Are you logged in with cookies in YT-DLP/JDownloader though?

Here's the bitrate distribution for format 251-17 (Original upload language (EN), Opus) for "What's new in Ardour 9.0?"

Pretty wide distribution, from <60kbps all the way up to >250 kbps. But again, only if logged in.


Here's the command I used:

Code:

yt-dlp --cookies cookies.txt --no-mtime --remote-components ejs:github -f 251-17 "https://www.youtube.com/watch?v=DEFFplnYbxA"
python bitrate.py "What's New in Ardour 9 0 [DEFFplnYbxA].webm"

Here's the Python script (bitrate.py) I wrote because I couldn't find a bitrate distribution plotter:

Code:

import sys, subprocess

args = "ffprobe -v error -select_streams a:0 -show_packets -show_entries packet=size,duration_time -of csv=p=0".split()
data = subprocess.run(args + sys.argv[1:], capture_output=True)
# print(data.stdout)

packet = [x.split(",".encode("utf-8")) for x in data.stdout.splitlines()[1:-1]]
packet = [[float(x[0]), float(x[1]) * 8] for x in packet]
packet = [x for x in packet if x[0] > 0 and x[1] > 0]
frames = [x[1] / (x[0] * 1000) for x in packet]
# print(frames)

import plotext, os

plotext.hist(frames, os.get_terminal_size().columns)
plotext.theme("dark")
plotext.xlabel("Bitrate (kbps)")
plotext.show()

Needs ffmpeg (ffprobe), plotext for Python, and cookies from a throwaway Google/YT account.

Spoiler: Older version

Code:

import sys
import subprocess

args = "ffprobe -v error -select_streams a:0 -show_packets -show_entries packet=size -of csv=p=0".split()
data = subprocess.run(args + sys.argv[1:], capture_output=True)
# print(data.stdout)

frame = 1 / (1000 * 0.02)
packets = [round((int(x) * 8) * frame) for x in data.stdout.splitlines()[1:-1]]
# print(packets)

import plotext

plotext.hist(packets, 64)
plotext.theme("dark")
plotext.xlabel("Bitrate (kbps)")
plotext.show()

 

Last edited: Feb 7, 2026

However that is what we get from yt, filtered or not, like it or not, and we have to deal with it.

The spectrograms you see are made with Fakinthefunk, and if you resample an audio from 44 to 48kHz it will flag it, it means that the capability of codec bitrate is not fully exploited. Try yourself with known files, re-encoding them.

The examples I downloaded are not randomly taken, but from an official artist source, to be sure they are "professionally mastered", (or I hope so).
As you can see not one file is flagged as fake, video or audio (and it happens a lot with random yt clips/audio).
So the bitrate codecs are consistent with audio, no matter how low the bitrate itself is.
However I'll download clips from random users to verify this, I want to be sure.

 

No. I'm not logged in.
You mean that if logged in yt (I don't know about JD cookies) you get better quality?

 

I can't even download Opus format unless I'm logged in.

Here's how many formats get advertised depending on whether I'm logged in or not:

Code:

yt-dlp -F "https://www.youtube.com/watch?v=DEFFplnYbxA" | wc -l
153

yt-dlp --cookies cookies.txt -F "https://www.youtube.com/watch?v=DEFFplnYbxA" | wc -l
218

That's only advertised, doesn't mean I can download all of those (I can, if I'm logged in).

 

Here yes, (128kb) but I'm not logged in (see the list available).

 

I don't use JDL enough to know what it's actually downloading/how to get it to show more formats. Are you sure that's constant bitrate and not "target bitrate"?

I just downloaded "128kbps Opus" in JDL and this is what the bitrate distribution shows for me:

 

I don't think there is any conversion, downloads are very quick, also for videos (it would take quite a bit to convert, at least for video).
Also if you see the mediainfo logs above posted it states constant bitrate for all audio files (.aac, .opus, video embedded or not).
If you use JD you can download all formats at once, just click Add additional variants.
I'll send you the video link in PM because this matter is becoming interesting.

 

So instead of choosing between one loss-less and another loss-less format, you choose one flavor of crap (opus) over another (mp3). Makes a lot of sense...

Loss-y == artifacting and less music bits period. The format is inconsequential.

 

Last edited: Feb 7, 2026

I basically use YouTurd to watch old movies, 50's-60's movies late at night on my laptop, so the sound isn't very good. the few clips posted here the audio is usually ok. I don't really care which platform is better if I only listen to something once, I'm not saving it to listen to it daily for the rest of my life, if the audio is clean and clear I'm ok with it. Just my opinion so who cares.

 

Last edited: Feb 8, 2026

You should at least read carefully, first.

And BTW, to be in topics, we're discussing YT audio quality, that much is what we get.

 

Last edited: Feb 7, 2026

YT <> quality. At least for audio anyway. If you want quality audio you go somewhere else. It's not on YT.

 

The quality of spotify "FLAC" is no better sounding than the MP3 stuff. Wherein it is highly suspect that if spotify does offer "FLAC" it's simply encoded MP3 crap in a FLAC format.

Qobuz STREAMS true HD audio 24bit/96K, 44K/16bit, up to 24bit/192K etc... Which means you can record straight off the Qobuz stream to whatever medium you want to store. If one buys tunes from Qobuz it's in HD WAV, FLAC, DSD, DXD format. Higher res than was ever available on CD. It doesn't get any better for just a slightly higher price point @ month than spotify. Tidal has a limited selection comparted to Qobuz who has more stuff in HD or at least CD quality that any I have seen anywhere else. Some even complain about the Qobuz selection saying that it's not a wide as spotify - but personally a wider selection of crap sounding audio is not what I am after in the first place.

 

@xorome found that yt constant 128kb opus is actually a variable encoding, though mediainfo states constant.
And was extracted from an official mastered artist channel, to be sure is not casual vinyl/tape ripped audio.

 

In your histogram the y axis is the number of frames, right?

 

Yep.

 

Upencoding leaves artifacts does it not? Especially if the source is substandard? I don't hear that coming from there personally. What would finally be the point in doing that - the resulting audio would definitely be subpar.

Something that does happen (I have experienced it myself) in the case of recordings that are rare or where master tapes are unavailable, wherein a vinyl recording is the only source available, the higher resolution streaming versions of those recordings might be very high res rips from vinyl. They are not upsampled but simply captured at high resolution and sourced for streaming. Apparently in the case of Qobuz at least, they do nothing to manipulate audio in that sense. What they get for streaming is what the original record companies and studios give them.

 

Last edited: Feb 8, 2026

MP3 encoding

- Assume the highest possible bit rate and the filters you need.
MP3 mainly has difficulties with the high frequencies.
Don't waste bandwidth with 32 kHz instead of 44.1 kHz. Your song will sound better.

- Don't overcompress everything with a compressor/limiter.
Leave some of the dynamics of the song so that the encoding algorithm has something to work with.

- Set the encoder to “Best possible quality”. This will allow you to get the best possible results. It takes longer, but it's worth it.
- Remember: MP3 encoding makes the resulting material slightly hotter than the original mix in most cases

Limit the output level of the material intended as MP3 to - 1 dB, instead of the usual -0.1 dB or - 0.2 dB. Avoid digital overloads (overs).

- A high-quality MP3 encoder is LAME.
- 160 kbs: The lowest bit rate that is acceptable for files
- 320 kbs: The highest quality, but shows the largest files, but is hardly distinguishable from a CD.