks Within the left menu panel. Then select Nova Reel 1.1 As a mannequin and enter your warning to generate video.

Amazon Nova Reel 1.1 affords two modes:

• Automated Multishot – On this mode, Amazon Nova Reel 1.1 accepts a single request of as much as 4,000 characters and produces a video of a number of photographs that displays that indicator. This mode doesn’t settle for an entry picture.
• Multishot Handbook – For individuals who need extra direct management over the composition of a video, with handbook mode (additionally referred to as graphic script mode), you’ll be able to specify a novel message for every particular person shot. This mode accepts an elective begin picture for every shot. The photographs should have a decision of 1280 × 720. It may possibly present pictures in base format64 or from a Amazon Simo Storage Service (Amazon S3) location.

For this demonstration, I exploit the AWS SDK for Python (Boto3) To invoke the mannequin utilizing the Amazon API Bedrock and Beginning operation To start out an asynchronous invocation and generate the video. I used Getasycinvoke To confirm the progress of a video technology work.

This Python script creates a 120 -second video utilizing MULTI_SHOT_AUTOMATED mode Tasktype parameter From this textual content message, created by Nitin Eusebio.

import random
import time

import boto3

AWS_REGION = "us-east-1"
MODEL_ID = "amazon.nova-reel-v1:1"
SLEEP_SECONDS = 15  # Interval at which to verify video gen progress
S3_DESTINATION_BUCKET = "s3://"

video_prompt_automated = "Norwegian fjord with nonetheless water reflecting mountains in good symmetry. Uninhabited wilderness of Big sequoia forest with daylight filtering between huge trunks. Sahara desert sand dunes with good ripple patterns. Alpine lake with crystal clear water and mountain reflection. Historic redwood tree with detailed bark texture. Arctic ice cave with blue ice partitions and ceiling. Bioluminescent plankton on seashore shore at evening. Bolivian salt flats with good sky reflection. Bamboo forest with tall stalks in filtered mild. Cherry blossom grove towards blue sky. Lavender subject with purple rows to horizon. Autumn forest with crimson and gold leaves. Tropical coral reef with fish and colourful coral. Antelope Canyon with mild beams by way of slim passages. Banff lake with turquoise water and mountain backdrop. Joshua Tree desert at sundown with silhouetted timber. Iceland moss- coated lava subject. Amazon lily pads with good symmetry. Hawaiian volcanic panorama with lava rock. New Zealand glowworm cave with blue ceiling lights. 8K nature pictures, skilled panorama lighting, no motion transitions, good publicity for every setting, pure coloration grading"

bedrock_runtime = boto3.shopper("bedrock-runtime", region_name=AWS_REGION)
model_input = {
    "taskType": "MULTI_SHOT_AUTOMATED",
    "multiShotAutomatedParams": {"textual content": video_prompt_automated},
    "videoGenerationConfig": {
        "durationSeconds": 120,  # Have to be a a number of of 6 in vary (12, 120)
        "fps": 24,
        "dimension": "1280x720",
        "seed": random.randint(0, 2147483648),
    },
}

invocation = bedrock_runtime.start_async_invoke(
    modelId=MODEL_ID,
    modelInput=model_input,
    outputDataConfig={"s3OutputDataConfig": {"s3Uri": S3_DESTINATION_BUCKET}},
)

invocation_arn = invocation("invocationArn")
job_id = invocation_arn.cut up("/")(-1)
s3_location = f"{S3_DESTINATION_BUCKET}/{job_id}"
print(f"nMonitoring job folder: {s3_location}")

whereas True:
    response = bedrock_runtime.get_async_invoke(invocationArn=invocation_arn)
    standing = response("standing")
    print(f"Standing: {standing}")
    if standing != "InProgress":
        break
    time.sleep(SLEEP_SECONDS)

if standing == "Accomplished":
    print(f"nVideo is prepared at {s3_location}/output.mp4")
else:
    print(f"nVideo technology standing: {standing}")

After the primary invocation, the script periodically verifies the state till the creation of the video has been accomplished. I cross a random seed to acquire a distinct consequence each time the code is executed.

I execute the script:

Standing: InProgress
. . .
Standing: Accomplished
Video is prepared at s3:////output.mp4

After a couple of minutes, the script is accomplished and prints the Amazon S3 output location. I discard the output video utilizing the AWS CLI:

aws s3 cp s3:////output.mp4 output_automated.mp4

That is the video that generated this discover:

Within the case of MULTI_SHOT_MANUAL Manner as a tasktype parameter, with a a number of photographs request and an outline for every shot, it’s not mandatory so as to add the variable durationSeconds.

Utilizing the indicator for a number of photographs, created by Sanju Sunny.

I run the Python script:

import random
import time

import boto3


def image_to_base64(image_path: str):
    """
    Helper perform which converts a picture file to a base64 encoded string.
    """
    import base64

    with open(image_path, "rb") as image_file:
        encoded_string = base64.b64encode(image_file.learn())
        return encoded_string.decode("utf-8")


AWS_REGION = "us-east-1"
MODEL_ID = "amazon.nova-reel-v1:1"
SLEEP_SECONDS = 15  # Interval at which to verify video gen progress
S3_DESTINATION_BUCKET = "s3://"

video_shot_prompts = (
    # Instance of utilizing an S3 picture in a shot.
    {
        "textual content": "Epic aerial rise revealing the panorama, dramatic documentary type with darkish atmospheric temper",
        "picture": {
            "format": "png",
            "supply": {
                "s3Location": {"uri": "s3:///pictures/arctic_1.png"}
            },
        },
    },
    # Instance of utilizing a domestically saved picture in a shot
    {
        "textual content": "Sweeping drone shot throughout floor, cracks forming in ice, morning daylight casting lengthy shadows, documentary type",
        "picture": {
            "format": "png",
            "supply": {"bytes": image_to_base64("arctic_2.png")},
        },
    },
    {
        "textual content": "Epic aerial shot slowly hovering ahead over the glacier's floor, revealing huge ice formations, cinematic drone perspective",
        "picture": {
            "format": "png",
            "supply": {"bytes": image_to_base64("arctic_3.png")},
        },
    },
    {
        "textual content": "Aerial shot slowly descending from excessive above, revealing the lone penguin's journey by way of the stark ice panorama, artic smoke washes over the land, nature documentary styled",
        "picture": {
            "format": "png",
            "supply": {"bytes": image_to_base64("arctic_4.png")},
        },
    },
    {
        "textual content": "Colossal vast shot of half the glacier face catastrophically collapsing, monumental wall of ice breaking away and crashing into the ocean. Gradual movement, digital camera dramatically pulling again to disclose the large scale. Monumental waves erupting from impression.",
        "picture": {
            "format": "png",
            "supply": {"bytes": image_to_base64("arctic_5.png")},
        },
    },
    {
        "textual content": "Gradual movement monitoring shot shifting parallel to the penguin, with snow and mist swirling dramatically within the foreground and background",
        "picture": {
            "format": "png",
            "supply": {"bytes": image_to_base64("arctic_6.png")},
        },
    },
    {
        "textual content": "Excessive-altitude drone descent over pristine glacier, capturing violent fracture chasing the digital camera, crystalline patterns shattering in gradual movement throughout mirror-like ice, digital camera easily aligning with floor.",
        "picture": {
            "format": "png",
            "supply": {"bytes": image_to_base64("arctic_7.png")},
        },
    },
    {
        "textual content": "Epic aerial drone shot slowly pulling again and rising increased, revealing the huge infinite ocean surrounding the solitary penguin on the ice float, cinematic reveal",
        "picture": {
            "format": "png",
            "supply": {"bytes": image_to_base64("arctic_8.png")},
        },
    },
)

bedrock_runtime = boto3.shopper("bedrock-runtime", region_name=AWS_REGION)
model_input = {
    "taskType": "MULTI_SHOT_MANUAL",
    "multiShotManualParams": {"photographs": video_shot_prompts},
    "videoGenerationConfig": {
        "fps": 24,
        "dimension": "1280x720",
        "seed": random.randint(0, 2147483648),
    },
}

invocation = bedrock_runtime.start_async_invoke(
    modelId=MODEL_ID,
    modelInput=model_input,
    outputDataConfig={"s3OutputDataConfig": {"s3Uri": S3_DESTINATION_BUCKET}},
)

invocation_arn = invocation("invocationArn")
job_id = invocation_arn.cut up("/")(-1)
s3_location = f"{S3_DESTINATION_BUCKET}/{job_id}"
print(f"nMonitoring job folder: {s3_location}")

whereas True:
    response = bedrock_runtime.get_async_invoke(invocationArn=invocation_arn)
    standing = response("standing")
    print(f"Standing: {standing}")
    if standing != "InProgress":
        break
    time.sleep(SLEEP_SECONDS)

if standing == "Accomplished":
    print(f"nVideo is prepared at {s3_location}/output.mp4")
else:
    print(f"nVideo technology standing: {standing}")

immediate = "A easy crimson triangle transforms by way of geometric shapes in a journey of self-discovery. Clear vector graphics towards white background. The triangle slides throughout unfavourable house, morphing easily right into a circle. Pan left because it encounters a blue sq., they carry out a geometrical dance of shapes. Monitoring shot as shapes mix and separate in mathematical precision. Zoom out to disclose a sample shaped by their actions. Restricted coloration palette of main colours. Exact, mechanical actions with good geometric alignments. Transitions use easy wipes and geometric form reveals. Flat design aesthetic with sharp edges and strong colours. Remaining scene exhibits all shapes combining into a posh mandala sample."