#1
|
|||
|
|||
Sound quality comparison between Youtube and Vimeo
Hi everyone,
I am about to release a new project, and I would like to upload the content to Vimeo. However, I am not yet convinced by the audio quality. In order to compare this with Youtube I uploaded the same video file to both Youtube and Vimeo. On the devices that I tried, the sound of the Vimeo video seems distorted, especially the higher notes. I contacted the Vimeo support about this. Although they are very helpful and answer quite rapidly, they cannot reproduce the issue, which makes it difficult for them to help me. So I am wondering if it is just me who feels that the Vimeo sound quality is bad, or if this is also experienced by others. I am wondering about your opinion on this. What do you think about the quality difference between the videos below? Do you hear distortion on the higher notes? Video on Youtube Video on Vimeo Your opinion would be very much appreciated! |
#2
|
|||
|
|||
I had a listen to both streams on a Mac>RME>Bryston>Dynaudio system and didn't detect unusual distortion on either platform. I did prefer the Vimeo, by a small margin, detecting a difference in definition and overtones.
All platforms add their secret sauce and compression to material that doesn't fit their individual requirements. It pays to do some experimenting with uploading different master renders at varying bit rates/sample rates and loudness levels. Vimeo Audio Delivery recommendations: Audio guidelines Channels: Stereo For best results, we recommend 2-channel Stereo audio. For any number of channels above 2, we will downmix your audio to 2 channels. Codec: AAC-LC (Advanced Audio Codec)For best results, we recommend using AAC-LC (low complexity) for the audio codec. Data rate: 320 kbit/s For best results, encode your audio at constant rate of 320 kbit/s. Sample rate: 48 kHz For best results, set your audio sample rate to 48 kHz. If your working setting is already less than or equal to 48 kHz, leave it as is. For any audio uploaded with a sampling rate over 48kHz, we will resample your audio to 48 kHz or below. YouTube Audio Delivery Recommendations: Encoding specifications for music videos The features described in this article are available only to partners who use YouTube's Content ID matching system. The media files you deliver to YouTube for music videos must conform to these specifications. Audio Profile Attribute Specification Codec FLAC Linear PCM Sample rate 44.1kHz recommended. Higher sample rates are accepted but not required (for example, 48kHz or 96kHz). Bit depth 24-bit recommended, 16-bit acceptable Channels 2 (stereo) Although it is not recommended, YouTube accepts compressed audio. YouTube transcodes from the delivered format; audio quality is much better when transcoded from a lossless format compared to re-compressing a lossy audio format. If you must deliver compressed audio, use these specifications: Codec: AAC-LC Sample Rate: 44.1Khz Bit Rate: 320kbps or higher for 2 channels (higher is always better; 256 kbps acceptable) Channels: 2 (stereo) |
#3
|
||||
|
||||
Those are interesting audio requirements. I'd guess the majority of video editing/creation apps ("average" folks use) are not rendering video with audio streams in either of those formats, if you can even specify the audio format, or working with audio that was even initially recorded in a non-lossy for fixed bitrate format.
I think the point is that the closer the D/A'd audio is to the original, the more likely that their [extreme] processing of it will not do even more damage.
__________________
"I know in the morning that it's gonna be good, when I stick out my elbows and they don't bump wood." - Bill Kirchen |
#4
|
||||
|
||||
opps duplicate
__________________
Enjoy the Journey.... Kev... KevWind at Soundcloud KevWind at YouYube https://www.youtube.com/playlist?lis...EZxkPKyieOTgRD System : Studio system Avid Carbon interface , PT Ultimate 2023.12 -Mid 2020 iMac 27" 3.8GHz 8-core i7 10th Gen ,, Ventura 13.2.1 Mobile MBP M1 Pro , PT Ultimate 2023.12 Sonoma 14.4 Last edited by KevWind; 01-26-2022 at 09:14 AM. Reason: Dup |
#5
|
||||
|
||||
Quote:
What "devices" did you try ? Often on my old laptop speakers I would get high frequency distortion that was not present in the earbuds on that laptop or my studio system I don't know about vimeo but I believe most of the pro's upload 24 bit 48kHz WAV audio to Youtube on a Video file that is 1080 or 4 k I'll listen on my laptop and my studio system and report back
__________________
Enjoy the Journey.... Kev... KevWind at Soundcloud KevWind at YouYube https://www.youtube.com/playlist?lis...EZxkPKyieOTgRD System : Studio system Avid Carbon interface , PT Ultimate 2023.12 -Mid 2020 iMac 27" 3.8GHz 8-core i7 10th Gen ,, Ventura 13.2.1 Mobile MBP M1 Pro , PT Ultimate 2023.12 Sonoma 14.4 Last edited by KevWind; 01-26-2022 at 09:19 AM. |
#6
|
||||
|
||||
I just spent ~$50 to buy Apple's compressor app, which I've always wanted anyway... It gives you the ability to create outputs with whatever you want for video and audio quality. Here's 3 outputs - one built-in and 2 custom ones (using the previously spec'd audio streams). You can see the difference in file size. And FWIW, the audio stream in the Apple Devices "Best Quality" (built-in) option is AAC 160kbps, and yes, that's the *stereo* stream.
For the test FCP project, I could have used 720p, which I probably did on the upload way back when (didn't keep that file), because the video was cropped. I'm not going to use Apple built-in outputs anymore, though. p.s. It says Compressor requires 11.5 or later Mac OS. It does connect to FCP so you can send your projects there for batch processing to multiple outputs (as I did in this example) and you don't need to keep FCP open once you've done that. Handy. p.p.s. This video is just around 3min.
__________________
"I know in the morning that it's gonna be good, when I stick out my elbows and they don't bump wood." - Bill Kirchen Last edited by keith.rogers; 01-26-2022 at 12:44 PM. |
#7
|
||||
|
||||
Quote:
You Tube seems to like 44.1 instead of 48 for some reason, but I doubt that makes much difference compared to the other mangling they do. For my videos, I usually use 48kHz, 24 bit. The video editors I use (Final Cut, Premiere, After Effects), will all let you specify the audio file quality and other aspects in fine detail. And I also often use the approach of loading the final movie into Logic, synched with the original audio mix. So I end up exporting the Logic full-quality audio rather than trusting the video encoder.
__________________
Music: Spotify, Bandcamp Videos: You Tube Channel Books: Hymns for Fingerstyle Guitar (std tuning), Christmas Carols for Fingerstyle Guitar (std tuning), A DADGAD Christmas, Alternate Tunings book Online Course: Alternate Tunings for Fingerstyle Guitar |
#8
|
|||
|
|||
Thank you for your input so far!
Quote:
To me the Vimeo version does for examply not sound like a slighly suboptimal recording, or a hypothetical difference between 44.1 and 48 kHz. Instead, the amount of distortion that I hear sounds like if something went wrong during mastering, like exceeding the 0 dB level. It makes me hesitating as I would very much like to release my recordings, but I am afraid that more listeners might face the distortion issue, so perhaps I should wait for it to be solved. Last week I did some experimenting with uploading different file formats and bit rates, but I did not experiment with multiple loudness levels, I will give that I try. Thank you for the suggestion. Quote:
Changing formats and bit rates did not make a difference unfortunately. I am very curious about your findings when you have had a listen on your studio system! |
#9
|
||||
|
||||
I also hear no significant difference between the two. I would not be able to distinguish between the 2 in a blind test. I checked the waveform coming from vimeo, and there's no visible clipping. Not sure what in your playback path could cause vimeo to sound drastically different for you, but something would seem to be going on. Have you tried different browsers? Listened on a device instead of computer? Tried a different computer?
__________________
Music: Spotify, Bandcamp Videos: You Tube Channel Books: Hymns for Fingerstyle Guitar (std tuning), Christmas Carols for Fingerstyle Guitar (std tuning), A DADGAD Christmas, Alternate Tunings book Online Course: Alternate Tunings for Fingerstyle Guitar |
#10
|
|||
|
|||
Thank you Doug! (Although it is not related to this topic, I have enjoyed many of your music and demo videos over the past years.)
Interesting that you also did not hear a significant difference between the two videos. Maybe I should just release my recordings as I am starting to think that it is not a general issue. Quote:
I don't know the details about how platforms like Vimeo spread their files over the world. Probbably most of you are located in the US whereas I am located in Europe, I am wondering if it might be a regional issue related to different versions in the cloud for example. I will have a chat with the Vimeo support about this theory. |
#11
|
||||
|
||||
That's an interesting thought. Seems quite possible! This is a downside to online streaming, we really have no control, and also little insight into what's going on behind the scenes.
__________________
Music: Spotify, Bandcamp Videos: You Tube Channel Books: Hymns for Fingerstyle Guitar (std tuning), Christmas Carols for Fingerstyle Guitar (std tuning), A DADGAD Christmas, Alternate Tunings book Online Course: Alternate Tunings for Fingerstyle Guitar |
#12
|
||||
|
||||
Quote:
I recorded the streamed audio off both videos and there is definitely a very slight bit of distortion in what I'm getting from Vimeo. If you open the original video in something like Handbrake, you'll see the embedded stereo type and bitrate (or equivalent), and just importing/opening the video in something like Audacity will show you if it has any peaks over 0dBFS. p.s. this is not an equivalent example, but it shows the suggested formats uploaded to the two services. This is the original project, which was rendered/uploaded as 720p with the poor 160kbps AAC audio, but, alas, it doesn't really benefit from better quality rendering . Still, I don't think there's as much audible difference as I heard on OP's samples. Also, some waveform stats for the audio in the two files, which have the same video settings, but different in the 48kHz/16-bit audio (FLAC vs. AAC 320kbps constant bitrate, and the images are in that order). YouTube video with FLAC audio Vimeo video with 320kbps AAC audio p.p.s. This link has slightly different specs for audio content in video uploads to YouTube. (It does not mention the "content matching" program, FWIW.) https://support.google.com/youtube/a...ppy=%2Cbitrate Some curious is the recommendation for 384kbps (compressed/lossy) stereo. That's not a common bitrate you'll see offered, as 320kbps is typically the highest. I am sure *I* cannot hear the difference between 320 and 384kbps!
__________________
"I know in the morning that it's gonna be good, when I stick out my elbows and they don't bump wood." - Bill Kirchen Last edited by keith.rogers; 01-27-2022 at 12:50 PM. Reason: add YouTube spec link. |
#13
|
|||
|
|||
Thank you for the input Keith!
In my opinion your video on both Youtube and Vimeo sounds great! I would be very happy if I could obtain a similar sound quality. Actually I have not yet looked into statistics like the ones that you posted, but that is definitely something I will do. I am installing Handbrake at the moment. For now I would like to share a recording of my computer audio, which is how I hear the sound. The fragment of about 6 seconds is played twice. The first one is Youtube, the second one is Vimeo: Youtube-Vimeo.flac Last edited by Martijn; 01-26-2022 at 05:43 PM. |
#14
|
|||
|
|||
Even though it doesn't seem to apply in this case, everyone should be aware of the recommended average Loudness Units Full Scale (LUFS) If you audio track is too loud, various platforms will do different things to you audio, and usually not in a good way. Rather than using a compressor (and this is different from file compression to make a smaller file) or a soft limiter, most platforms will simply do a hard clip to whatever their respective specs are.
Ian Shepherd, a mastering engineer who is the host of The Mastering Show podcast, has a website where you can do temporary uploads to determine how you file will fare on the different platforms. Here's the link: https://www.loudnesspenalty.com
__________________
'47 000-18 '49 00-17 '91ish Deering Tele prototype '02 Goodall GC '20 Gibson Southern Jumbo Deering Maple Blossom '62 Danectro Longhorn Bass UAD Apollo x8p, Apollo Twin Genelec 8351B's Studio Monitors Genelec 7370A Sub Lauten Audio LT-386 |
#15
|
||||
|
||||
I had a listen when you first posted and could consistently detect a harshness in the Vimeo version that was not present in the Youtube one.
This is still the case in your recent sample as you hear them, but strangely I can no longer hear this in the original. I could not distinguish between them even with the expectation that I would still hear it the same. I am in France, but maybe more than just geographical differences... |
|
Tags |
sound quality, vimeo, youtube |
|