Now all that’s left is Spotify to release lossless music streaming so that all major streaming platforms have the option. As already announced, Spotify HiFi will be launched in some markets later this year, with the same promise as the other services: to offer the sound experience the way the artist imagined it, as if he were in the studio.
But is this catchphrase of the lossless really what it says it is? Do I need to pay more to get the full experience? Is current streaming bad? calm down, the Techblog spoke with audio experts to understand the real importance of high quality music in streaming services. Let me go ahead: it doesn’t make any difference, unless you have your ear trained and the equipment pricks the galaxies.
Tidal, Deezer, Apple Music, Spotify and Amazon Music
Let’s remember that the lossless project on streaming platforms started with Tidal in 2015, announced by rapper Jay-Z. The launch campaign also had the participation of other important artists to highlight the streaming and differential that would bring the studio’s quality to the subscriber’s ears.
In 2015, Deezer also adopted the slogan it sustains to this day. To the Techblog, Marcos Swarowsky, general director of Deezer in South America, highlighted that “it is a differentiated experience that takes the listener to the world of the highest sound quality. It’s like listening inside the recording studio.”
Last month, Apple announced without fanfare or without a film production the addition of the lossless catalog to Apple Music and released the alternative to users without raising the price, unlike the previous ones. With that, we have the following scenario:
- Tidal: charges R$16.90 for the normal plan and R$33.80 for the HiFi (lossless) plan;
- Deezer: charges R$16.90 for the regular plan and R$26.90 for the Deezer HiFi;
- Apple Music: charges R$ 16.90 for subscription, no more for lossless.
Even though Spotify hasn’t joined in, we know that the HiFi plan will have an additional cost. Amazon Music — another streaming available in Brazil, but without the lossless mode here — charged US$ 15 for the HD (lossless) plan and US$ 10 for the regular plan, but after Apple’s announcement, it merged the two plans at the lower price .
O Techblog asked Deezer if there is any intention to merge the Premium and HiFi plans, but the company has not commented.
But, after all, what is a lossless song?
Basically, it is the existing audio quality on a CD, which brought (was) files in WAV format (of WAVEform or waveform). Today, streaming services are trying to do the opposite of going back to the origins rather than reducing data consumption. To understand, it is necessary to go back in time and who helps along this path is audio specialist and co-founder of Kuba, Leonardo Drummond, from the channel Mind the Headphone, on Youtube.
Leonardo explains to Techblog that, in vinyl, there is the movement of the needle going through the grooves. This movement is analogous to variations in air pressure that generate sound, which generates electrical signals reproduced later by the speakers, so it is an analog medium. The popularization of digital audio starts with the CD, according to him:
Digital audio is a more efficient way to store this information, but instead of doing it through physical grooves that deteriorate over time, are subject to dust and have technical and physical limitations, the CD does this through data digital, in a process called sampling.
During music recording, the microphone sends electrical signals to a digital-to-analog converter (the DAC) that measures the electrical wave to create a digital representation of this variation. Thus, there are two characteristics that affect this signal: the sampling rate and the bit depth.
The sampling rate is how many measurements of that signal are taken per second and bit depth is how many bits of information there are in each sample. Then you create the audio of a CD, for example, which has the standard 44.1 kHz at 16 bits. It means that every second 44,000 measurements were taken of this original signal and each one of these samples has 16 possible values.
According to the Nyquist–Shannon theorem, in order to be able to sample and reconstitute a given frequency range, you will need the sampling rate to be twice the highest frequency. We humans listen from 20 Hz to 20 kHz, so for me to perfectly sample and reconstitute this range, I’m going to need a sampling rate of 40 kHz.
According to the co-founder of Kuba, higher frequencies are useful for studios, because you have a larger area to work. Otherwise, for auditing, the benefits are “debatable, to say the least”.
Bit depth is related to the signal-to-noise of a recording, a process known as quantization. Leonardo helps explain:
Think of an analog wave embedded in a grid. The number of vertical lines is the sampling rate, the number of measurements taken per second. The number of horizontal lines is the number of bits, the possible values. But there will be situations where the analog wave will pass between two possible values. So I can have a value of 3 or 4 and the analog wave will pass at 3.5, it is necessary to round this value.
This rounding generates noise, distortion. And what you do, in this process, is add a random noise that is underneath, which is the quantization noise. It is inherent in any digital recording.
There are three common bit-depth values: 8, 16, or 24. This number determines the “possible distance between the loudest sounds and the noise floor”. But, still according to Leonardo, there’s nothing to worry about, because it’s a noise lower than that of a needle traversing a vinyl.
On a CD, the theoretical signal-to-noise ratio is 96 decibels. With some advanced techniques, it reaches 120 decibels. The point is that for you to hear the background noise of CD audio, you’ll have to play music at the volume of a chainsaw in your ear.
That said, it can be concluded that the quality existing on a CD is enough for any human being, considering the audible frequency range.
So I want lossless, not MP3
Calm down, it’s not there. Today, even though it is a lossy compression format, MP3 is already much more developed and different from that popular in the 2000s, when the internet speed was slow to download lossless files or the device’s storage space for playback was small.
To get an idea, an MP3 song at 256 kb/s can have about 6 MB, while in lossless format it can reach 36 MB with the same three minutes (considering 48 kHz at 24 bits). If you consider that there are songs of five minutes or more, such files can take up a lot of storage when this is an issue.
And there are several compressions for MP3. Spotify itself starts with files from 24 kb/s up to 320 kb/s. The difference in these compressions is how much of the original audio will be lost. According to Leonardo, what MP3 does is cut, in some possible levels, the information that the human being is less likely to hear, in order to reduce the file.
Imagine that in the case of a WAV, saying it’s 44,100 measurements per second and each measurement is 16 bits. If you do the math, it’s a pretty big file. MP3 limits this amount of bits, so it will say “look, the maximum this file will have is 128 kb/s”. This compression is kind of flattening that signal from the extremes.
Alexandro Azevedo, director of Audio-Technica in Brazil, gives examples of the results of these compressions:
In the beginning there was this very different compression rate, so the quality was very affected, that is, the audio was affected to the point where you could see that the voice was muffled, the bass was not consistent, the sound was more artificial.
However, both Leonardo and Alexandro agree that users are already well served with a 320 kb/s file.
In the current formats, FLAC, ALAC, AAC, if you work from 320 kb/s upwards, you have very good quality. Anything that manages to work at 320 kb/s will affect the sound quality much less.
MP3 uses very sophisticated algorithms to dispense with information that is not audible. Of course this bit rate will affect that. A 128 kb/s file is very easy to distinguish from a WAV file, for example. So if you take a lay person and put a 128 kb/s audio and 320 kb/s audio in front of him, there are people who won’t notice the difference, but there are many lay people who will notice, if you tell them “it’s okay pay attention to this and that”.
And when is the amount of kilobits per second important?
Just during audio transfer. Thinking back to the 2000 internet, a file of 70MB or more could take days to download, while a 3MB file would take a few hours.
Bringing it to current times, this concern meets the consumption of mobile internet bandwidth, thinking of those who have more limited packages to listen to. Spending 50MB or 70MB of the monthly allowance on a song isn’t all that smart.
What about Wi-Fi? It could stream without compromising the operator’s data package. Here we come to the point, if MP3 compressions deliver quality at 320 kb/s, why do unnecessary streaming?
Lossless for what (or who)?
Theoretically, with MP3, inaudible portions of the raw material are removed so that the user has a smaller file, but with the quality he could want, considering the compression at 320 kb/s.
I don’t think it’s impossible for people to tell the difference between MP3 and lossless, but I think there are very few people and very sophisticated systems. I have a phone that costs R$ 8 thousand here in Brazil, I’ve had a speaker system at home for R$ 80 thousand, I’ve never heard any difference between anything above an MP3 at 320 kb/s. I never listened.
For me, this lossless thing is marketing. Apple’s own case, AirPods Max does not accept lossless files, because the codecs it has for Bluetooth are AAC and SBC. It doesn’t have aptX HD, it doesn’t have LDAC, it will only transmit capped files. Same case for Spotify HiFi and Tidal, it’s marketing. These are companies wanting to have more expensive plans so that people pay a little more for a difference.
Leonardo Drummond, co-founder of Kuba.
And now another X enters the equation: Bluetooth
The codecs Leonardo mentioned are the “file translators” supported by Bluetooth technology. AAC tends to be the evolution of MP3. aptX HD is a Qualcomm proprietary codec that theoretically supports up to 576 kb/s. LDAC is Sony’s codec that, according to the brand, allows the transmission of high resolution files (Hi-Res) at a maximum of 990 kb/s over Bluetooth.
In the case of Apple Music, for example, none of these codecs with higher transfer rates are supported by the iPhone, iPad or Mac. Or when they are, in other devices, it is also up to the phone to have the support for the technologies.
But what percentage of people who subscribe to such streaming services have these headphones? In the case of aptX HD and LDAC, third parties would need to pay licenses to Qualcomm and Sony.
And in this topic of Bluetooth, one must also consider the path that companies took to save internal space and smartphone components: remove the P2 input for headphones, the only analog input that was available and allowed to listen to lossless audio via cable , no workarounds.
How many people have enabled headphones, excluding Apple’s (still), with aptX HD or LDAC to try to hear any difference?
Alexandro Azevedo, director of Audio-Technica in Brazil, tells the Techblog that it is better to bet on cable if you want to take advantage of the lossless. “The lossless format is frustrating when it comes to Bluetooth. Bluetooth is practical, but it is a very volatile transmission system, due to the distance and the significant drop in quality. You can’t have lossless because it doesn’t support such a high, lossless transfer rate,” he explained.
Attention should be turned to another point.
Some say they see the difference between listening on Spotify, Apple Music, Deezer HiFi and Tidal, but this statement needs to be made carefully. Comparing streaming from different platforms without paying attention to some points is a mistake.
For Leonardo, “high resolution audio tackles the wrong problem. The problem with low quality is not because the resolution is low, CD audio is sufficient for listening, the problem is the quality of the recordings.”
“The addition of lossless to Spotify, to Apple Music, is marketing to me. It won’t make a difference for 99.99% of people, most who say they hear a difference don’t, if you put them in a blind test, they won’t pass”, he continued.
And why the difference between one and the other?
The difference may be related to the settings of each application. Spotify, for example, is installed with the volume normalization option enabled, unlike Tidal, “that means it [Tidal] it’s not forcing the recordings to have a similar volume,” said Leonardo. Therefore, it gives the impression that the sound is louder on the second platform.
Another highlight is regarding mastering. Apple Music has Apple Digital Masters, Tidal has MQA (Master Quality Authenticated). On Tidal’s support page, it is quoted as “a way to compress digital music without limitations to deliver guaranteed master quality sound”.
Current streaming offers more than enough quality at maximum quality considering the files. One can go to Tidal or Apple Music because these streams have different masters, then it’s all right because it’s focusing on the real problem, mixing and mastering. But it’s important to say that it’s not the files making that difference.
Blind test to compare is indisputable
Alexandro Azevedo, from Audio-Technica, argues that comparison is the best way to explain the difference between a compressed and a lossless file:
Many times I would do a demo, for example an MP3 at 96 kb/s and do the same demo as an MP3 at 320 kb/s and then finally another demo with WAV or FLAC or any other full and uncompressed format. And then the person “wow, what a difference”. From 96 kb/s to 320 kb/s it is blatant. The best way I see you explaining this to a layman is exactly with the demo. That would be the perfect scenario.
Leonardo Drummond complements with the idea that the user needs to do the blind test, so as not to be fooled by “lossless” stamps or other trinkets of the application. “Only if you look and see writing that is of higher quality, you think you are hearing a difference”, he reinforced.
There are many people saying that there is a lot of difference [entre os streamings]. Then you will see FLAC and it is an extraction of a vinyl, vinyl has a different sound because it is another mix. To find out if you hear a difference or not, you have to do the blind test and more than that: if you downloaded a FLAC, convert it to MP3 at 320 kb/s, using a good converter with a good codec and then the person will discard the possibility of being different mixes.
Observation: FLAC is a lossless audio compression codec. They are bigger than MP3 and smaller than WAV, which would be the raw.
NPR (National Public Radio, a US non-profit organization) published an article, in 2015, when Tidal started to make noise, to address this lossless and MP3 difference. In one excerpt, it is quoted that “when it comes to audio quality, file size is not everything”. In the publication, there is a test with six songs, each song with three versions. The listener/reader needs to identify which one is lossless.
Leonardo Drummond has also conducted a blind test on his channel, the Mind the Headphone. If you want to try it (remember that you need to use wired headphones):
What to focus attention and investment on?
A good headset. That was the answer given by the two.
We have a question of cost-benefit versus purchasing power, which also includes the cultural aspect, which is the following: we give little value to the importance of audio in our lives. Any box is fine, any phone is fine, why am I going to spend R$1,500 on a phone if there is a phone for R$200.
But audio is part of our lives 24 hours a day, 7 days a week and for our entire life. It’s important to take care of your hearing aid, and good equipment helps with this.
Alexandro Azevedo, Director of Audio-Technica in Brazil.
What about spatial audio?
Despite considering it “cool and interesting”, Leonardo Drummond clarifies that he has few things available in this format, besides being necessary to have a headphone with surround support. Deezer, Apple Music and Tidal offer spatial audio in some songs, on specific equipment.
About choosing among the available music streaming platforms, the co-founder of Kuba recommends focusing on the one that best suits the user’s needs, whether in interface, integration with other products he uses, curation for recommendations or masters that most please the listener. “Anything else is getting scabies to scratch yourself, which probably won’t make a difference. The focus is music”, he highlighted.
None of the experts sign a lossless plan. And you, also prefer to save?