advertisement


Quality of You Tube music videos

Pretty sure, as Tony said earlier, it is down to being able to see the performance.
Just think about concerts that you have been to where, in "hifi" terms, the sound is pretty rough but the overall impression is great :)
I don't think Tony messes with new lossy codecs enough. They are seriously impressive now.
 
I don't think Tony messes with new lossy codecs enough. They are seriously impressive now.

For music I am a vinyl, CD and SACD buyer. I have no subscription service, nor any interest in such things beyond YouTube and iPlayer, though I do have, and hardly use, Amazon Prime (mainly for free same/next day delivery).

These services can sound great though, I thoroughly enjoy the little system I use in the TV room (Audio Synthesis passive pre, Quad 303 and Spendor S3/5R). It certainly sounds more than good enough to gain real enjoyment from whatever I’m watching, explore new music etc, though I’m always pleased when the record or whatever lands and I play it on the main system in the other room.

My view of streaming is it is ‘good enough’ in most respects, but for those of us with good systems who really care about mastering to the point of hunting down and comparing specific versions of a title there are certainly gains to be had. It is the key area where the ‘source first’ rule can not be denied. Mastering > everything else.

That said no way will I knock YouTube. I spend far more time there than I do on terrestrial TV these days and I’m subscribed to around 300 channels spanning music, electronics, vintage computing, politics etc. The sound quality often surprises me as to just how good it is despite the data compression being employed. Many of the music channels, e.g. Boiler Room, Tiny Desk, JHS Show etc can sound stunningly good, but again that is a ‘source first’ thing, and is down to good mic placement, mixing and post-production.
 
I always use yt-dlp -x and it's almost always opus. Make sure you have ffmpeg installed.

Code:
yt-dlp --audio-format best -x (URL)

I also generaly go for 'best'. Yes, I keep the ffmpeg family handy. (Often more than one version in my user space so I can choose one.) What does ffmpeg's 'probe' report about the format you get?
 
I don't think Tony messes with new lossy codecs enough. They are seriously impressive now.

What isn't clear to me is what format of audio was used to *upload* to YT. Having a copy of that should be 'best' as it avoids an extra lossy transcode. But I suspect that isn't what happens when I see much the same list of audio formats listed for a video.
 
Hmmm... The above tripped my 'curiousity' so I had a look at the options for a music video put on YT by the RVWSoc. Yes, one of the audio-only versions is using 'opus' at a higher bitrate - and sample rate - than the 'best' (sic) video (which isn't opus format audio). So I got both and compared them. Most obvious differences are that the opus audio is offset in time by about 0.03 sec from the video's audio. Also 48k sample rate for the best opus when the best video is 44k1.

I'll see if I can find out what was uploaded. Maybe do an aligned comparison, albiet needing to upsample them both I suspect to deal with the clashing rates before alignment.

Has anyone already analysed this?
 
I have found plenty of decent sounding content on YouTube. I do use a very good DAC and if I bypass this the sound is much less enjoyable. Agree with the comment re how the brain works when your both watching and listening so a fair comparison for SQ purposes would be to listen only. What I love about this hobby is the only person you need to please is the listener (you) so if it sounds good to you that’s all that matters….
 
I like this. Can't work out if the sound is exceptional or just matches the video in a very pleasing way. King Crimson: "Starless"

 


advertisement


Back
Top