Perfectly mixed song sounds bad on mobile devices from Youtube

[bencuri]

This is a song I mixed recently. I am not a professional, but I tried to mix it in a way so I enjoy listening to it, so that it sounds pleasing on my system. I optimized it for my own HiFi, that consists of a pair of 90's KEF passive small towers (about 1m high) and an amp made by my father who is an audiophile and built quality stuff into it. I am listening to it in a big living room.

This is the version that sounds very good on my system:
https://drive.google.com/file/d/1vkRHoR ... sp=sharing

The problem is when I upload this to Youtube and listen to it on mobile devices on their own built in speakers, it does not sound very good. It becomes middle dominated, the secondary instruments in the background like secondary guitars and the marimba comes to the front too much and starts disturbing the rest of the song. And the whole thing is very silent, too. I set it to -14 lufs, but it is very silent compared to most stuff in the genre on youtube. If I add an adaptive compressor and push the volume up, it still sounds bad. It is louder, but the middle becomes too loud as well and the mobile speakers start distorting.

I decided to put an EQ in the Stereo Out channel in Logic, and pulled down the middles. It sounds somewhat better on mobile platforms, but still not too satisfying. But better. However, that version is awful on the HiFi: it sounds terrible, as if I was listening to a tiny 50 year old radio.

Do you have any suggestion how to solve this issue? The problem is: I like the HiFi mix so much that I don't want to change that. It sounds very satisfying to me. If I imagine that I downgrade it with additional mastering for mobile, and then people will load that stuff from youtube into their big HiFi: NOOOO. I cannot let that. Is there any trick to make it sound good on both platforms and keep as much as the original mixing as possible?

Here is another song to compare, from the same genre. It is much louder, and the balance of frequencies is much better. The annoying is, however: if I feed this kind of mixing into the HiFi, it sounds bad. Not just because of the low bitrate (I have the CD version of this, so I know), but because of the way it was mixed.

https://www.youtube.com/watch?v=WZYXS-uBDL4

[alexis]

There are several reasons it could sound different:

1) The speakers on your mobile device may be partially "monoizing" your track because the L/R speakers are close together.

This can make the "middle sound louder"

If you can listen to the YouTube on headphones you can assess how much the mobile device is affecting the sound, subject to #2, below

2) YouTube algorithms might be messing with your mix (compression/ limiting, others) enough to sound quote different

This may be helpful:
https://www.soundonsound.com/techniques ... g-services

3) Your mix may have been constructed to sound perfect by compensating for the room you mixed in possibly not being treated sufficiently - so listening elsewhere may sound quite different

https://www.soundonsound.com/techniques ... oming-room

I think you will likely need to change your mix if you want it to sound better on other systems and in other locations.

These may help also:

https://www.soundonsound.com/techniques ... avoid-them

https://www.soundonsound.com/techniques/mix-mistakes

[The Elf]

It seems that you are up against the fundamental problem of mixing on a playback system that is flattering your audio, likely because it is emphasising low and high frequencies in a typical 'hi-fi' way. When you mix on such a system your mixes will be LF/HF shy, since your ears are being given misleading information.

One way to fix this is to replace your monitoring system with something that is more neutral. Another way is to listen to lots of reference material on your current monitors, such that you become aware of how a commercial mix sounds, and make your mix sound similar. You can also load a reference track alongside your own audio and compare the two as you work on the mix.

Whatever you do as a mixer, once it walks out into the world, you have no control of how your audio is going to be heard. People will play it back on all kinds of super hi-fis, tin boxes, laptops, earbuds, mono and stereo systems... They will have badly set equalisers, poorly placed speakers, mis-wired, broken and abused systems... your mix needs to sound of its best on all of them.

Does 'Bohemian Rhapsody' sound bad because it's being played on laptop speakers? It certainly won't sound the same as playing it on large speakers, or quality headphones. But it will give of its best for the playback system being used. Consumers don't buy different versions of a song mixed for their specific playback system.

So what you are aiming for is a mix created to a 'commercial average' - a mix that will work, of its best, on *anything*, even if that playback medium is not ideal. To do this you need a monitoring system that does not flatter your audio, or you need to become so familiar with your monitoring that you can compensate for it. After that you could look at spectrum analysers, or even consider EQ-matching processors, but unless you can trust (or translate) what you are hearing you will always be on the back foot.

[OneWorld]

WOW, takes me back to the days of the Bhundu Boys, the catchy 'bubbly guitar' lines. Love it.

Sorry I cannot contribute on the production/mixing etc but I listen to a lot of this type of music on YouTube, well all manner of African music, Tiwa Savage, Mose Fan Fan, Charlotte Dipanda, Drumboss, FuseODG, Fatoumata Diawara and many more - and they all play very well on my setup here at home. Here's hoping you can get a result that represents your music at its best.

I agree with one of the points made by Elf - use a system that is flat, and doesn't 'flatter' any particular sound because it could well sound more than acceptable on one system and un-acceptable no another. I found this out when I mixed on headphones, that did flatter the sound, but could never figure out when on the sound set I have in my 'music room' but absolutely woeful in another room a different setup, there was no vitality in the mix, it sounded mushy, bassy, muffled. Then one day I plugged the 'studio' headphones into my living room amp, and even a commercially produced CD sounded as if I was listening with a pillow over my ears. This was despite the headphones being 'studio' as opposed to 'consumer' headphones, and up to that point I had mixed to favour the 'sound' of the headphones, they were supposed to be 'flat' sounding, neutral, but simply didn't sound as if they were neutral, they didn't represent the bass end too well, so I upped the bass (in the studio mix) played the tune on my HiFi, car and there was nothing but bass, too muffled. A change of set up improved the issue quite considerably.

To my surprise, some songs, when online, sounded even better than the mix I had uploaded!

After listening to your song a few times I just listened to Charlotte Dipanda - Sista

https://www.youtube.com/watch?v=w5lVfBZItt0

And the first thing that struck me was the 'purity' of the mix, she sounds as clear as the birds singing in the trees, and as I expected, there is more separation, with the guitars mixed hard left and right, the instruments and voices allowed to breath rather than suffocate.

[MOF]

You could release a track as the single, that was (is?) common, it was for radio play purposes. The album version of that track can be your hi-fi mix.

[Drew Stephenson]

This is the version that sounds very good on my system:
https://drive.google.com/file/d/1vkRHoR ... sp=sharing

Listening here there are two things that jump out to me, first is that the elements in the sides could be quite a bit louder, second is that the kick is overpowered compared to the rest of the kick.

[James Perrett]

My first thought is that the bass end is far too loud and is taking up energy that could be used elsewhere. I would guess that your hifi speakers are rather bass light compared to proper studio monitors and that you've boosted low frequencies that they can't actually reproduce.

Your file also comes out at -16.3 LUFS with a peak of -4.9dBFS so you could easily boost the level by 4dB and possibly more if you were to tame the bass.

I'd disagree a little with Drew about the sounds at the sides - the vocals are important in this style of music and I think the balance is about right although I might change the level of the odd vocal phrase if I was being really fussy and I would spread the backing vocals across the stereo spread rather than have them all in the middle. I might also move the centre guitar to the left as the guitars are a bit unbalanced. The backing vocals also sound a bit dry compared to the lead vocal. I think the mix between vocals and guitar mainly works well in mono.

[The Elf]

I've now had a listen to your audio. You have some heavy bass going on between 40-70Hz that is soaking the energy from your mix. I can well imagine that smaller speakers are going to find that hard to live with. That bass is also going to determine the maximum level of your mix. I would tame that bass to the point where it levels out with the mids.

[Hugh Robjohns]

That all sounds to me like a classic case of bad mix room monitoring and/or acoustics, resulting in poor transportability of the mix due to having embedded compensation for the room.

[PedroTheThird]

There is no way around getting propper studio monitors and at least a minimum of acoustic treatment. You can't mix on hi-fi speakers and expect it to translate well to other speakers and listening conditions. Then there is the fact that smartphone speakers are pretty much the worst speakers you could use to listen to music because of their limited bandwith and high distortion (which pushes mid frequencies even further).

[Sam Spoons]

There is no way around getting propper studio monitors and at least a minimum of acoustic treatment. You can't mix on hi-fi speakers and expect it to translate well to other speakers and listening conditions. Then there is the fact that smartphone speakers are pretty much the worst speakers you could use to listen to music because of their limited bandwith and high distortion (which pushes mid frequencies even further).

^^^

[MarkOne]

There is no way around getting propper studio monitors and at least a minimum of acoustic treatment. You can't mix on hi-fi speakers and expect it to translate well to other speakers and listening conditions. Then there is the fact that smartphone speakers are pretty much the worst speakers you could use to listen to music because of their limited bandwith and high distortion (which pushes mid frequencies even further).

While I totally agree on the installation of some acoustic treatment, I do take issue with the second statement "You can't mix on hi-fi speakers"

ATC, PMI and Focal (three that come to mind, off the top of my head) market reference grade speakers to both studios and well heeled hifi owners, that I'm more than sure would be up to the task of producing a fantastic mix. My own experience of my own KEF LS50s in comparison to Adam and IK near fields, show them to be almost brutally revealing in a mixing situation.

Of course, not all speakers marketed to the hifi world are made equal, so I understand what you are saying. But not as a blanket statement

[MarkOne]

And of course, you can get away with zero treatment if you invest in a good set of headphone

[alexis]

And of course, you can get away with zero treatment if you invest in a good set of headphone

Or a not-that-good set of headphones, but that are Slate VSX?

[tea for two]

https://drive.google.com/file/d/1vkRHoR ... sp=sharing

Hi there Ben.

I imported this wav into logic did a brief reduction handling of bottom also making Kick less prominent, lesser muddied the mids clearing up a bit of bloat there, raising the top a trifle : resulting clearer mix easier to make louder.
You can do this easily in a few minutes in your DAW with its included plugins EQ, Filtering.

Here is another song to compare, from the same genre. It is much louder, and the balance of frequencies is much better.

https://www.youtube.com/watch?v=WZYXS-uBDL4

I couldn't listen on repeat as it just hurts my ears. It's alright in a loud party dance hall I suppose.

One of my references from CD would be this song I heard in the 90s. I have listened on repeat from CD.
Mundeke by the Afrigo Band.
https://m.youtube.com/watch?v=EN-b8Y7_u ... FuZA%3D%3D
I heard it on the ending scene of this filum, I purrchased the soundtrack CD just for Mundeke.
https://m.youtube.com/watch?v=-14BrlXn7FQ

[bencuri]

I agree your reference songs have better quality mixes.

Meanwhile I also created another version, it sounds great and loud on mobile, but it is too harsh for better stuff like headphones:
https://drive.google.com/file/d/1d9BvfT ... sp=sharing

I am trying to arrive to a compromise, reducing the sharp voice. It is just a bit tricky because I EQ-ed it in the Spectrum Filter of Goldwave and I don't find a plugin with similar controls for Logic. A bit difficult to copy the setting that way.

[James Perrett]

I think you are possibly approaching this from the wrong angle. I feel that you have boosted frequencies you can't hear when it is actually better to cut the problem frequencies that are obscuring what you want to hear.

And, as I said before, your overall level is a fair bit lower than a fully mastered track would be. Maybe it is time to find a friendly mastering engineer.

[PedroTheThird]

Hi-fi speakers are designed to make the input sound good (Swallowing little mistakes) while studio monitors are designed to sound honest (Shoving mistakes right in your face). Sure, it's possible to use studio monitors "for fun" if you perceive a honest sound as good but that's a special case.

And of course, you can get away with zero treatment if you invest in a good set of headphone

Unfortunately that's not the case. An example why would be the difference between ILD- and ITD-based stereophony (and mixtures) which rules out entire microphone recording techniques (AB, ORTF, EBS) and even genres for which you need them (like classic). And that's just stereophony, frequency and dynamics perception are another can of worms. Headphone mixes never translate well to speakers no matter how good the headphones are and how well they have been calibrated.

To make things clear: I'm not against using hi-fi speakers nor headphones as secondary systems, quite the opposite. It's good to check things on more (and very different) devices afterwards. But using hi-fi speakers (or headphones) as the main tool never works out (If binaural material is excluded). It introduces plenty of issues you end up spending way too much time fixing in case you later work on the mix with studio monitors. Often enough you even have to start from scratch.

Work efficient and effective. Start with studio monitors in a room with acoustic treatment, then make final adjustments with the help of hi-fi speakers, headphones, the car test and so on. Not only will your mixes translate much better to all kinds of devices but you will finish them faster as well.

[bencuri]

You mean the frequencies over 20K? That is just the rush of work. I forgot cutting the unnecessary off.

The level you hear was set by Youtube. It was originally higher, but I uploaded it to Youtube, to see how it handles it. Sometimes Youtube sets the volume way too silent. In this case I think it is okay. Doesn't reduce that much. I downloaded the file from there again, and that is what you see now. The processed one.

[Sam Spoons]

I doubt very much that anybody on here can hear 20kHz or that most peoples replay system can reproduce it. Harshness is usually considered an excess of energy between 2kHz and 5kHz so two to three octaves below the theoretical top end of human hearing. This corresponds to the 'intelligibility range for speech so is of particular importance for the speakers in mobile phones/devices.

[sc1460]

That all sounds to me like a classic case of bad mix room monitoring and/or acoustics, resulting in poor transportability of the mix due to having embedded compensation for the room.

Agree with Hugh comment, indeed this has always been the biggest problem in translating from studio to tinny mobile phones. In a mixing session I find I absolutely need to move from speaker to speaker very frequently: Studio monitors then check on a budget Bluetooth speaker, and I use Mix2mobile app to stream from DAW to mobile phone, also then listen on headphones. In a way I’m really mixing for multiple devices simultaneously in a treated room, so it’s a complete compromise but I settle for the mix that best plays as possible across them all and not for just one set of speakers.

But here’s the thing about your mix: I just listened to it on my Audeze LCD-X headphones and sure perhaps I would make some changes in instr levels, stereoimaging, EQ, duration, verse/chorus energy levels etc , but when I then listened on my iPad speakers the translation was not as bad as you seem to be making out. I recognised the core substance of the mix that I heard on the audiophile headphones on the crappy iPhone speakers as well.

Yeah I’m sure everyone will have a different opinion on the mix, but for me the translation problem was less to worry about than other possible challenges in the mix.

just my humble two penneth, cheers
H

[bencuri]

That all sounds to me like a classic case of bad mix room monitoring and/or acoustics, resulting in poor transportability of the mix due to having embedded compensation for the room.

But here’s the thing about your mix: I just listened to it on my Audeze LCD-X headphones and sure perhaps I would make some changes in instr levels, stereoimaging, EQ, duration, verse/chorus energy levels etc , but when I then listened on my iPad speakers the translation was not as bad as you seem to be making out.

It is very different on my end. When I listen to this on High End HiFi (not only mine at home but I went to the audiophile store and we tested on 2 systems), it generally sounds good. So it is absolutely not the case that it is my room, HiFi brand, setup that specifically projects it well, but it seems so far that the original mix does sound good on HiFi (at least I am 100% satisfied). On mobile or anything low end, it sounds boring. I don't have active stuff at the moment, have a headphone that is very similar in character to listening to active monitors, it does sound good there. I am going tomorrow to the studio where there is active monitor and passive HiFi paralelly, I will listen there how it sounds on the monitor.

[James Perrett]

You mean the frequencies over 20K? That is just the rush of work. I forgot cutting the unnecessary off.

I probably phrased it badly. What I meant is that you are boosting the sounds that you think are too quiet. Instead, try cutting the sounds that are too loud because these sounds are masking the sounds that you want to hear.

When you go to the studio you should also take along some commercial recordings that sound good to you on your hifi so that you have a known reference to compare things to. I'll bet that commercial recordings don't have as much going on below 100Hz as your recording does.

[bencuri]

I doubt very much that anybody on here can hear 20kHz or that most peoples replay system can reproduce it. Harshness is usually considered an excess of energy between 2kHz and 5kHz so two to three octaves below the theoretical top end of human hearing. This corresponds to the 'ineligibility' range for speech so is of particular importance for the speakers in mobile phones/devices.

For the latest version I linked that I mentioned sounds good on mobile, I used this curve to master it.
https://drive.google.com/file/d/1d9BvfT ... drive_link

The portion that you can see between 3.5K-30K (okay slash off the unnecessary above 20K), is what defines whether it will sound good or not on mobile. When I start bringing that portion lower, it starts to loose the presence on mobile, and becomes more and more boring, as the piercing S-es in the commentator voice are getting reduced, along with the snare. So I found to make this sound good, that portion is very necessary to be at the correct level. The problem is: at the moment for any higher grade speaker this is too harsh. You need to bring it down. But as soon as you bring it down it starts loosing the presence on the mobile. It is a kind of Catch 22.

[James Perrett]

But using hi-fi speakers (or headphones) as the main tool never works out (If binaural material is excluded). It introduces plenty of issues you end up spending way too much time fixing in case you later work on the mix with studio monitors.

I wonder how much experience you have in proper commercial studios? I use speakers that were sold as hifi speakers as my main monitors and have done for over 30 years. Before I owned them, they had been used as main monitors in two other working studios including (apparently) mixing an album by a very well known guitarist. The mixes and masters that I produce on them translate well to all kinds of other systems.

There are no hard and fast rules about hifi speakers and monitors - it totally depends on the exact models of each that you are talking about.