 |
Welcome to the Piano World Piano Forums
Over 3 million posts about pianos, digital pianos, and all types of keyboard instruments.
Over 100,000 members from around the world.
Join the World's Largest Community of Piano Lovers
(it's free)
It's Fun to Play the Piano ... Please Pass It On!
|
|
|
|
67 members (amc252, brennbaer, accordeur, antune, anotherscott, AndyOnThePiano2, benkeys, 9 invisible),
1,790
guests, and
317
robots. |
|
Key:
Admin,
Global Mod,
Mod
|
|
|
|
Joined: Jan 2010
Posts: 1,643
1000 Post Club Member
|
1000 Post Club Member
Joined: Jan 2010
Posts: 1,643 |
How's it sound is everything. Compare yours to one's you think were done well.
re: 44.1 vs 48, it's interesting that YouTube suggests 44.1 for video. But if that's what they're using you just need to decide who should do the sample rate conversion... You or YouTube.
|
|
|
|
|
Joined: Aug 2010
Posts: 188
Full Member
|
|
OP
Full Member
Joined: Aug 2010
Posts: 188 |
How's it sound is everything. Compare yours to one's you think were done well.
re: 44.1 vs 48, it's interesting that YouTube suggests 44.1 for video. But if that's what they're using you just need to decide who should do the sample rate conversion... You or YouTube. This is weird, because when you look at this page: https://support.google.com/youtube/answer/1722171?hl=en... it says, under the "Audio codec" section, that the recommended sample rate is 96khz or 48khz. What should one do? I mean this is conflicting information it seems. Can it really lead to some kind of disaster if I upload 48 instead of 44.1?
Yamaha Clavinova CLP-645 Polished Ebony
Korg Triton Studio 61
Korg Pa800
|
|
|
|
|
Joined: Jan 2010
Posts: 1,643
1000 Post Club Member
|
|
1000 Post Club Member
Joined: Jan 2010
Posts: 1,643 |
No, no disaster. LOL YouTube's software is simply going to do a sample rate conversion if you upload in 48k, 96k, 192k or basically something other than 44.1k.
The only question is, is their conversion algorithm better than the one you're using? Try it both ways, now test to see if you can hear the difference. If you personally can't, it matters little.
|
|
|
|
|
Joined: Feb 2010
Posts: 5,870
5000 Post Club Member
|
|
5000 Post Club Member
Joined: Feb 2010
Posts: 5,870 |
I'm always using Amplify instead of Normalize so i dont know about Normalize. But from the manuals I understand that Normalize does an extra DC offset removal before plus the amplify.
Does your signal have a significant offset?
|
|
|
|
|
Joined: Aug 2010
Posts: 188
Full Member
|
|
OP
Full Member
Joined: Aug 2010
Posts: 188 |
Now I'm reading all over Google that ALL YouTube videos get sampled down to 44.1, even if you upload 48 or 96. This is now starting to irritate me a bit, lol.
Wouldn't the safest way be to upload 44.1 then?
Yamaha Clavinova CLP-645 Polished Ebony
Korg Triton Studio 61
Korg Pa800
|
|
|
|
|
Joined: Jan 2010
Posts: 1,643
1000 Post Club Member
|
|
1000 Post Club Member
Joined: Jan 2010
Posts: 1,643 |
Two schools of thought on this. You can avoid sample rate and but depth conversion if you record in the delivery format throughout your project. But some people feel if you start with a better format some of that must come across when you dither down and convert the sample rate.
As usual, what I'm trying to elude to, how it sounds to you is more important than numbers.
|
|
|
|
|
Joined: Aug 2010
Posts: 188
Full Member
|
|
OP
Full Member
Joined: Aug 2010
Posts: 188 |
Two schools of thought on this. You can avoid sample rate and but depth conversion if you record in the delivery format throughout your project. But some people feel if you start with a better format some of that must come across when you dither down and convert the sample rate. Well, I record in 96 khz with 32-bit float setting in Audacity, and then I export to various other sample rates with Audacity. I always keep the original .aup (Audacity format) files, and also keep the highest quality export, 96 khz 24-bit. As usual, what I'm trying to elude to, how it sounds to you is more important than numbers. Heh, it sounds fine to me prior to uploading. But I want to make sure it stays that way after I upload it. My original intention was to upload 48 khz 24-bit, but I'm trying to figure out what's the point if YouTube will really downsample it to 44.1 khz? I know some quality will always be lost with YouTube, but I'm trying to keep that to a minimum, and that's why I ponder whether uploading 44.1 khz is the way to go. I still can't seem to find a definitive answer whether YouTube really does this or not. I'd also hate for the downsampling on YouTube's part to lead to some kind of sync issues. Anyway, the most recent related forum post I could find is from February 2015: https://productforums.google.com/forum/#!topic/youtube/GXGn6Cac7_c
Last edited by Stephano; 07/23/16 11:44 PM.
Yamaha Clavinova CLP-645 Polished Ebony
Korg Triton Studio 61
Korg Pa800
|
|
|
|
|
Joined: Jan 2010
Posts: 1,643
1000 Post Club Member
|
|
1000 Post Club Member
Joined: Jan 2010
Posts: 1,643 |
Sounds like YouTube dithers down to 44.1. So who's software sounds better dithering down? Audacity or YouTube?
|
|
|
|
|
Joined: Sep 2011
Posts: 3,756
3000 Post Club Member
|
|
3000 Post Club Member
Joined: Sep 2011
Posts: 3,756 |
Sounds like YouTube dithers down to 44.1. So who's software sounds better dithering down? Audacity or YouTube? ElmerJFudd, I think I know what 'dither' is, but what is 'dithering down' (apart from being a Status Quo song)?
Roland HP 302 / Samson Graphite 49 / Akai EWI
Reaper / Native Instruments K9 ult / ESQL MOR2 Symph Orchestra & Choirs / Lucato & Parravicini , trumpets & saxes / Garritan CFX lite / Production Voices C7 & Steinway D compact
Focusrite Saffire 24 / W7, i7 4770, 16GB / MXL V67g / Yamaha HS7s / HD598
|
|
|
|
|
Joined: Jan 2010
Posts: 1,643
1000 Post Club Member
|
|
1000 Post Club Member
Joined: Jan 2010
Posts: 1,643 |
Just an industry expression to go from a higher bit depth to a lower. Here is pretty straight forward explanation google revealed, http://www.darkroommastering.com/blog/dithering-explainedIt also explains the problems that arise when you go from higher to lower (ie. 24bit to 16bit) and that there are different algorithms you can choose from in most DAW software - not sure if it is offered in Audacity which the OP is using. Regardless, the point is that it's nice to have 24bit or better recordings at 48k or better but in the end we are forced to dither down and convert to lower sample rates for delivery. It seemed briefly with digital downloads (after people's connection speeds to the internet had become adequate) that there was interest in better quality audio. But now that the tech industry is pushing streaming over downloads, we are back once again to lesser audio quality. As far as audio quality goes, we were better off with CDs - but we are being pushed toward quantity of recordings vs. quality and the average listener seems to prefer this because they can listen to what they want rather than what they own. Of course we shouldn't confuse audio quality with music quality - if the writing and performance are lousy no amount of bit depth or sample rate is going to improve that! 
|
|
|
|
|
Joined: Feb 2010
Posts: 5,870
5000 Post Club Member
|
|
5000 Post Club Member
Joined: Feb 2010
Posts: 5,870 |
16 bit and 48kHz seem good enough for a final rendering of your performance. The 24 bits 96kHz quality is mainly for the recording master and editing phases. Probably only very few people with top quality equipment and ears may have use for more than 96dB snr.
The mp3 compression bothers me more than the 48kHz restriction.
FAIK all video web services, also eg vimeo, RE-compress both audio and your video. So you upload already compressed audio, which youtube then de-compresses, and then re-compresses again witha slightly different format. This is the recipe to degrate your audio quality...
Youtube audio has problems with stutter anyway and is not a good platform for high quzlity audio. Vimeo seems slightly better (no stutter).
|
|
|
|
|
Joined: Sep 2011
Posts: 3,756
3000 Post Club Member
|
|
3000 Post Club Member
Joined: Sep 2011
Posts: 3,756 |
Just an industry expression to go from a higher bit depth to a lower. Here is pretty straight forward explanation google revealed, http://www.darkroommastering.com/blog/dithering-explainedIt also explains the problems that arise when you go from higher to lower (ie. 24bit to 16bit) and that there are different algorithms you can choose from in most DAW software - not sure if it is offered in Audacity which the OP is using. Regardless, the point is that it's nice to have 24bit or better recordings at 48k or better but in the end we are forced to dither down and convert to lower sample rates for delivery. It seemed briefly with digital downloads (after people's connection speeds to the internet had become adequate) that there was interest in better quality audio. But now that the tech industry is pushing streaming over downloads, we are back once again to lesser audio quality. As far as audio quality goes, we were better off with CDs - but we are being pushed toward quantity of recordings vs. quality and the average listener seems to prefer this because they can listen to what they want rather than what they own. Of course we shouldn't confuse audio quality with music quality - if the writing and performance are lousy no amount of bit depth or sample rate is going to improve that! Thanks for the explanation, and I agree with you. It's just that dither (at least as I understand it) is a serious problem as it introduces instability into the timing, albeit in tiny resolution levels. I suppose I was wondering if going from high bit rates to lower can cause more either than sticking at 44/16 to begin with.
Last edited by toddy; 07/24/16 10:34 AM.
Roland HP 302 / Samson Graphite 49 / Akai EWI
Reaper / Native Instruments K9 ult / ESQL MOR2 Symph Orchestra & Choirs / Lucato & Parravicini , trumpets & saxes / Garritan CFX lite / Production Voices C7 & Steinway D compact
Focusrite Saffire 24 / W7, i7 4770, 16GB / MXL V67g / Yamaha HS7s / HD598
|
|
|
|
|
Joined: Jan 2010
Posts: 1,643
1000 Post Club Member
|
|
1000 Post Club Member
Joined: Jan 2010
Posts: 1,643 |
Agreed, the reason Google/Youtube is suggesting that you upload 24bit/44.1 WAV recordings is because they are going to reduce file size with compression, and if you start with a lossy format it's in effect going to get "re-lossied" - if that's even a term. The YouTube conversion algorithms will dither down your bit depth but if you upload 44.1k your sample rate will stay in tact, avoiding having them do a sample rate conversion.
I think the stuttering you are referring to is simply that YouTube needs to buffer on your computer to play smoothly, and since YouTube is so heavily used (and your ISP may not be giving YouTube preference on their network) sometimes the service just fails to buffer adequately. But YouTube has the largest audience, so that's the trade off.
|
|
|
|
|
Joined: Aug 2010
Posts: 188
Full Member
|
|
OP
Full Member
Joined: Aug 2010
Posts: 188 |
So to avoid issues, uploading 44.1 khz 24-bit on YouTube would be the safest way to do it?
Last edited by Stephano; 07/24/16 10:48 AM.
Yamaha Clavinova CLP-645 Polished Ebony
Korg Triton Studio 61
Korg Pa800
|
|
|
|
|
Joined: Jan 2010
Posts: 1,643
1000 Post Club Member
|
|
1000 Post Club Member
Joined: Jan 2010
Posts: 1,643 |
Yeah, this is a decent approach. You can also try doing the export in 16bit/44.1 in Audacity and compare after upload. Compare is important. You have to happy with the FINAL results, how it sounds in the delivery format (which in this case will be compressed YouTube video).
|
|
|
|
|
Joined: Feb 2010
Posts: 5,870
5000 Post Club Member
|
|
5000 Post Club Member
Joined: Feb 2010
Posts: 5,870 |
I think the stuttering you are referring to is simply that YouTube needs to buffer on your computer to play smoothly, and since YouTube is so heavily used (and your ISP may not be giving YouTube preference on their network) sometimes the service just fails to buffer adequately. But YouTube has the largest audience, so that's the trade off.
If anything, a bigger company should have a less fluctuating load than the smaller companies (general rule of statistics). So I assume it's a bug somewhere.
|
|
|
|
|
Joined: Aug 2009
Posts: 1,623
1000 Post Club Member
|
|
1000 Post Club Member
Joined: Aug 2009
Posts: 1,623 |
Please consider that the peak level in your recording very likely only happens one time. Using that single peak as a benchmark for 0.00db is a good practice. For other more dense mixes, that are compressed, there can be long stretches that are at peak. Soundcloud and youtube as well as personal mp3 conversions do exhibit problems in those sections that aren't evident when listening to the original file.
For a straight piano solo recording wave that is not further subjected to a vst type compression, using the peak that happens 1 or maybe 2 times in your recording is a good practice.
Some engineers do recommend just keeping a max of -0.1db as a standard.
As many here have said, does it sound good?
If you can post 0.00db with no problems, that is the best you can get.
|
|
|
|
|
Joined: Jan 2010
Posts: 1,643
1000 Post Club Member
|
|
1000 Post Club Member
Joined: Jan 2010
Posts: 1,643 |
I think the stuttering you are referring to is simply that YouTube needs to buffer on your computer to play smoothly, and since YouTube is so heavily used (and your ISP may not be giving YouTube preference on their network) sometimes the service just fails to buffer adequately. But YouTube has the largest audience, so that's the trade off.
If anything, a bigger company should have a less fluctuating load than the smaller companies (general rule of statistics). So I assume it's a bug somewhere. Actually, it's YouTube's popularity that has ISP's throttling them. YouTube and other popular video services like Netflix are competing with your providers for your viewing time and bandwidth.
|
|
|
|
|
Joined: Aug 2010
Posts: 188
Full Member
|
|
OP
Full Member
Joined: Aug 2010
Posts: 188 |
As I'm preparing to upload my first finished recording, there's something I "think" I noticed.
Is it at all possible that an audio file in 44100 Hz 24-bit is just a very (almost unnoticable) tiny bit louder than the exact same audio in 16-bit?
Maybe I'm imagining things? I'm talking about the exact same recording. Both files were created from the same Audacity file, with no difference between them whatsoever.
It was normalized at -1.0, but now I'm thinking if -0.1 would be the better solution, as it would be as loud as possible and I wouldn't have any red clipping warnings. Either way, the 24-bit vs. 16-bit loudness really intrigues me.
Last edited by Stephano; 07/27/16 11:12 PM.
Yamaha Clavinova CLP-645 Polished Ebony
Korg Triton Studio 61
Korg Pa800
|
|
|
|
|
Joined: Aug 2009
Posts: 1,623
1000 Post Club Member
|
|
1000 Post Club Member
Joined: Aug 2009
Posts: 1,623 |
24 bit does sound better than 16 bit, there's 8 more bits of sound at any given moment.
Can you post the 96/24 track too?
BTW believe it or not - if your going to sample rate convert to 44.1 kHz, your tracking sample rate is suggested to be set to 88.2khz.
|
|
|
|
Forums43
Topics223,390
Posts3,349,244
Members111,632
| |
Most Online15,252
Mar 21st, 2010
|
|
|
|
|
|
|
![[Linked Image]](https://www.pianoworld.com/Uploads/files/RedDotMedalSmall.jpg)