Audio Source: Start Recognition does not work

[alexandremix]

Just installed this, tried to use recognition mode and i dont see it start for no reason.

I even tried https with browser , it asked permission on chrome correctly but then nothing starts at all.

[AnshulJ999]

Hey @alexandremix !

Thanks for opening this issue. I saw your Reddit comment as well. I'm not entirely clear what's causing this issue, but you have tried both HTTP and HTTPS URLs?

http://localhost:9012

and

http://localhost:9013

And accepted the self-signed cert warning on HTTPS?

If you are on localhost only, I recommend using the HTTP link since that is still secure. For LAN or remote access, HTTPS is required.

If you're still facing issues, please do the following:

1. Let me know how you have installed SyncLyrics, some details of your machine, and the exact links you are using to access SyncLyrics.

2. Backend logs from SyncLyrics. Wherever you installed it, there will be a /logs folder with 'debug' logs.

3. Let me know your Chrome version too, and please try a different browser once.

More frontend logs from the browser's DevTools would also be helpful.

That can allow me to investigate further and see if I can fix this on my end.

Thanks!

[alexandremix]

Yo, Ive been trying this for 1 hour now. still no luck now downloading version 1.3 to see if it works. i even tried the env variable to that recognition website.

ive tried https with browser. ive tried directly , the green button never changes

ran this in cursor and it said to me:

Root Cause
The application starts in "--reaper mode" (line 38 in logs):
Audio recognition started (--reaper mode)
But the reaper.py module fails to initialize and keeps reporting:
system_utils.reaper - WARNING - reaper.py:44 - Audio recognition not available
This blocks ALL audio recognition (both backend system audio AND frontend microphone), which is why:
The buttons appear to do nothing (they actually do send requests)
The WebSocket connects but immediately disconnects
The status stays at "--"
This Is a Beta Bug
The audio recognition engine (shazamio or the underlying fingerprinting library) failed to initialize in this v2.1.0-beta Windows build. The capture system works (it detects your Voicemeeter device), but the actual recognition engine doesn't start.

[alexandremix]

also runnning as admin. always. disabled firewall and antivirus

[alexandremix]

version 1.3 worked ! the green button changes! to red and shows stop recognition! awesome

[AnshulJ999]

Wait version 1.3? I'm more confused now haha.

Please use version 2.0.5. That was the last stable version.

v2.1.0 is beta and may have issues, which is why I didn't mark it as stable.

But if you're having issues on 2.0.5 and 1.3 is working for you, that's very strange haha. The rest of the app has come a long way since then.

[alexandremix]

Ive tried all new versions beta and non beta, can confirm on windows the latest version to work with audio recognition with microphone is version 1.8

[AnshulJ999]

Hey @alexandremix !

Thanks for testing everything, and sorry you had to go through all this.

It's still quite a strange issue that I can't make sense of, but if v1.8 is working well for you and you'd like to stick to it no issues.

If you'd like me to investigate this further, please share some logs and the other details from my initial comment, then I will look into it.

[alexandremix]

I had to work yesterday and today . But I'm willing to help discover this . This is a great tool that can be useful

[AnshulJ999]

Hey @alexandremix ! Thanks a lot!

If you are able to help, please test the latest v2.0.5 release and attach the following info:

1. After trying the Audio Recognition feature a few times in the same session, please attach the 'debug.log' and 'app.log' from the logs folder. This should be in the same folder where you ran the EXE from. You can also just zip up the whole logs folder and send it here.

2. To confirm, you're using SyncLyrics EXE on Windows? I need the:

a) Windows version
b) Browser and version (Chrome or another, and its version)
c) The exact link you used to access SyncLyrics. Like localhost or LAN IP or another.
d) Some details of your machine and which audio device you are trying to use that kept failing. SyncLyrics supports both 'frontend' and 'backend' devices; have you tried both?

For clarity, backend means a Mic/Loopback attached to the server, and frontend means the browser mic on the client. If you use SyncLyrics on the same machine, it might be the same device, but using 'backend' is better since it does not require a websocket.

With all this, I can investigate this issue further to see if I can pinpoint the problem. In my own tests I haven't been able to recreate this, and I confirmed from other users that it's working for them too.

Thanks!

[alexandremix]

ip: http://100.84.168.58:9012/

but it still happens with localhost 9012

clicking start recognition never goes red. always green. does not matter if its browser or via the exe.

does not matter the browser. brave chrome edge etc .

Im using voicemetter but i did also just try the direct connection with microphone with voicemetter disabled.

debug.log
synclyrics.log

[AnshulJ999]

Hey @alexandremix !

Thanks so much for sharing the logs, they were super helpful. This allowed me to pinpoint the issue and fix it.

Please try the new v2.1.1-beta version whenever convenient for you. I'm hoping this one works for you.

If you still face issues, let me know and please attach the new logs so I can continue debugging.

Cheers :D

[alexandremix]

Hey @alexandremix !

Thanks so much for sharing the logs, they were super helpful. This allowed me to pinpoint the issue and fix it.

Please try the new v2.1.1-beta version whenever convenient for you. I'm hoping this one works for you.

If you still face issues, let me know and please attach the new logs so I can continue debugging.

Cheers :D

can confirm is working.

more questions: why is the audio level seems like its updating each 5 seconds for eg (not like instant for eg 15 fps or something)

also: when the music was fetched trough shazam or music recognition methods the controlls are there + the progress bar 0.50 / 0:00

it does not have the progress bar working.

also, would it be possible to fetch a music manually? write its name? to get lyrics? and for eg click on lyrics manually on the "timing" when as a workaround if for any reason the recognition is not working?

ill share with you a custom CSS that ive implemented so only the current lyrics appear for a VJ setup on a Giant ledwall
like festival / clubing:

/* 1. IMPORT & GLOBAL SETUP */
@import url('https://fonts.googleapis.com/css2?family=Anton&display=swap');

body {
background: #000000 !important;
overflow: hidden;
margin: 0;
display: flex;
align-items: center;
justify-content: center;
height: 100vh;
width: 100vw;
}

/* 2. HIDE ALL COVER ART & BACKGROUND LAYERS */
#album-art,
#album-art-link,
.album-art,
#background-layer,
.background-layer,
#art-zoom-img-a,
#art-zoom-img-b,
#next-up-art,
.next-up-art,
#background-overlay {
display: none !important;
background-image: none !important;
background: transparent !important;
opacity: 0 !important;
}

/* 3. CLEAN THE UI (Remove all side elements) */
#track-header,
#bottom-nav,
.top-right-controls,
#provider-info,
#player-ui-container,
.progress-container,
.provider-info,
#next-up-card,
#audio-source-modal,
#settings-panel {
display: none !important;
}

/* 4. LYRIC TYPOGRAPHY (VJ MODE) */
.lyric-line:not(#current) {
display: none !important;
}

#current {
font-family: 'Anton', sans-serif !important;
text-transform: uppercase;
font-size: 8vw !important;
color: #ffffff !important;
text-align: center;
letter-spacing: 2px;

width: 95vw; 
max-width: none !important;
white-space: normal; 
line-height: 1.1;

position: absolute;
top: 50%;
left: 50%;
transform: translate(-50%, -50%);
z-index: 9999;

text-shadow: 
    0 0 10px rgba(255, 255, 255, 0.8),
    -3px 0 #ff00ea, 
    3px 0 #00ffff; 

animation: lyric-glow 2s infinite ease-in-out;

}

/* 5. INSTRUMENTAL & ERROR BLACKOUT /
/ This ensures the screen is black during "♪" or "Lyrics not found" /
#current:empty,
#current[data-content="♪"],
#current[title*="♪"],
#current[data-content*="not found"],
#current[data-content*="Not Found"],
#current[data-content*="no lyrics"],
#current[data-content*="No lyrics"],
#current[title*="not found"] {
display: none !important;
opacity: 0 !important;
visibility: hidden !important;
font-size: 0 !important;
text-shadow: none !important;
}

/* 6. ANALOG VJ TEXTURE (SCANLINES) */
body::after {
content: " ";
display: block;
position: absolute;
top: 0;
left: 0;
bottom: 0;
right: 0;
background: linear-gradient(rgba(18, 16, 16, 0) 50%, rgba(0, 0, 0, 0.25) 50%),
linear-gradient(90deg, rgba(255, 0, 0, 0.04), rgba(0, 255, 0, 0.01), rgba(0, 0, 255, 0.04));
z-index: 10000;
background-size: 100% 4px, 3px 100%;
pointer-events: none;
}

/* 7. ANIMATION (GLOW ONLY) */
@Keyframes lyric-glow {
0% { filter: brightness(1) blur(0px); opacity: 0.9; }
50% { filter: brightness(1.4) blur(0.5px); opacity: 1; }
100% { filter: brightness(1) blur(0px); opacity: 0.9; }
}

[alexandremix]

If you like we could discuss further some other ideas / software integrations, let me know if theres a way i can directly reach out

[AnshulJ999]

Hey @alexandremix !

Glad to hear it's working now!

So the audio level meter feeling slow: if you use a backend source, it updates only once per cycle, and the cycle time is 5-6 seconds. If you use a frontend microphone, then it updates continuously and looks like a real VU meter. This is just a limitation of how the audio is recorded and can't be fixed. Backend source takes the RMS value of a full 5-6 second clip and sends that to the frontend.

Shazam does not reliably return song duration, which makes the progress bar behave a bit awkwardly at times. Like it knows the current position of the track, but not the duration, hence the 0:00 you saw.

Manual lyrics finding is not supported right now, and probably not planned either. It'd require a lot of extra work to build a lyrics browser that lets you search all providers simultaneously and then select a song. There's already other apps that cover this use case, so I'm not sure it's worth it.

The CSS you shared looks very cool!

Feel free to share any other ideas you have here! If you want to talk privately, you can DM me on Reddit or email me too. Email: anshulaj999@gmail.com

Thanks!