The wav file size depends on the sample rate, bit depth, length, and the number of channels is also important. It doesn't matter how the file is created, the size only depends on these factors. At least as far as I know.
Took this Spotify vs lossless audio test yesterday with an outcome that was exactly as expected: I am unable to separate the sheep from the goats so to speak.
Disclaimer:
I took the test with my HD6XX plugged directly into my computer as I am awaiting my new amp. And yes, I know the sound card on my computer is, well, crap.
The result:
I hit the head on the nail on a mere 56% of the tests - including correct answers by coincidence.
What is more, on a more general level I heard no or only minimal discernible difference in my perception of the actual quality of the music samples.
My conclusion:
I am not saying that someone with a better hearing, more resolving gear or golden ears wouldn't benefit from lossless over lossy. I did not perceive the lossless music samples as 'better' than the 320 mp3. At least not by a margin that gives reason to favour Tidal over Spotify for me when inserting the much better user interface, more extensive catalogue (to my taste), Spotify Connect etc. into the equation.
Weird how Musepack never took of as there codec of choice, Since HA's latest blind test shows it being transparent on anything at 192kbps. I can still hear weak pre echo on some stuff at V0/320 while V2 ~ V4 is ugly.
Spotify uses Ogg Vorbis. Not too many streaming services are exclusively mp3 now. I know Deezer and Amazon Music have lower quality tiers that are using it, but most others have been using AAC, which offers a higher coding efficiency with both stationary and transient signals due to the variations in blocksize usage and a simpler filter bank.
Spotify uses Ogg Vorbis. Not too many streaming services are exclusively mp3 now. I know Deezer and Amazon Music have lower quality tiers that are using it, but most others have been using AAC, which offers a higher coding efficiency with both stationary and transient signals due to the variations in blocksize usage and a simpler filter bank.
Is there public knowledge regarding what's the transparency's threshold for Ogg Vorbis? As far as I'm concerned, on Spotify I can't discern any difference from 192Kbps (with a pair of Sennheiser HD800s and another half dozen headphones I tried).
Is there public knowledge regarding what's the transparency's threshold for Ogg Vorbis? As far as I'm concerned, on Spotify I can't discern any difference from 192Kbps (with a pair of Sennheiser HD800s and another half dozen headphones I tried).
At q9 level for Spotify's higher quality tier (pre-lossless option), I would expect it to be transparent to the lossless original. Not much testing has been done since 2014, but up to then I didn't see any results that concluded with anyone identifying a difference at the higher bitrates. Other than an occasional issue with pre-echo from instruments such as castanets, even the lower quality level q5 (160 kbit/s) is often cited as being transparent from the lossless version.
Back in 2000 when Christopher Montgomery and the Xiph team first began work on the codec, they had already identified a few potential deficiencies with MP3 and were working toward providing a solution to them.
Hydrogenaudio did a 192kbps multi codec test, Where Vorbis was 4,9. At 160kbps Q5 it still sounded better than QAAC/Lame on few samples that both give up on, so I don't by AAC being 4,94 for a minor gripe.
I would think that with music, if you put it at 320 VBR, and it could go above that, it would be because of super audible frequencies you can't even hear.
Only If the track averages 216kbps at a 160kbps target. AAC/Ogg are still MDCT codecs even at 320 ~ 500kbps something could fail. I stopped using Vorbis ages ago when some samples would still show the puffy wind artifact and to stop It needed 384kbps VBR, When AAC fails it sounds like MP3 but uglier.
With Musepack It doesn't have those issues, Heck It wasn't made for under 192kbps yet It outperforms both AAC/Ogg at 160kbps VBR. Would've been a wipe out if MPC was the popular one, Since It took till 2012 for AAC/Ogg/MP3 to come close to MPC @ 192.
These days any device with Android/ios & PC can play MPC with a 3rd party app. Since xHE AAC & Opus are too busy touting 96kbps in 2021, Opus performs even worse with minimal electronic ambient than the other 3 by having the bit rate reach 260kbps while MP3 is transparent at V5.
I'm lucky. For me, AAC at 160kbps is transparent. I've done dozens of ABX comparisons, with many different headphones, many different styles of music. I cannot tell FLAC from 160 AAC. I'm 54 though, and my very high end hearing is gone....so that probably explains it.
I'm glad I don't hear the difference, as I can fit a lot more music on my computer and DAP's. And I have a LOT of music.
High frequencies with AAC can be an issue below 192, but above that it's only a case of momentary artifacting. I only found one recording that required more than 192. It's pretty rare, but audiences applauding and massed strings with lots of upper frequency sheen are the hardest to encode.
High frequencies with AAC can be an issue below 192, but above that it's only a case of momentary artifacting. I only found one recording that required more than 192. It's pretty rare, but audiences applauding and massed strings with lots of upper frequency sheen are the hardest to encode.
I record relaxation sounds, like thunder showers. So far, the most life-like results have come from recording to DAT at 16/44.1. Any copying to non-lossless makes the showers sound like white noise.
I wonder, how 'bad' (or good) is DAT longplay? (32KHz 12bit). That bitrate is also used by those super small Sony NT cassettes.
I can't find much about it since it's pre-internet technology.
I wonder, how 'bad' (or good) is DAT longplay? (32KHz 12bit). That bitrate is also used by those super small Sony NT cassettes.
I can't find much about it since it's pre-internet technology.
I wonder, how 'bad' (or good) is DAT longplay? (32KHz 12bit). That bitrate is also used by those super small Sony NT cassettes. I can't find much about it since it's pre-internet technology.
I record relaxation sounds, like thunder showers. So far, the most life-like results have come from recording to DAT at 16/44.1. Any copying to non-lossless makes the showers sound like white noise.
If that is convenient enough, that works. But one of the advantages of AAC is that it is able to dynamically adjust data rate beyond 320 with VBR. So if you encoded your rain sounds at AAC 320 VBR, it would probably increase the data rate to the point where it could accurately reproduce the sound. I encode early 20th century 78s with tons of tiny clicks and those do very well with AAC, not so good with Frauenhofer.
I suppose you don't have a huge library of different rain sounds, so it probably doesn't matter though. You can just do those as FLAC and encode music as high data rate lossy.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
