Turn Your Photos Into a Music Video With AI Lip Sync Tech

In this post, I’ll share how I created this music video from just a few static photos.

Assets

To make a music video showing someone singing, you’ll need the following:

  • One or more photos of the people who will be singing (usually JPG format)
  • A song (usually MP3 format)
  • Background video footage

Tools

To make the video, I used the following tech tools.

  • Adobe Photoshop Image Editor
  • Audacity Audio Editor
  • Capcut Pro Video Editor (you can use this as an audio editor as well)
  • YT-DLP to download videos and music from YouTube
  • UVR Online to split a song into stems and extract the vocals as a separate audio file
  • HeyGen AI video generator to convert a photo into a lipsynced video
  • Topaz Photo AI to upscale low-res photos

There are many alternatives to the tools above, but I like these the best.

Instructions

Get a song audio file

One easy way to get your song audio file is to find a song on YouTube and download it. I wanted this song:

I downloaded the audio MP3 file using the YT-DLP command line tool. The URL of the song was https://www.youtube.com/watch?v=D0ru-GcBIr4, which shows a video ID of D0ru-GcBIr4. So, to download the audio, I downloaded YT-DLP and ran the following command:

yt-dlp --extract-audio --audio-format mp3 --audio-quality 0 https://www.youtube.com/watch?v=D0ru-GcBIr4

If you are a simpleton and this looks confusing to you, then just use the online version of YT-DLP, which currently looks like this:

Get background video footage

Since the name of the song I made into a music video was 你本來就很美, which means “You are beautiful”, and the original music video showed beautiful beach scenes, I looked for similar beach footage on YouTube to use as the background for my video. I liked this video (https://www.youtube.com/watch?v=0ZBqnOeIxbQ):

Since the video ID was 0ZBqnOeIxbQ, I downloaded it from YouTube using YT-DLP using the command

yt-dlp 0ZBqnOeIxbQ

You don’t have to put a video in the background of your music video; you can also just put one or more still photos, but I think a background video looks better.

Separate vocals from the song’s audio

To improve lipsyncing and transcribing, we’ll need to separate the song into stems, where each track is a separate sound, e.g. vocals, instruments, etc. Go to UVR Online and upload the song’s audio file (MP3). You get process up to 12 minutes of audio per day for free.

When done, download the vocals track. Here’s what I got.

Note: I used Audacity to trim the audio to remove the silent sections to speed up AI-processing in Heygen and because Heygen has a duration limit per video.

Choose photos to lipsync

The photos that work best for lipsyncing are ones that

  • show the subject facing forward
  • are hi-res

These are the photos I picked for my video.

This last photo shows me looking to the side to add variety to the video. The AI lipsyncing results aren’t perfect, but it was acceptable, so I kept it.

Remove background from photos

I used Adobe Photoshop to remove the background from photos. For example, when I open a photo in Photoshop, I see a floating toolbar with a button called “Remove background”.

Clicking on it add a layer mask to the image layer

which causes the background to be transparent.

If the mask isn’t perfect and you see some of the background showing or some of the subject removed, you can edit the mask by

  1. clicking on the mask thumbnail in the layer,
  2. clicking on the paintbrush tool,
  3. adjusting the size of the paintbrush,
  4. changing the color of the paint to either white or black
  5. painting on the are of the image you want to show or hide

Change the photo aspect ratio to 16:9

Since I wanted to put my music video on YouTube, I wanted the video to be landscape format, 16:9, and 4K. I used Adobe Photoshop’s “crop’ tool to convert my portrait, 9:16 photos to 16:9. Notice that when I do that, I lose much of the photo, like the top of the hat and the shirt in the screenshot below.

To remedy this, first I expand the canvas wider than the original image and crop it.

Since my arms are cut off, I select those two areas and click the “Generative Fill” button.

The generative fill produces 3 variations. I picked one that looked the best.

This made my image 19:6 image look like this:

Change the background to neon green

Since we’ll want to change the background of our singing subjects to show our background video, we’ll need to put our photos on a green background so we can chroma key the green background out when editing the video in Capcut. In photoshop, change the foreground color to neon green.

Then, use the paint bucket tool to paint the background green. You may need to create a new layer positioned below the subject layer.

I repeated these steps for all other images, as necessary.

Upscale the images

Since my target video platform is YouTube and many devices like TVs support 4K, I upscaled some of my images to 4K. I used Topaz Photo AI to do this, but there are many alternatives that may be cheaper. With Topaz Photo AI, you can also do other things like sharpen, denoise, etc.

Create an avatar in Heygen

Now that we’ve collected and prepared all of our assets, we can create convert our photos to videos with lips synced to the lyrics. I created an account with Heygen and paid for cheapest monthly plan ($29 / month). This outputs 1080p videos only. If I wanted to created 4K videos, I would need the next plan at $39 / month. However, I think Capcut was able to upscale my 1080p footage to 4K using AI.

In Heygen, click on “Avatars” and create a new avatar by uploading your photos.

Create lipsync video

In Heygen, click the “Create video” button, choose “Landscape” for the orientation, and change the default avatar to one of your avatar’s “looks”.

The next part is important. For the script, click the “Audio” link and “Upload Audio”.

Upload the audio vocals file to Heygen, which has the backing music removed.

You can leave “Voice Mirroring” turned off so that the voice will be the voice in the uploaded audio.

Heygen will then try to transcribe the audio and display the words (lyrics) for the AI to produce the lipsync. Depending on the quality of the uploaded audio file, this may or may not work. In my case, it worked, and Heygen even detected the language as Chinese.

Click the “Generate” button and choose the appropriate settings.

Heygen will take some time render the video, which will show up under “Projects”. Repeat the above process for all photos (avatar looks).

Note: sometimes, some portions a video will have good lipsync results while others may not. In this case, you can try recreating the lipsync video.

Create the final video

To create the final video, I used Capcut. Though Capcut can be used for free, I paid for “Pro” access so I could use some of the paid features that I needed. I won’t go through all steps in detail since there are many general video editing tutorials online.

  1. Import (drag) all media (audio, video) to the “Media” pane (top left) in Capcut
  2. Drag your media to the timeline (bottom pane) to the appropriate tracks

The screenshot below shows the following tracks from bottom to top

  1. Audio track (complete song, not just the vocals)
  2. Main video track (background video showing different beach scenes)
  3. Other video tracks (since the lipsync results weren’t perfect for the duration of each Heygen-produced video, and because I wanted to show different versions of me signing different parts of the song, I chopped the videos into sections where the lipsync results were good)
  4. Text track (I copied and pasted the Chinese lyrics so viewers who read Chinese can read the lyrics as the video plays).

The tracks behave like layers in Photoshop, so media on higher tracks appear above media on lower tracks, which is why the tracks containing the green screen lipsync videos are above the track containing the background video.

Remove the green screen

To remove the green background from the lipsync videos,

  1. click on the video clip in the timeline
  2. in the top-right pane, click Video > Remove BG, Chrome key, and then, using the color picker, click on any area of the green background in the “Player” pane in the middle. The green background with suddenly disappear, revealing the media in the lower track (the background video).
  3. you may see some green artifacts around the edge of the subject. To clean up them up, slide the following sliders until you see good results: “Clean up edge”, “Feature edge”, ” Intensity”.

To improve the final video, you can do some other things as well, like

  • add a crossfade (“mix”) transition between adjacent video clips
  • Add a “Fade in” and/or “Fade out” animation to clips that are not adjacent to any other clips (go to “Animation” > “In” > “Fade In” and “Animation” > “Out” > “Fade Out”)
  • adjust the color of a video clip by going to Video > Adjust > Auto Adjust or Manual Adjust

Read my other post on how to work with Capcut.

Here’s another example of a music video I created using the same method.

HeyGen Avatar IV

HeyGen recently released Avatar IV, which allows you to create more realistic lipsync videos. To use it, make sure to click on Avatar IV in Heygen to use it. Here are some example inputs and outputs. The results are definitely more realistic!

INPUTS

There are 2 inputs:

  1. Image
  2. Script

For the image, just upload one image.

For the script, you can either type something or upload an audio file. Since we want to lip sync to a song, we’ll upload an audio file of the song containing just the vocals.

In the examples below, all script inputs will use the following audio.

INPUT

OUTPUT

INPUT

OUTPUT

INPUT

OUTPUT

INPUT

OUTPUT

Notice in the last test HeyGen can’t animate the background correctly.

There is yet another way to create a music video. It requires more effort, but the result may be more interesting. It uses Heygen, Krea, and Runway AI’s ACT-TWO feature.

Generate a Driving Performance Video

Follow the same steps above to create a lip sync video in Heygen to be used to drive another video. The background doesn’t have to be green. Make sure the resulting video is as follows:

  • Well-lit with defined facial features 
  • Single face framed from around the shoulders and up
  • Forward-facing in the direction of the camera

Since we will give Runway Act-Two a “character video”, the most important things in the “driving performance” video are the facial features and lip movements. What clothes you wear and the background are irrelevant as Runway Act-Two will just use your facial expressions in the “driving performance” video.

For example, here’s a “driving performance” video I created.

Split the driving performance video into 30-second clips

Runway’s Act-Two feature only lets you create videos that are 30 seconds long. So, we’ll have to split the driving video into a series of 30-second clips. If your driving video is 2 minutes long, you’ll end up with 4 clips, e.g.

  1. Clip 1 (0:00 – 0:30)
  2. Clip 2 (0:30 – 1:00)
  3. Clip 3 (1:00 – 1:30)
  4. Clip 4 (1:30 – 2:00)

Train a model of your face in Krea

Go to https://www.krea.ai/train and follow the instructions to train a model of your face. You will need a subscription to do this. Use the Flux model.

Train the AI model: Upload multiple hi-resolution photos of yourself (ideally around 40 with different angles and lighting) to the “train” section of Krea AI. You can specify whether you’re training a style, object, or character. You can choose “character” or keep the default setting.

Use the Flux model: Once your model is trained, click “Use with Flux” to generate images based on the trained model.

Generate a bunch of images of yourself singing

  1. Add your trained style: In the Flux model, click on “add style” and select your newly trained style from the “my styles” category. 
  2. Adjust influence: Use the provided slider to control how much your trained face style influences the generated images. Increasing the slider will make your face more recognizable, while decreasing it will reduce the resemblance. 

You can now use various prompts in Krea to generate diverse images incorporating your trained face. For example, you can make an image of you

  • singing in front of a mic facing the right with the camera close up
  • singing while holding a mic with the camera at a distance
  • playing drums
  • playing an electric guitar
  • etc

Here’s an example prompt with a reference image.

Text prompt

Subject is singing on stage in front of a band with a large audience in front of him. There are many multi-colored lights illuminating the stage.

If you’re unsure what prompt to enter, ask Chat-GPT to write a prompt for you specifically for Krea.ai.

Uploaded “image prompt”.

Output

Krea generated 4 images. After experimenting a few times, I chose this image.

Generate a bunch of 30-second videos

Use Runway’s image-to-video feature (not Act-Two) to create a series of videos based on the images generated in Krea. Here’s an example.

Input

You can create a video that is at most 10 seconds long. Since our song is 2 minutes long, we’ll need to create twelve 10-second videos, unless we use b-roll for parts of the final video.

Output

Create a lip sync video

In Runway, click on Act-Two. In the Performance section on top, upload your first driving performance video. In the Character section on the bottom, upload your first character video that you generated in the previous step. Click the button to generate the lip-synced video.

Runway will take the facial expressions and lip movements from your driving video and apply them to the character video. The resulting video will be the same duration as driving performance video – in this case, 30 seconds.

Here’s another example using a simple character video for demo purposes.

One-Day Visit to Madinah by Bullet Train from Jeddah, Saudi Arabia

Madinah is one of the two holiest cities in Islam after Mecca, and it’s only 2 hours away by bullet train from Jeddah, the 2nd largest city in Saudi Arabia. While visiting the kingdom, I was able to complete a one-day trip to Madinah from Jeddah with the following itinerary.

Itinerary

TimeActivity
10 AM – 12 PMRide bullet train from Jeddah to Madinah
12 AM – 3 PMVisit the Prophet’s Mosque (Masjid Al-Nabawi)
3 PM – 6:30 PMHave lunch
Visit the Qiblatain Mosque
Visit the 7 Mosques
Visit Mount Uhud
Visit the Quba Mosque
7:30 PM – 9:30 PMRide bullet train from Madinah to Jeddah

Costs

ActivityCost
Round trip bullet train ride between Jeddah and Madinah$155
6-hour car + driver/tour guide$120
Lunch at a Turkish restaurant for 3 people$43

Transportation

Bullet train

We purchased tickets online at https://rail.ninja/ for $72 per person. We chose the economy class because the business class didn’t seem worth the extra cost. The train travels at 300 km/h (186 mph). It’s clean, fast, convenient, and comfortable. There are 2 train stations in Jeddah, one in the north at the airport and one in the south.

Me at the train station in Madinah

Car + driver/guide

We thought about taking an Uber from place to place, but with so many places to visit, we didn’t want to waste time and try to figure everything out ourselves, so we decided to hire a driver for about 6 hours who could show us around. We were referred to Ibrahim, a Madinah local. He picked us up at the train station, took us to all the tourist spots, came with us inside the Prophet’s Mosque to show us around and explain things, had lunch with us and helped explain things to the waiter, and eventually dropped us off back at the train station. He doesn’t speak English, but I was able to communicate with my somewhat broken Egyptian Arabic I had learned while studying at the American University in Cairo, Egypt. If you don’t speak Arabic, you can always use Google Translate on your phone in “conversation” mode.

Ibrahim, our driver and guide in Madinah

Trip at a Glance

A friend dropped us off at the train station at the Jeddah international airport at 9:30 AM.

Left = train station, right = Jeddah International Airport

The train station was very spacious, clean, modern, and not crowded at all.

Train station at Jeddah International Airport

Your train e-ticket will come with a QR code. Just scan the QR code to open the turnstiles. And, of course, there is a Starbucks there with the Starbucks logo right above the arrivals and departures screens.

As you can see below, economy class was clean, comfortable, and spacious.

The bullet train travels at 300 km/h (186 mph).

On the way to Madinah, the scenery is mostly just desert.

Ibrahim, the driver/guide, picked us up at the train station in Madinah.

Our first stop was the Prophet’s Mosque (Al-Masjid Al-Nabawi).

There was convenient and ample parking below the mosque.

Below is a Google Earth view of the mosque.

There are many restrooms all around the mosque and huge umbrellas to provide shade during the day.

The Prophet’s Mosque is big. Ibrahim, our driver/guide, came with us inside to show us around and help guide us.

People from all over the world visit Madinah and the Prophet’s Mosque, like the person on the left, who may be from somewhere in Africa.

As we explored the mosque, I asked Ibrahim about various aspects of the mosque.

The interior design of the mosque was very elaborate with huge chandeliers.

I especially like the design of the striped arches.

People can be seen praying, relaxing, reading, and chatting in their ethnic attire.

Being from the US, I normally wear a baseball cap and simple Western clothes, but I decided to dress more like a local for this trip. See photo below. The headscarf is commonly worn in Arabia whereas the black turban on top is more north African. I bought both on Amazon. When I tried the headscarf on, it kept moving around and looked messy, so I used safety pins to hold the shape I wanted. I also wanted a visor to provide some shade on my face, so I cut a visor off of an old had, stuck it in between the folded scarf, and secured it using safety pins. For the black turban on top, I didn’t want to mess around with folding and wrapping it on me, so I bought one that came pre-shaped that you wear like a hat. It had 2 straps on the back to tighten it to your head.

Inside the mosque, there were many water dispensers contain Zamzam water.

All over the mosque, both inside and out, you can find free Zamzam water, which was necessary considering the very hot climate.

I must have drank 10 of these cups of water. Next time, I would bring a water bottle with me.

If you visit Prophet Muhammad’s tomb, it will likely be very packed with people as shown in the photo below.

All around the mosque there are numerous golden doors with intricate designs.

After visiting the Prophet’s mosque, we went for lunch. I wanted Turkish food. Ibrahim took us to a place somewhat distant from the touristy Prophet’s Mosque area where it wasn’t crowded and the prices were more reasonable. Though I was able to read the Arabic menu, I didn’t understand what everything was. Luckily, Ibrahim was able to explain everything to both me and the waiter, who happened to be Egyptian. Apparently, there are many Egyptians working in Saudi Arabia.

We ordered a mixed plate that was enough for all 3 of us, including the driver, Ibrahim. The food was acceptable and fair for the price.

We then went to the Qiblatain (Two Qiblahs) mosque, where prophet Muhammad was commanded to move the qiblah (prayer direction) from the Aqsa mosque in Jerusalem to the Haram mosque in Mecca.

Since it was super hot, I asked our driver, Ibrahim, to take us to get some ice cream.

I ordered a mix of pistachio and vanilla ice cream with pistachio shavings on top. Pistachio seems to be very common in the Middle East, like in Turkey.

We then went to the Seven Mosques, a historical collection of small mosques.

We then went to Mount Uhud, a historical site where the prophet Muhammad had a battle with the polytheists of his Quraysh tribe.

Here’s a photo of me (right) with our driver, Ibrahim (left).

Our last stop on our one-day Medina tour was to the Quba mosque, the first mosque to exist. This was one of the busier mosques we visited. Many people were praying and reading the Quran, and there were groups of kids learning something.

After that, Ibrahim took us back to the train station in time for us to catch our 7:30 PM train ride back to Jeddah.

Video

Watch the full video of my one-day excursion to Madinah.

How to Securely Install a Toilet to a Floor

As a real estate investor, I’ve had to replace many toilets. Normally, I do that while doing a complete renovation. The problem with many toilets is

  1. the bolts to the toilet flange in the floor can get loose
  2. the floor may not be perfectly level

These issues can cause a toilet to slide, rotate, and tilt.

Fixing a rocking toilet due to an uneven floor

After renovating many bathroom floors with both tile and vinyl, it’s unfortunate that despite a contractor’s best effort to create a perfectly level floor, a small imperfection could cause a toilet to not sit evenly on the floor. The easily way to fix this is by installing plastic toilet shims on the LOW side of the toilet. Make sure to install it on the LOW side as the weight of the toilet will help keep the shim in place.

Fixing a loose toilet due to loose bolts

The bolts that come with many toilets are usually the cheap kind. Sometimes, they work just fine, but depending on your toilet flange and toilet, they may easily become loose, which is what happened to me. To remedy this, I replaced the cheap toilet flange bolts with the Fluidmaster SetFast 5/16 in. x 2-1/4 in. No-Cut Brass Closet Toilet Bolt Kit. Price at $8, which is about double the cost of the cheap bolts, these bolts don’t require any cutting and they securely hold your toilet in place. Unlike some other toilet bolts that come with a small head, the bolt head the Fluidmaster bolt is thick and wide, so the bolt remains within the toilet flange slot no matter what.

When you tighten the acorn nut, the height of the threaded stud decreases automatically.

The yellow lock washer is thick, which holds the bolt in place vertically as you place the toilet over the bolts. This is way better other toilet bolt kits, which either don’t come with a lock washer, or the lock washers are thin and flimsy.

Stur: A Healthy and Convenient Flavored Drink Option

It seems like there are just more and more drink options available when you visit a supermarket. And while some of them are healthy, most of them are not. The ones that aren’t contain either too much sugar or artificial ingredients like sweeteners and food coloring. For a while, I’ve been making green tea sweetened with all-natural monk fruit mixed with Stevia. The problem with that is I’m often too busy to prepare it and it takes time for the tea to brew. Fortunately, there is now both a healthy and instant flavored drink alternative called Stur.

I’ve been trying Stur now for about a week, and I’m pretty impressed. It checks all the checkboxes for a healthy drink in addition to being super easy to prepare: just get a glass of cold water, squirt the liquid into the water, and drink. You don’t even need to refrigerate the small bottle, so you can bring it with you to restaurants where you can save a lot of money by not paying a huge markup for flavored drinks.

There are many flavors to choose from. So far, I’ve tried

  • Blue and blackberry
  • Strawberry watermelon
  • Fruit punch

and all 3 were good.

You can buy the ultimate variety pack on Amazon for $45.

Video Editing with Capcut

In this post, I’ll share my process for how to make a traditional video in Capcut. Specifically, I’ll show how to make a video that:

  • comprises a collection of video and image clips
  • has a simple crossfade (mix, fade-in/fade-out) transition between clips
  • has background music with a crossfade between music clips
  • improves audio quality
  • uses AI to stabilize some shaky video footage
  • improves the color and lighting of some clips
  • animates some images with a subtle zoom effect
  • animates some images to create rolling credits
  • freezes a frame in a video clip
  • has text overlays

Download and install Capcut

Visit the Capcut website, download Capcut and install it.

Import video and image footage

  1. Open Capcut
  2. Create a new project
  3. Click the default project name at the top center and rename it to something descriptive, e.g. My Vacation
  4. Click the “Media” tab at the top left and then click “Import” tab below it.
  5. Click the “Import” button or drag your footage (audio, video, images) to the assets pane.

Use the filter dropdown to filter assets by type (video, audio, image, etc).

Use the sort dropdown to sort assets, e.g. by date created, etc. I find that sorting by “Time created” and “Earliest to latest” puts my assets in the order I want them in so I can just select all assets and drag them to the timeline.

Drag some or all of your footage to the timeline at the bottom

Drag the zoom slider at the top right of the timeline pane to zoom the timeline in and out.

Trim images

When you drag images to the timeline, the duration will be whatever the setting is under Options, which defaults to 5 seconds. If you want to change the default to, say, 3 seconds, go to Menu > Settings > Edit (tab) > Image Duration and change the value to 3.

Click on an image in the timeline. In the right pane, you will see various options to edit the image.

The playhead is a visual marker that indicates the current position or frame being viewed in the preview pane. Drag the playhead to anywhere above the image in the timeline. You will see a preview of the image in the center in the preview pane.

Click the play button in the preview pane to preview how the image will appear relative to the rest of the video project.

If the image duration is too short or too long, you can extend the duration by dragging the left or right edge of the clip in the timeline to the left or right. You can also position the playhead at a particular time and click one of the following buttons:

  1. SPLIT: to split the image clip into two clips
  2. DELETE LEFT: to delete the portion of the image to the left of the playhead
  3. DELETE RIGHT: to delete the portion of the image to the right of the playhead

To see the duration of an image, zoom out on the timeline and look at the duration label on the clip in the timeline. For example, the image clip below is 3 seconds long.

The video clip below is 23 seconds and 8 frames long.

Trim Videos

Trimming video clips on the timeline is similar to trimming images. Click on a video clip in the timeline. You will see a preview of the video in the preview pane. In the right pane, you will see various options to edit the video clip.

In the right pane, click “Speed” to see the video clip’s duration. If you change the duration value here, the video clip’s speed will change accordingly. If you don’t want to speed up or slow down the video, you’ll need to trim the video clip the same way you trim image clips:

  • drag the left end of the video clip in the timeline to the right
  • drag the right end of the video clip in the timeline to the left
  • split, delete left, or delete right the video clip relative to where the playhead is in the timeline

Speed up or slow down video clips

If your video clip is too slow or too fast, you can speed it up or slow it down. As in the previous step, click the video clip in the timeline and then click the “Speed” tab in the right pane.

Speed up

To speed up the video clip, e.g, 2x for twice the speed, just drag the speed marker to the right. The duration of the clip will change accordingly.

Slow down

To slow down the video clip, e.g. 0.5x for half the speed, drag the speed marker to the left. The duration of the clip will change accordingly. Note that if your video clip was shot at 30 frames per second (fps) and you slow down the clip by 50% (0.5x), the resulting video will stutter due to an insufficient number of frames. This is why you should shoot video at a high frame rate, e.g. 60 or 120 fps, if you know you want to play it back in slow motion, which is usually done for action scenes.

If you didn’t shoot at a high fps, you can use Capcut to smoothen the slow-mo effect by either

  • frame blending
  • optical flow

The results are not as good as a high-fps video, but they’re better than without smoothening, especially using the “optical flow” option.

Add transitions between clips

There are many transition effects available. I prefer to use the simple crossfade (mix) transition, which gradually fades out one clip and fades in the next clip. To add this transition between two adjacent clips, click the “Transitions” tab in the left pane, type “mix” in the search field, and drag the mix transition icon down to the timeline between two clips. If necessary, zoom the timeline out. You’ll see a semi-white section between the two clips representing the transition. If you move the playhead to that transition area in the timeline, you can see a preview of the transition in the preview pane, showing a blending of the two clips.

If you want to apply the same transition to all clips, click one transition in the timeline and then, in the right pane, click the “Apply to all” button.

Add background music

  1. Click the “Audio” tab in the left pane.
  2. Click the “Import” button to import songs (MP3) from your computer.
  3. Drag a song from the list of songs down to the timeline below the main video track.

As with still images and video clips, you can trim audio clips in the same manner.

If you want to crossfade two adjacent songs,

  1. put the 2nd song on a separate track below the first song’s track
  2. zoom in on the timeline and scroll to where the two songs meet
  3. click the 1st song and, in the right pane, set a fade out duration of, say, 1 second
  4. click the 2nd song and, in the right pane, set a fade in duration of, say, 1 second
  5. position the playhead just before the end of the first song
  6. preview the audio crossfade in the preview pane

Notice the black curve in the audio tracks showing the fade effect.

When you add audio or music track, if your video clips contain audio, you’ll hear audio from all tracks containing audio. If you want to mute all audio from all clips in the main track, click the audio icon as shown below.

If you want to adjust the volume or mute just a single video clip, click on the clip in the timeline, then in the right pane, click the “Audio” tab and drag the volume slider.

Improve audio quality

The audio in your video clips may contain background noise as well as vocals. To improve the audio quality, select the relevant clips and choose from the following options in the “Audio” panel.

  • To reduce background noise, e.g. humming or the sound of an engine, check the “Reduce noise” checkbox
  • To improve vocals, check the “Enhance voice” and/or “Isolate voice” checkboxes. Note that the “Isolate voice” options strips out all audio, leaving only vocals.”
  • To normalize the loudness of all background music clips, select all clips and check the “Normalize loudness” checkbox.
  • Slide the volume slider to increase or decrease the volume as needed.

Adjust color and lighting

You can adjust the color and lighting for both image and video clips individually.

  1. Click on a clip in the timeline
  2. Move the playhead to that clip
  3. In the right pane, click “Adjust” > “Basic” > “Auto adjust”
  4. Slide the “Auto adjust” intensity slider until you like how the clip looks in the preview pane.

You can also manually adjust the color by tweaking various color settings in the right pane.

I find adjusting color is especially helpful for brightening a dark image or video. Here’s how one clip looks before and after applying “auto adjust”.

Before
After

Stabilize a shaky video

If your video footage was taken with a camera that doesn’t include mechanical stabilization (like a gimbal) or software stabilization, then the resulting footage could be annoyingly shaky. Capcut can try to stabilize your footage at the expense of cropping a portion of the video. To stabilize a shaky video,

  1. click on a video clip in the timeline
  2. in the right pane, click the “Video” tab and check the “Stabilize” option.

Animate images

For certain photos, I like to apply a subtle zoom-out animation effect to them. To do this,

  1. click on the image clip in the timeline
  2. move the playhead in the timeline to where the image clip is
  3. click the up arrow key to move the playhead to the beginning of the clip
  4. in the right pane, click “Video” > “Basics”, change the “scale” value to 200% and click the diamond icon to set a keyframe
  5. click the down arrow key to move the playhead to the end of the clip
  6. in the right pane, click “Video” > “Basics” and change the “scale” value to 100%
  7. drag the playhead from the beginning to the end of the clip. You should see the image zoom out. You can also click the play button in the preview pane to preview the animation.
Playhead is at the beginning of the clip
Scale at 100%
Scale at 100%
Scale at 200%
Scale at 200% with keyframe set
Playhead at end of clip

Create rolling credits

There are different ways to create rolling credits. The way I’m about to show you involves slowly animating an image’s position upwards. Therefore, you’ll need to create a tall image with the content you want in it, like this

  1. Add this image to the timeline
  2. Select the image in the timeline
  3. Position the playhead in the timeline where the image is
  4. Click the up arrow key to move the playhead to the beginning of the clip
  5. In the right pane, click “Video” > “Basics”
  6. Change the “scale” value until you like how the credits image looks in the preview. In the screenshot below, I set it to 500%.
  7. Change “position” Y value to a value that moves the top of the image near the bottom of the preview pane. In the screenshot below, I set it to -3300.
  8. Click the diamond icon to set a keyframe
  9. Click the down arrow key to move the playhead to the end of the clip
  10. in the right pane, click “Video” > “Basics” and change the “position” Y so that the bottom of the image is near the top of the preview pane. In the screenshot below, I set the value to 3300.
  11. Drag the playhead from the beginning to the end of the clip. You should see the image roll up. You can also click the play button in the preview pane to preview the animation.

Freeze a frame in a video clip

Sometimes, you may want to freeze a frame in a video clip for a few seconds. For example, in the screenshot below, there’s a video clip of a vehicle moving along a road and a sign showing various destinations. To freeze the frame at the point when the sign is shown,

  1. move the playhead to the timestamp where you want to freeze a frame
  2. click on the video clip in the timeline
  3. click the “freeze” button above the timeline

The video clip will be split at that point and you will see a still image of the frame added to it. You can then adjust the duration of the image. Make sure not to add any transitions between the video clip and the still image. You can group the two clips or create a component to prevent accidentally adding a transition between them.

After freezing the frame, the video clip is split and a still image of the frame is appended

Add a text overlay

There will be times when you’ll want to overlay text on a scene to describe a location or what’s happening as in the following screenshot.

To do this,

  1. click the “Text” tab in the left pane
  2. drag the “Default text” item to the timeline on a track above the clip where you want the text to appear
  3. drag the ends of the text clip in the timeline to increase or decrease the duration
  4. click on the text clip in the timeline and, in the right pane, edit the text and various other properties, e.g. background color and background opacity
  5. in the preview pane, drag the text overlay to where you want it. you can also drag the corners to resize it.

Add a marker on the timeline or a clip

When editing a video, it can be helpful to add markers at certain points to remind you where a specific moment is. For example, you may want to align some a music clip to a specific moment. To do this,

  1. move the playhead to the timestamp where you want to add a marker
  2. click the marker button above the timeline

You will see a light blue marker added on the timeline.

Export your video

Click the light blue “Export” button at the top right to export the video.

Export a still image of a video clip

  1. move the playhead to the timestamp where you want to freeze a frame.
  2. click the hamburger menu (3 horizontal lines) at the top-right corner of the preview pane
  3. click “Export still frames”
  4. if you don’t want the still image added to your project, uncheck the “Import project” checkbox

Create a fast-rewind effect followed by a slow-motion effect

For fast-moving scenes, you may want to replace the scene in slow motion. One way to do this is to show the original clip (1) followed by a fast rewind of the clip (2) followed by a slow-motion version of the clip (3) as shown in the screenshot below.

To do this,

  1. copy and paste the original clip twice and place the copies after the original clip on the timeline
  2. click the 2nd clip to select it and then click the “Reverse” button above the timeline
  1. Change the speed of the 2nd clip to something like 5x.
  1. click the 3rd clip to select it and then change the speed to something like 0.2x.

Easily Remove Grout Haze From Tile

When applying grout to tile, you’re supposed to wait 15-30 minutes before wiping the grout of the tile while leaving the grout between each tile. This is easier said that done. No matter how good you are, you’ll probably be left with a thin layer of grout on your tile. If you use a light-colored grout on light-colored tile, you won’t notice the haze except at certain angles in certain lighting. The haze can resemble hard water stains. If you use a dark-colored grout on light tile or a light-colored grout on dark tile, the haze is more prominent, and your tile can look dirty. See example below.

Before

You can try to scrub with soap or even a steam cleaner, but that will have zero effect. Luckily, there is an amazing chemical product that can relatively easily remove the grout haze from tile. With Aqua Mix 1 Qt. Cement Grout Haze Remover by Custom Building Products, you just mix the liquid with water, pour some one the tile, wait a while, then rub the haze off. For tougher haze, don’t mix with water. In my case, I poured the liquid at full strength without adding water into a spray bottle, sprayed the liquid on the tile, then wiped the haze off with a rag. As most comments say on the Home Depot product page, this stuff is “amazing”!

After
Aqua Mix 1 Qt. Cement Grout Haze Remover by Custom Building Products

Quickly and Easily Remove Weeds From Your Driveway

To easily remove weeds from control joints in your concrete driveway, use an angle grinder with a wire wheel. I use a 4.5″ Ryobi brushed, cordless angle grinder. Don’t use a brushless angle grinder because it can’t be used for this purpose. I prefer using a knotted wire wheel. A non-knotted wire wheel works as well, but I find it better for cleaning control joints after removing large weeds using the knotted one.

Knotted Wire Wheel

Non-knotted Wire Wheel

To prevent kickback, stand and drag the tool to the right as shown in the photo below.

If the angle grinder flange lock nut is stuck, use either a pipe wrench or a thin wrench to unlock it. For convenience, clamp the angle grinder to a stationary vise.

English Muffin Breakfast Recipe

A few years ago, I flew to Korea on Hawaiian Airlines and was served a warm English muffin containing a turkey patty and egg. As cheap and simple as it looked, I wasn’t expecting much, but it turned out to be so good, I had to try to recreate it. I think I figured it out. Here’s my recipe.

Ingredients

Instructions

  1. Spray some cooking oil on a frying pan.
  2. Fry the egg in the shape of the muffin. I like to use this egg pancake frying pan that I got on Amazon. Optionally, pierce the yolk.
  3. Sprinkle some salt and pepper on the egg and cover the pan so the top of the egg gets cooked.
  4. Slice the muffin into 2 halves.
  5. Heat the muffin in a microwave for 1 minute to warm and soften it up. This is especially necessary if the muffins were refrigerated, cold, and hard.
  6. Since the turkey patties are small, defrost 2 of them in a microwave for 2 minutes and 30 seconds.
  7. Spread some hummus on all inner sides of each muffin half.
  8. Cut one of the turkey patties in half so that one and a half muffins can cover most of the muffin.
  9. Put 1.5 turkey patties next to each other, cover it with half a cheese slice, and microwave for 25 seconds.
  10. Assemble the muffin as shown in the photos below.

A Collection of AI Tools for Various Purposes

This is just a collection of AI tools I’ve used that I’ve found useful. With so many AI tools sprouting up, this list will likely be updated regularly as time permits.

AI Portal

Image Editing

Audio Editing

Speech

Video Editing

Text Editing

  • ChatGPT
  • Grammarly

Code Editing

  • Cursor
  • Bolt.new
  • Bolt.diy
  • Claude 3.7 Code
  • GitHub Copilot
  • Coedium’s Windsurf
  • Locofy.ai
  • Vercel v0
  • Dora.ai
  • Relume
  • replit
  • lovable

Install AI Apps

Color Palette

Virtual Staging

Different Ways to Build a Component-Based Static Website

Static websites are fast and ideal for many types of websites like blogs, marketing websites, and more. When building websites, you should always use a component-based approach for code simplicity, maintenance, and reusability. Here are a few ways to build a static website using components.

  1. Web Components – No framework
  2. Eleventy (11ty) – Very simple and flexible static site generator that supports various template languages (Handlebars, Nunjucks, etc)
  3. Astro – Very similar to Eleventy, with the advantage of being able to load premade components from React, Vue, Svelte, etc.
  4. Svelte – More like React, but better. Outputs static files. Has native benefits like CSS optimizations and the ability to create interactive apps, if necessary.

The following video does a great job in comparing and demonstrating building a static site using web components and Svelte.

▲