Why YouTube says no to an MP3

YouTube's own list of supported formats is all video: MP4, MOV, AVI, WMV, WebM and so on. Its help page says it straight out: "Audio files, like MP3, WAV, or PCM files, can't be uploaded to create a YouTube video. You can use video editing software to convert your audio file to a video." The only place an audio file goes is as an extra language track on a video you've already uploaded (YouTube Help).

So whatever route you take, the thing you upload is a video with your audio in it.

Way 1: a still image in Clipchamp (free, already on Windows 11)

Clipchamp is the official Windows video editor. On Windows 11 you can find it from Start or by searching for it in the taskbar, and on Windows 10 it's a free install from the Microsoft Store (Microsoft Support). The free plan exports at up to 1080p with no watermark (Clipchamp pricing).

1. Start a new video and click Import media to add your MP3 and an image. Clipchamp takes video, images and audio (Microsoft Support).
2. Drag both onto the timeline, the image on the video track and the MP3 under it.
3. Stretch the image to the end of the audio. An image starts at four seconds long. Click it on the timeline and drag its right hand trim handle until it lines up with the end of the MP3. There's no box to type an exact length, so zoom the timeline in near the end to get it right (Microsoft Support).
4. Click Export at the top right and choose 1080p. It saves an MP4 at 30 frames a second (Microsoft Support), and Clipchamp can upload it to YouTube for you from the same screen.
5. Upload the MP4 in YouTube Studio like any other video.

What image to use: a 1920 x 1080 picture so it fills YouTube's 16:9 player. YouTube's own advice is to avoid adding padding or black bars to the video itself (YouTube Help). A square cover art file will sit in the middle with bars either side, so either put it on a 1920 x 1080 background first or let the bars be.

For reference, the encoding YouTube recommends (written for its Content Manager partners, but it's the standard reference) is an MP4 with H.264 video, AAC audio at 48 kHz, and 384 kbps for stereo audio (YouTube Help). Clipchamp's MP4 export is already in that shape, so you don't need to change anything.

Way 2: let YouTube do it from your podcast's RSS feed

If you have an audio podcast with an RSS feed, you don't need to make any videos. YouTube can read the feed and "use your podcast's show art to create a static-image video and upload it to your channel on your behalf", and new episodes upload automatically from then on (YouTube Help).

1. In YouTube Studio, click Create, New podcast, Submit RSS feed.
2. Accept the RSS terms, paste your feed URL and verify it with a code sent to the email address in your feed. If your channel doesn't have advanced features yet, you'll be asked to verify your identity.
3. Pick which episodes to bring in: all of them, those since a date, or only future ones. Uploads can take a few days, and they arrive as private videos you then publish (YouTube Help).

Things to know first:

• It's not in every country. The channel owner's location decides it, and YouTube keeps a list of about 100 countries where it works (YouTube Help).
• No ads in the audio. YouTube's page says podcast content uploaded this way "cannot contain advertisements", and branded content and host read promos have to be declared (YouTube Help). If your episodes carry dynamically inserted ads, read that page closely.
• It's always the same still picture. Fine for people who listen in the background, not much to look at.

Way 3: a waveform video, for something that moves

A still image is fine for a full album track or an hour long episode that people leave running. For a short clip, a song you're sharing, or anything that has to stop someone scrolling, a picture that moves with the sound does more. That's what a waveform video (an audiogram) is: your cover art and title, with bars or a ring that move to the audio, and often captions.

This is what I built Wayvl for. It's free and runs in your browser, and the audio never leaves your device: the decoding, drawing and encoding all happen in the tab.

1. Drop in your MP3 (or WAV, M4A, AAC, OGG, FLAC, or a video file).
2. Trim it to the part you want, up to five minutes.
3. Pick 16:9 for a normal YouTube video (1920 x 1080), or 9:16 for a Short (1080 x 1920). 1:1 and 4:5 are there too, and you can export several sizes in one go.
4. Choose a template or a wave style, add a title, your cover art and your handle.
5. Turn on captions if you want them, then Export MP4 and upload that to YouTube.

A few things in it worth knowing about:

• Captions made on your device. Speech to text runs in the browser with Whisper, so the audio still isn't uploaded. Tap any word to fix it, and download the captions as SRT, VTT or text if you'd rather upload them to YouTube separately.
• Songs work too. Captions on singing used to come back empty because the model answers music with a [Music] tag, so there's a song mode now. And if you have the lyrics, lyrics mode lines your pasted words up with the singer.
• 11 wave styles, gradient fills, templates, and a brand kit that remembers your look in that browser.

Where Wayvl doesn't fit

• Five minutes at most. Longer clips take too much memory to encode in a browser tab. For a full podcast episode or a long track, use Way 1 or Way 2.
• Big files. Expect roughly 1 to 1.5 MB per second of video at the default quality, so a three minute song is a few hundred MB. YouTube takes that fine, it just takes a while to upload.
• It needs a recent browser to export: Chrome or Edge on a computer, Chrome on Android, or Safari from iOS 16.4 and macOS 13.3. Firefox can preview but not export.

When a still picture is enough

• Full episodes and long mixes, where people listen with the tab in the background: a still picture or the RSS route is fine, and it's far less work.
• Clips, Shorts and singles you want people to actually watch: motion and captions give them a reason not to scroll past.

One thing a video upload won't do: YouTube Music

Uploading a song as a video puts it on YouTube. It doesn't make it a proper release on YouTube Music. For that YouTube points you to a distributor: "Even if you upload your music videos to YouTube through your artist channel, you'll need to have your music audio files and metadata delivered to YouTube" through one, and an Official Artist Channel needs at least one release delivered by a distributor or label (YouTube Help, YouTube Help).

And only upload audio you own or have the rights to. Someone else's song with a picture on it is still their song.

Wayvl is free to use at wayvl.com.

What are you putting on YouTube: songs, a podcast, or something else? I'm keen to hear what people need from the audio side.

Sources

• YouTube Help, Supported YouTube file formats, Recommended upload encoding settings and Video resolution and aspect ratios
• YouTube Help, Deliver podcasts using an RSS feed, Publish episodes from an RSS feed and where RSS delivery is available
• YouTube Help, Music aggregators and Official Artist Channels
• Microsoft Support, Clipchamp: What is Clipchamp, editing a video, image duration and exporting
• Clipchamp, pricing and free plan
• Wayvl's site and build notes (25 and 26 Sep 2026), for the formats, the five minute limit, file sizes, browsers, captions and lyrics mode

This post first appeared on sundrymill.com.

I build a Windows screen recorder, and the first vertical trailers I made for it were honestly confusing to watch. Here's why that happens, the free ways around it, and the layout I ended up building.

Why vertical screen recordings look bad

It's one number. A 9:16 crop of a 1920 x 1080 screen is 608 pixels wide. That's 32 percent of the screen, before you zoom in at all.

So a plain centre crop throws away two thirds of the picture, and whatever you click near the edges (menus, the send button, the sidebar) simply isn't in the video. That's what went wrong with my first trailers: the cursor and half the words were out of frame.

The other way round is just as bad. Fit the whole wide screen into a 1080 pixel wide frame and everything shrinks to 56 percent, so 14 px text ends up about 8 px tall in the video, with a big empty band above and below it. On a phone that's unreadable.

Every method below is a way of dealing with that trade.

What size to make it

All three apps take a 9:16 vertical video, and 1080 x 1920 is the size most people export at.

• YouTube Shorts are vertical or square videos up to three minutes long (since 15 October 2024), at up to 1080p (YouTube Help, YouTube Help).
• Instagram Reels take anything from 1.91:1 to 9:16, at least 720 pixels and 30 frames a second (Instagram Help).
• TikTok recommends 9:16 vertical for its ads, at 540 x 960 or more (TikTok Ads).

Way 1: record something narrow in the first place

The simplest fix is to not record a wide screen at all.

1. Make the window narrow. Drag your browser or app down to a tall, thin window. Most websites switch to their phone layout, with bigger buttons and text that reads at phone size. In Chrome you can also open DevTools and use Toggle device toolbar to pick a phone sized width (Chrome for Developers).
2. Record just that area. Press Windows + Shift + R to open the Snipping Tool's recorder, drag a rectangle around the window, then select Start (Microsoft Support). The rectangle is free hand, so it won't be exactly 9:16. Get it close and fix the last bit in your editor.

Good for: websites and apps that have a mobile layout. Bad for: desktop software that needs the full width, like a spreadsheet or a video editor.

Way 2: a vertical canvas in OBS

OBS records whatever its canvas is, and the canvas doesn't have to be wide. Under Settings, Video, the Base (Canvas) Resolution is "the amount of space you have to fit your sources" and the Output (Scaled) Resolution is what the recording comes out at (OBS Knowledge Base). Set both to 1080 x 1920, add a Display Capture or Window Capture source, then hold Alt and drag the source's edges to crop it down to the part of the screen you want.

If you also want a normal wide recording of the same session, the free Aitum Vertical plugin gives OBS "a separate canvas to manage your vertical content", with its own Record vertical button that can start and stop with the main recording (OBS forum, Aitum Vertical). It needs OBS 29 or later.

The catch with both: the crop is fixed while you record. If you click something outside it, it isn't in the video. So you either keep your clicks inside the box or move the crop as you go.

Way 3: crop a normal recording in your editor

Record the full screen as usual, then reframe it.

• Clipchamp (free, already on Windows 11): click the size button on the toolbar above the preview and pick 9:16. The video sits in black bars until you click Fill, which crops it to fill the frame (Microsoft Support).
• CapCut desktop: choose Ratio and switch to 9:16, then try Auto reframe in the Basic section, or use Crop above the timeline and pick 9:16 (CapCut, CapCut).

Either way you start from the centre of the screen. To follow the action you have to split the clip wherever you click somewhere new and move the frame for each piece, and that's where the time goes.

The stacked layout, by hand

The layout that reads best on a phone, in my experience, is two panes: the whole screen small at the top, so people know where they are, and a close up underneath, so they can actually read it. You can build it in any editor that does layers. Put the recording on the timeline twice, shrink the top copy to fit the width, and crop and scale the bottom copy to the area you're working in. Then move the bottom one every time the action moves.

Keep it inside the safe zone

The apps draw over the top and bottom of your video: tabs at the top, the caption and account name at the bottom, and the like, comment and share buttons down the right.

• Meta asks Reels advertisers to leave "at least 14% of the top, 35% of the bottom and 6% on each side" free of anything important (Meta Ads Guide). On a 1080 x 1920 video that's about 269 px at the top and 672 px at the bottom.
• TikTok says its safe zone depends on the shape, the caption length and the format, and gives downloadable safe zone files rather than fixed numbers (TikTok Ads).
• YouTube doesn't publish Shorts safe zones that I could find.

Those are ad rules, but the same buttons sit over normal posts. My own rule of thumb in Framok is to keep the top 160 px, the bottom 400 px and a 140 px column down the right clear on a 1080 x 1920 export.

And make text big before you hit record. Browser zoom at 150 to 200 percent costs nothing and makes a huge difference on a phone.

The automatic way: Framok's Stack

This is the problem I ended up building into Framok, my screen recorder for Windows. You record the full screen once, and the 9:16 export is the stacked layout from above, done for you.

• The whole screen on top, almost the full width of the video, with a rounded box showing where the camera is looking.
• The close up underneath, following your clicks and your typing. Before each click it settles on the button, row or field you're about to press and holds still until the result has shown.
• Inside the safe zone. Both panes stay clear of the tabs, the caption line and the buttons on the right. On a 1080 x 1920 export the top pane is 984 x 553 and the close up 892 x 783, and together they take 70 percent of the frame height.
• A bigger cursor in tall videos, since that's the thing people follow on a small screen.
• Two presets: Reels, TikTok, Shorts (9:16 Stack) and Feed post (4:5 Stack). Files come out named after their shape, like recording-9x16-stack-1080p.mp4.
• Fill and Fit are still there. Fill crops to the frame and follows the action, framing the whole button or field under each click so words don't get cut off. Fit is the whole wide picture in a band.

When I checked Stack on a demo take, none of the 24 click moments had the cursor or the clicked thing out of frame.

Where Framok doesn't help

• Only recordings made in Framok. Stack needs the click and cursor timing Framok saves while it records, so it can't reframe a video from OBS or your phone.
• No captions yet. Lots of vertical videos get watched with the sound off, so you'll still want captions from Clipchamp, CapCut or the app you post with.
• Windows only. Windows 10 or 11, 64 bit. A Mac version has a waitlist at the bottom of framok.com.
• No free trial on the download from framok.com. A 7 day trial comes with the Microsoft Store version, which isn't out yet, and paid licences have a 14 day money back guarantee.
• The installer isn't code signed yet, so SmartScreen warns the first time. Click More info, then Run anyway.

Framok is $39 once for the life of the product, or $5 a month and cancel any time. Both cover 2 of your own PCs.

See Framok

How do you do your vertical cuts at the moment: narrow window, OBS, or cropping after? Keen to hear what's working.

Sources

• YouTube Help, YouTube Shorts up to three minutes and Upload YouTube Shorts
• Instagram Help, Reels sizes
• Meta Ads Guide, Instagram Reels video ads
• TikTok Ads Manager, Video ads specifications, updated June 2026
• Microsoft Support, Use Snipping Tool to capture screenshots and Change the aspect ratio of a video in Clipchamp
• OBS Knowledge Base, OBS Studio overview, and the Aitum Vertical plugin page
• CapCut, How to edit vertical videos and How to crop a video in CapCut PC
• Chrome for Developers, Device Mode
• Framok's changelog (0.7.7 to 0.7.11) and build notes from 24 and 25 Sep 2026, for the 32 percent figure, the pane sizes and the 24 click check

This post first appeared on sundrymill.com.

Silent demos are cooked imo. A click sound tells the viewer something happened even when they're half watching on a phone, and typing sounds make a form fill feel like a person is doing it. Here's how to add them for free, then the automatic way I built into my screen recorder.

Where to get mouse click and keyboard typing sounds you're allowed to use

Search "mouse click sound effect" and you get pages of YouTube uploads with "no copyright" in the title. A title isn't a licence. Get sounds from somewhere that says what the licence is.

• Your editor's own library first. Clipchamp has sound effects under the content library, Audio, Sound effects, and says all of its stock audio is royalty free for personal and commercial use (Clipchamp). Worth a search for "click" or "typing" before you download anything.
• Kenney. Interface Sounds and UI Audio are free packs of clicks, switches and ticks, released as CC0. That means no credit needed, commercial use fine. Two of Framok's click styles come from these packs.
• BigSoundBank. Joseph Sardin's library of real recordings, including a MacBook Pro keyboard and a typewriter. The free sounds are CC0, and the licence page says a credit is appreciated but nothing is mandatory.
• Freesound. Huge, but every sound has one of three licences (Freesound FAQ). After a search, filter the licence to Creative Commons 0 and you're safe. Attribution means you credit the author in your description. Attribution NonCommercial means you can't earn money from the video, so if your channel is monetised, skip those.

Grab several takes of the same key, not one. You'll need the variety later.

How to add click sounds to a video by hand

This works the same in Clipchamp, CapCut, DaVinci Resolve or Premiere.

1. Import the sounds and give them their own audio track under the video, so you can move and mute them without touching your voice.
2. Find the frame where the button reacts. Zoom the timeline in around the click and step through frame by frame (in DaVinci Resolve the left and right arrow keys step one frame, Blackmagic Design). The moment to hit is when the button changes, the menu opens or the checkbox ticks. The cursor arriving over the button is too early.
3. Put the start of the click sound on that frame. Trim any silence off the front of the file first, or every click will land late.
4. Copy it to the next click and repeat.

The honest downside is that last step. When I built the automatic version, my 34 second test take had 57 clicks and key presses in it. That's 57 sounds to place by hand for half a minute of video.

How to add typing sounds to a video

Typing is where hand placed sounds go wrong, because one key sound repeated fast sounds fake straight away.

• Long runs of typing: lay a typing clip across the whole run and cut it where the typing stops and at every pause. Nobody can count keystrokes at normal speed.
• Short entries (a search, a name): place single key sounds, rotating three or four different ones and never the same one twice in a row.
• Give Space and Enter their own sound. They're bigger keys and sound heavier. A typewriter bell on Enter is a fun one if the video suits it.
• Nudge each one a little. Tiny changes in volume (and pitch, if your editor does it) stop a burst sounding like a machine gun.
• Match the speed. Fast typing wants short, light hits. Slow typing can take the fuller sound with a release.

How loud should click sounds be

Under your voice, always. If there's narration, the clicks are seasoning. Set them where you can hear them in a quiet gap and barely notice them while you're talking, then play it back on phone speakers as well as headphones, since your viewers will be on one or the other.

For what it's worth, Framok starts clicks at 80 percent and typing a notch below at 70 percent of each sound's level, and you drag both from there.

The automatic way: Framok

This is the bit I added to Framok, my screen recorder for Windows, on 19 Sep. Framok already logs every click and key while you record, for its automatic zoom, so the editor knows exactly when each one happened. Flip on Add sounds and it lays a click on every mouse press, a key on every key press, a soft tick per scroll notch and a quiet whoosh as the camera zooms in and out.

What's in it:

• Five click styles: Soft, Mouse, Mechanical, Chunky and Wooden.
• Four real keyboards, plus Soft: a Cherry MX Blue (clicky), a brown style board (tactile), a MacBook Pro, and a Hermes Precisa 305 typewriter whose end of line bell plays on Enter.
• Typing that matches your speed. Key presses under 90 ms apart get the shortest, lightest hits and no releases; over 250 ms apart they get the fuller ones. Every hit shifts up to 4 percent in pitch and 1.5 dB in volume, so it never sounds like a machine gun. Space, Enter and Backspace get their own sounds where the keyboard has them.
• On the right frame. The sounds play in the editor as you preview and follow your trims and speed changes. On export they're mixed with your mic and system audio. I checked it by exporting a real 34 second take twice, once plain and once trimmed with 1.5x speed: 57 hits measured each time, none more than a frame off.
• Nothing to credit. Every sound is from a CC0 or royalty free pack, and none of them needs a credit line in your video.
• It doesn't know what you typed. The recorder keeps the timing and the kind of key (letter, Space, Enter and so on), never the characters.

Where Framok doesn't help

• Only recordings made in Framok. The sounds come from the clicks and keys Framok logged while it recorded, so it can't add them to a video from OBS, your phone or the Snipping Tool. For those, use the manual way above.
• Per recording, not per click. You pick one click style, one keyboard and a volume for each. You can't mute a single stray click or add a sound where nothing was pressed.
• You can't load your own sound files yet. It's the built in sets or nothing.
• Windows only. Windows 10 or 11, 64 bit. A Mac version has a waitlist at the bottom of framok.com.
• No free trial on the download from framok.com. It asks for a licence key before it records. A 7 day trial comes with the Microsoft Store version, which isn't out yet, and paid licences have a 14 day money back guarantee.
• The installer isn't code signed yet, so SmartScreen warns the first time. Click More info, then Run anyway.

Framok is $39 once for the life of the product, or $5 a month and cancel any time. Both cover 2 of your own PCs.

See Framok

Do you add click sounds to your demos, or do you think they're too much? Keen to hear which way people lean.

This post first appeared on sundrymill.com.

I build a Windows app that does all of this with one key, so I've spent more time than most poking at these exact settings. The free way comes first, because for the odd call you don't need an app at all.

Why bother

Microsoft's own Teams help says that sharing your entire screen shows "notifications and other activity", and its fix is to turn on Do not disturb (Microsoft Support). That's the chat preview from your partner, the calendar reminder for the interview you haven't told your boss about, and a desktop full of screenshot-04.png files. Same story in a tutorial recording, except it's on YouTube forever.

1. Hide notifications while screen sharing: turn on Do not disturb

The fast way: press Win + N to open the notification centre and click the bell with the zZ on it. That silences notifications until you turn it off again (Microsoft Support).

The settings way: Settings, System, Notifications, Do not disturb.

Two things worth checking in there before a call:

• Priority notifications. Calls, reminders and any apps ticked on the priority list can still get through while Do not disturb is on. If the app that usually embarrasses you is on that list, take it off.
• Turn on do not disturb automatically. Windows can switch it on by itself during certain times of day, when duplicating the display, when playing a game, when using an app in full screen mode, and for the first hour after a feature update. The duplicating rule is for Windows' own Duplicate display mode, the one you use with a projector or a TV. As far as I can tell a Teams, Zoom or Meet share doesn't count as that, so on a video call turn it on yourself.

While you're in Notifications, Hide content when notifications are on lock screen is worth ticking too, if you ever step away mid call.

2. How to hide desktop icons on Windows 11 without deleting them

Right click an empty part of the desktop, choose View, and untick Show desktop icons (Microsoft Support).

Nothing is deleted or moved. The files are still in your Desktop folder, and ticking the same option brings the icons back exactly where they were.

3. How to hide the taskbar on Windows 11

Right click the taskbar, choose Taskbar settings, open Taskbar behaviors and tick Automatically hide the taskbar (Microsoft Support).

This takes the clock, the tray icons and every pinned app off screen with it, which is most of what gives away what else you have open.

The honest catch: it's auto hide, so the taskbar slides back up whenever your mouse touches the bottom edge of the screen. Keep the cursor away from the bottom during a demo. On a second monitor, the same page has Show my taskbar on all displays.

4. Swap a personal wallpaper (optional)

If your wallpaper is a family photo, Settings, Personalization, Background and pick Solid color for the call. Skip this one if your wallpaper is fine to show.

5. Share one window, not the whole screen

This does more than the other four put together. Teams offers to present your entire screen or a single window (Microsoft Support), and Zoom, Google Meet and Discord all let you pick one window or app in a similar way. With one window shared, pop ups from other apps and the rest of your desk never reach the other side.

What still leaks is anything inside that window. In a browser that means your other tabs, the bookmarks bar and whatever the address bar autocompletes. Open the tabs you need in a fresh window and share that one.

For a recording it's the same idea: record a window or a region instead of the full display if your recorder lets you.

The downside: putting it all back

That's four or five switches spread across Settings, the desktop and your call app, and every one needs flipping back afterwards. The one people forget is Do not disturb, and then they miss messages all afternoon. If you share your screen once a month, the checklist above is all you need.

Or do all of it with one key: Qulok

If you share or record most days, this is the bit I built Qulok for. Its Demo mode is the checklist above as one shortcut.

Tick the window you're showing and press Ctrl + Alt + Q. Qulok minimises everything else, hides the desktop icons, sets the taskbar to auto hide, silences notifications and, if you picked one, swaps in a clean wallpaper. Press the same keys again, or the Restore pill in the top right corner, or the tray icon, and everything comes back in the order it was, your own wallpaper included.

A few things that make it better than the manual version:

• Presets. Save a setup (which windows, which wallpaper, which toggles) and bring it back with Ctrl + Alt + 1 to 9.
• Its own buttons stay out of the share. The Restore pill and the timer are marked as excluded from screen capture, the same flag a password manager uses. I checked it with a real screen capture: with protection on, the only pixels in Qulok's gold were the launcher's own (599 of them), against 2,653 with it off.
• Crash safe. Qulok writes down what it changed before it changes anything, so if something dies mid session, the next launch puts your desk back first.
• Lock in, the other tab, is the same engine for focus: keep your work windows, hide the rest, run a timer or a Pomodoro in the corner.

What Qulok doesn't do

Worth knowing before you pay for it:

• It doesn't touch Do not disturb. Windows has no supported switch for it, so Qulok turns off the Windows notification master switch (the one in Settings, System, Notifications) and turns it back on at the end. Windows only rereads that switch when its notification service starts, and before version 0.6.4 that meant Silence notifications did nothing until your next sign in. Since 0.6.4 (26 Sep) Qulok restarts that service when a session starts and ends, which takes under two seconds and needs no admin rights. Notifications already in the notification centre stay there, and one that's on screen at the moment you start may disappear a little early.
• The taskbar still peeks. Qulok flips the same auto hide setting as step 3, so the taskbar slides up if your mouse hits the bottom edge.
• It can't hide browser tabs. Windows can't see inside a browser window. Open the tabs you need in their own window and keep that one.
• It never closes or blocks anything. Windows are minimised and come back. If you bring one back mid session, it stays back.
• Windows only. Windows 10 (version 2004 or later) and Windows 11, 64 bit. A Mac version is on a waitlist at the bottom of qulok.com.
• The installer isn't code signed yet, so SmartScreen warns once. Choose More info, then Run anyway. It's coming to the Microsoft Store at the same prices.

It's $5 a month after a 7 day free trial, or $29 once for life. Both cover 2 of your own PCs. The trial takes your card at checkout and charges nothing for 7 days; cancel before then at qulok.com/manage with your licence key and you pay nothing.

See Qulok

What else do you hide before you hit share? I'm always after one more thing to add to the list.

This post first appeared on sundrymill.com.