(mp3) compression integrity

Discussion in 'Mixing & Song Critique' started by audiokid, Jul 2, 2018.

  1. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    I could be right or wrong but at present I don't blame Soundcloud or any mp3 for that matter for swirly transients and spacial corruption (stereo compression artifacts) we hear. I think Soundcloud compression (possible any MP3 compression simply exposes (the bad in a mix) more. I'm more inclined to believe converters, conversion, mix processing and plug-ins are all part of the bad we continuously hear.

    I've done tests for my own learning and never once heard SC change the transients that were locked in the right locations prior to uploading. Once I stopped doing what I found to be exposing this in my mixes, other than just crap MP3 compression sound... it all uploaded and sounded right to me.
     
    kmetal likes this.
  2. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    @pcrecord Marco, hey! you could do a good video on this. Suggestions... Do a mono mix and see what happens before and after you upload it to SC, then go from there. Start spreading a mix, using various plugins and so on. I have a feeling we would learn quite a bit from this.

    Oh I wish I could participate but can't because I sold my converters and don't have a studio to work in now.
     
  3. Kurt Foster

    Kurt Foster Distinguished Member

    Joined:
    Jul 2, 2002
    Location:
    77 Sunset Lane.
    i'm beginning o think you just might be on to something here Chris.

    maybe it's latency within the DAW and adding plugs and tasks for the computer to execute exacerbates the problem?

    tape /console /outboard / 2 mix .... that's all at light speed. start running it through a processor and you have latency. latency has always been an issue for me. i hate it. one reason i hate working with digital audio.
     
  4. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    Indeed... A few years ago I actually started checking mp3 mixes similar to how we check a mix in mono. The mp3 process simply skins out problems making it easier to hear the remainder much the same as a null test can show in a remainder.
     
  5. pcrecord

    pcrecord Quality recording seeker ! Distinguished Member

    Joined:
    Feb 21, 2013
    Location:
    Quebec, Canada
    Home Page:
    Very nice suggestion ! I'll get on to it in a few weeks..
    I'm testing a ICon controler at the moment.. ;)
     
    audiokid likes this.
  6. kmetal

    kmetal Kyle P. Gushue Well-Known Member

    Joined:
    Jul 21, 2009
    Location:
    Boston, Massachusetts
    Home Page:
    this is a (un-mastered) mix from like 2012, of one of my old bands songs. there's some sibalance problems which i'll fix in samplitude if i ever use this mix for anything. besides a slight dulling overall, deoth flattening, and an a seeming couple db of volume, i dont notice any audible artifacts from the MP3 conversion, which i think was done on an iphone app, but my ears and speakers could be lying to me. we only used 5 or six mics on the drums, and stuck them in a fairly dead booth with no parallel surfaces, and no samples. its overall pedestrian mix and song-wise, but im curious if it exhibits mp3 related issues, just out of curiosity. I just uploaded the .wav to soundcloud to compare the two, and see if SC exhibits any swirly phase stuff on it.



     

    Attached Files:

  7. pcrecord

    pcrecord Quality recording seeker ! Distinguished Member

    Joined:
    Feb 21, 2013
    Location:
    Quebec, Canada
    Home Page:
    No this upload seems to be ok..
    I started to prepare the video, so far I've done a whole mix of a song from www.cambridge-mt.com/ms-mtk.htm
    I uploaded 12 versions of the song to soundcloud and they all seems to be ok. mp3, 16/44, 24/96, -11.7 and -7.5 LUFS, -0.48 and -0.0 bDTP.
    My next move is to upload near digital peaking versions but if I can't reproduce the problem, it might be because of Samplitude.. I'll try Sonar next...
    I'm surprise I'm having problems to make it sound bad.
    I'm now searching some posts I heard with the bad sound.
     
    kmetal and audiokid like this.
  8. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    I'm convinced its all to do with bad phase issues ITB. Has nothing to do with SC. Has all to do with artifacts, unstable transients created when processing ITB or OTB round trip processing prior to uploading or compression. I believe the mp3 compression simply relieves the truth rather than creating it. I started testing this a few years back using my 2 DAW system. It exposed bad code in plugins, sloppy work, excessive use of spacial gimmicks found in BS pro audio gear etc. Its a real eye opener.

    Thus... mix in mono to avoid being fooled. Less is more. :) Use compression to test your mixes for phase issues ;)
     
    kmetal and pcrecord like this.
  9. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    I'm convinced the mass of hybrid summing boxes, stereo gear that exclaims they sound wider than clean big rail are nothing more than a BS product strategically shifting the L/R just enough to create a widening effect. And I also think this about plugins. Bottom line. once you know how to do it, you know how they are building it into the algorithms and what to listen for and not to do.

    After I discovered this, I started listening to mastering quality hardware and realized why hardware had to be built so precise and why it cost so much. Then I realized the very best hardware was never as tight as digital. thus... sell the mastering gear and mix-down using 2 DAW's and uncouple the DAW's.

    And the rest is all about reverb which is why I have so much respect for Bricasti. It seems to keep the space in phase. The rest is history.

    The space we are all looking for comes from a good room, good performances in tune with each other that are in a common space all reflecting together like a finely tuned piano (symphony of sound) . This space can come from a digital processor but not just any processor.

    imho

    @pcrecord Go Marco, expose this even more.
     
  10. pcrecord

    pcrecord Quality recording seeker ! Distinguished Member

    Joined:
    Feb 21, 2013
    Location:
    Quebec, Canada
    Home Page:
    Any tip on easy way to produce those artifacts on soundcloud ?
     
  11. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    (edited)
    Good question that Bos and others may be able to help.

    At the time of testing, I had 4 ways to test this using my 2 DAW system. I heard (what I call the truth about gear) with help using the very best (stable) converters and a Dangerous Monitor ST and doing a lot of mix-down's and uploads. I bet I mixed down thousands of tracks of the same sources over the course of this project to prove "most of the gear" I owned could be emulated or simply eliminated through smart mixing ITB.
    The more I learned, the more wanted to specifically prove the theory that "mixing and mastering hardware" really wasn't worth investing in anymore. I was fortunate enough to get most of my mixing and mastering gear bartering and therefore did not fool myself into likely something I had to like because it example: cost $5000.
    Thus.... I didn't fall into the "support of purchase" trap, liking something because I bought it or because just because it had great praise did I believe it was worth keeping.

    All that said,

    I'm not sure what you've done so far but you could try mixing-down a mix in mono, uploading a mono mix to SC then go from there. See if SC changes the actual L/R of the mix. Other than it crushing the bits.... it should sum perfect in mono.

    NOTE: Which is another excellent function of the Dangerous Monitor ST. You can sum and study a mix online with the addition of a 2nd DA.

    It all evolved from there .

    (WARNING) exposing this will piss a lot of people off. :notworthy: You will be up against money and ego. You have my blessings. (y)

    imho
     
  12. kmetal

    kmetal Kyle P. Gushue Well-Known Member

    Joined:
    Jul 21, 2009
    Location:
    Boston, Massachusetts
    Home Page:
    one contributor i think [based on listening to my cousins mixes over the years] is sampled cymbals. take BFD. you get a live recorded sample done at 96k. then they [possibly] dither it down to 44.1, then data compress. then you play the data compressed files back into the daw, recorded at say 44.1 or 96. eq, compress, widen. mixdown to stereo. then use mastering processing on the 2-track, then mixdown to 44.1, use dither. then you transfer it to MP3, or soundcloud.

    in that case theres at least 2 additional compressions and/or conversions involved. i think thats part of it at least.
     
  13. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    Yes, all that would indeed create or change things but I think the real pursuit is to find out what happens to a stereo mix and why it happens. I'm convinced the SC compression is exposing / making it easier for us to hear our mistakes and how we are processing a mix to sound artificially wider, wrong. (thus the swirly end result)

    Example: Want to test a Crane Song STC 8 compressor.... make a perfect mix that sums and uploads perfect on SC then do it again with the STC8 and see what it does to the master once its uploaded. Do this with any analog product and I'm pretty confident you will question its integrity. Then start emulating products ITB and you will likely find out you can do it better, ITB. Hybrid mixing is a complete waste of time and money.
    Then start comparing hardware to software using null tests. Learn what gear and software does.
    I spent time time learning what products are doing rather than just accepting they do something because thats what they say its doing for me. I want to find out why and what that is. Or what I don't want it to do.

    Eventually I came to the conclusion, its smarter to buy a Bricasti over 10's of thousands of dollars in hardware and put that at the end of a (mix or master) and forget about all the hype distracting us. Don't get me wrong... I do believe good gear is the most beautiful thing during tracking but once ITB, other than the exceptional tools... stay ITB. And if you think its widening a mix chances are it won't sum in mono like you'd expect.
    Don't blame it on SC.

    imho
     
    pcrecord likes this.
  14. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    What are some of the converters raved about how they sound wider than most?

    Do Burl or the Antelope ADC really sound great because the chip they are using is extra special, worth the $6000 admission or are they playing with the end of the conversions DA algorithm? I suspect Burl is doing a smart trick with their converters. I don't think they are remotely worth the cost and would love to see how they hold their integrity up in a mono shootout of a good mix.

    But then there are acceptations where a Lavry AD simply just sounds smooth. Why is that?

    I can't help but roll my eyes thinking about the 10M Antelope Audio clock and them charging $5000 for it. I'm thinking about those endorsing that product... check your mix guys and don't be so naive.But this is another aspect of the intertwined reasons we put blame on compression.
    imho
     
  15. kmetal

    kmetal Kyle P. Gushue Well-Known Member

    Joined:
    Jul 21, 2009
    Location:
    Boston, Massachusetts
    Home Page:
    intriguing.

    i skipped the hybrid thing based on watching your post all these years, as you evolved from a fan/practitioner, to calling the bluff.

    since we dont seem to experience this kinda of degradation in an all analog system, it must come down to the conversion. ive never heard this kind of side effect on classic recordings played back on tape and vinyl, or even cd. i invested in the fraunhaufer codecs which may or may not be snake oil, but ill be curious run things thru their paces. if nothing else, at least ill have simplified bath conversion capability.

    the pricing is truly absurd. now i understand a good converter like that takes 10 years of R&D, and wont sell quickly or in large numbers. i dont have enough knowledge of how converters work to know what they;re doing. from what i do know, its not the chip really, its the analog section, stable power, and clocking that does it. i think Burl is catering to niche who loves the tape sound, by essentially mimicking a tape machines electronics, and having transformers built in. i dont think they;re DA side seems to be as acclaimed. i do suspect that much of it come from the digital algorithm and digital side of things, as the fundamentals of good analog electronics are well documented and proven.

    i suspect the days of outboard digital conversion are in fact numbered. besides networked audio, i think we will be seeing digital integration built into the mics and instruments. look how midi drum triggering changed drums... keyboards... its only a matter of time before my guitar pickup sends a digital impulse to an amp sim, and voice synthesis is becoming more and more common. theres really no reason why you pay 5k for a pair of quested's of barefoots, and they dont have digital conversion built in. really i think its a bit much these days, where everything is digital to have to invest in 250-500$ per channel in AD and DA. lol so says the poor musician.

    having experienced the difference in algorithms between DAWs back when i thought i was crazy for it, and hearing the difference between plugins, the antilogarithm seems to be king, not just in audio, but banking, military strategics, basically everything in the digital age of information.

    what a racket. people have experienced significant improvements in those old PTHD 1st gen interfaces, by clocking them to alesis adat machines. i cant imagine antelope is selling many of those things past the 2014's 64 bit, next gen conversion releases.

    it will be interesting when/if marco is able to pinpoint some things that SC is highlighting. it was insane to me the PTLE didnt have pluggin delay compensation built in, until the 64 bit version dropped.
     
  16. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    I believe both analog and conversion is responsible for this shifting but I think digital wins in the end because it has to finally decide what side to put the wrong bit on which is what the compressing of bits shows through process of elimination. You don't need to be a scientist to hear that swirly thing happening. I think we just didn't know it was the processing in our mixes (hybrid or digital) that is doing it in the first place. So, the key is to know what not to do when it comes to the "trade off" of preserving phase vs warming or over processing. What I suspect Ive learned out of all this is to get to the meat with as little processing as possible. Track better, process less and that included both analog and digital processing. It could be that the sample rate still hasn't got to a point digital knows what to do with what it can't convert. I have no idea but look forward to higher and higher sample rates. What I do believe is the swirlying we hear in an MP3 can be avoided with good mixing techniques.

    cool. At first I was blown away with hybrid mixing but as I started fine tuning my system, analyzing everything , which I personally believe I would never have discovered this without the specific mastering console and 2 DAW system.

    The good thing for us all... now that I know this, I trust my finding and believe we can all achieve awesome mixes ITB using a good DAW. I do however believe there are specific outboard products that I still want which include a outboard Bricasti M7 and Pulse Technique Pultecs . http://www.pulsetechniques.com/

    I suspect you are right.
     
    pcrecord and kmetal like this.
  17. kmetal

    kmetal Kyle P. Gushue Well-Known Member

    Joined:
    Jul 21, 2009
    Location:
    Boston, Massachusetts
    Home Page:

    very well put.

    my only question is are you suggesting the pultec is good for mixing, or summing to daw 2? im assuming its excellent for tracking.
     
    audiokid likes this.
  18. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    My final word on this topic: I'm looking forward to others going on from here. I'm not saying I wouldn't want on a nice SSL, Neve or API console. I would love this but only if you gave it to me.
    I also don't believe engineers banging out hit songs on consoles have an advantage either. I don't believe either format has mix advantage. But, how can I know that for certain anyway. I'm just a guy with an opinion that thinks we can achieve great mixes and master music on either format once we know what we are doing.
    I think the guy on the DAW is simply saving a $*^t load of money and most likely able to spin circles around the guys solely using a console. Digital audio mixing and editing is miles ahead of analog mixing. I think the mono button is pretty damn cool and if a process sounds goofy in mono, its best to find out why.
     
    kmetal likes this.
  19. audiokid

    audiokid Chris Staff

    Joined:
    Mar 20, 2000
    Location:
    BC, Canada
    Home Page:
    When I owned a Dangerous Master, of all EQ I used, Pulse Technique Pultec's were incredible for both tracking and mixing. They have a sound and presence that dials in effortlessly. They move a wave so smoothly. I wish I could afford one analog Pultec for every channel. I've tried plugin Pultec's and they were very disappointing.

    My next 2 DAW system may include 2 Pultec's on the second DAW. Not sure yet. I'm also not sure how I'm going to insert them either. Could be a simple summing amp or I may buy another Dangerous Master or something clean like that. Not sure. I sure like them though, they are special.

    This is special, really cool using one Pultec for the middle and the other for the sides. Bricasti on the end of this chain is all I need. No need for the NSEQ-4. ITB works better.

    MEQ-4-NSWQ-4-DM.jpg
     
    pcrecord and kmetal like this.
  20. kmetal

    kmetal Kyle P. Gushue Well-Known Member

    Joined:
    Jul 21, 2009
    Location:
    Boston, Massachusetts
    Home Page:
    not everyone will be able to aquire pultecs, but we all have the mono button!
     
    audiokid likes this.

Share This Page

  1. 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.
    Dismiss Notice