Jan 27, 2018607 views

Can you hear a difference?

Preamble: I am not looking to start yet another thread full of people calling each other idiots for liking OR disliking various levels of encoding/compression. To me the following comparison is incredibly apparent to MY EARS in MY SYSTEM and I genuinely want to see who does or does not feel the same way. This won't be for everyone because the tracks involved do cost money (one subscription the other a purchase) and no it's not a placebo controlled, double-blind case study. I am posting for fun... not science!
Personal Stance on High Resolution Audio... and other audio tidbits.
Over the past 15 years I've enjoyed a reasonable amount of "high resolution" music (PCM bit rates of 48kHz-192kHz and up to 24 bit depth as well as DSD's various formats). I don't buy a ton of it but I've been buying "better encoding" since the early days of HDCD (first major consumer format with encoding greater than 44.1kHz 16 bit aka Redbook). I am and have always been... on a budget. I buy gear for the long haul and am not afraid to spend a bit more than the average purchase for flexibility and long term viability.
I've purchased physical media such as DVD-Audio stereo and multi-channel, SACD in all its various incarnations, DTS Audio Discs, Blu-ray Audio Discs. For the most part I have generally noticed a difference but I am certainly not going to stand up on a soapbox and say it's for everybody. In addition to the physical side of digitally stored media I can also recall connecting to FTP servers that one got from IRC chat rooms of like minded music lovers and swapping .mp3's! Before anyone flames me on the ole FTP days... it wasn't yet illegal to swap digital media. In the early days it stood on the same legal footing as giving your friend a mix tape on a cassette/cd or dubbing from the radio. Windows didn't include a player that knew what the hell the .MP3 was and Napster was yet a long way out. Speaking of digital media I currently use Spotify Premium, locally stored phone media via Poweramp as well as jRiver Media Center with FLAC files to an outboard dac, and pre-amp/amp configuration at home. In the early days .mp3 came mostly in 96kbps and 128 kbps variety due quit simply to the slow ass connections we all had. As the bit rate increased and the codecs released later revisions (VBR was a game changer) the fidelity got better but still lagged behind all but consumer cassette tapes. Now we have 320kbps streaming at our finger tips and digital albums can be purchased in similar bit rates as well as lossless 44.1khz 16 bit and high resolution PCM, DSD/2x/4/xDXD etc.
If the album is from an artist I truly love, or one known to be immaculately recorded and an audio enthusiast staple I will think about picking up the high res version. Before I jump on the purchase I will typically look at reviews and see if a few key items are met:
1) Was it remastered for high resolution or in the case of multi-channel audio from the Quadrophonic tapes? I find that a lot of the benefit comes in the fact that the media was remastered with a known intended audience. And that audience is often quite picky and discerning so poor efforts don't sell in an already niche market. In some cases of previously well mastered media they decide they just need to "make it sound different" or use the same Redbook purposed master upsampled to the higher bit rate/depth. I blame this shady ass behavior for a lot of the negative perception of the various high res formats.
2) Is the price right? So much of the high res stuff is overpriced crap... but if you catch a coupon code or a sale it can make it into my "worth the purchase" range.
3) I don't buy high res Pop, Rap or EDM... for the most part, the dynamics are too low and the marginal benefit from the high resolution is often negated. I am not saying I don't enjoy those genres, I am just not paying a premium for them. Key high resolution genres for me are Acoustic/Instrumental, Jazz/Blues, some well recorded Classic Rock/Modern Rock/Hard Rock/Metal. Never been a big orchestral fan but I do own some high res Wagner.
Bottom Line: I think high res music does sound better than it's streaming or lossy counterparts but I also don't think it's enough of a difference that everyone will care to fork out the premium for it. I also don't see a huge difference between 192 and 320kbps or lossless Redbook FLAC and high resolution FLAC. The difference is there in some tracks but wholly masked to me in others by shite recording or some producer/engineer who just loves to add excessive layers of effects. I see pros and cons on both sides of the debate and everyone should make the right choice for their particular situation (aka wallet & ears).
The Question
Can you hear the difference between a lossy version of To the Sun and All the Cities in Between by Cities of the Sun and a high resolution version?
No these aren't from the same source up/downsampled because these days most users are presented with 5 options for consumption and I am again aiming for the practical not technical side of the debate.
1) Stream from a music service a la Spotify, iTunes, Google Play, Pandora
2)Stream from a hifi source like Tidal or Spotify (by invite ATM)
3) Purchase CD and rip to lossy or lossless format
4) Purchase digital lossy or redbook copy from Apple, Amazon, Google, etc...
5) Purchase high resolution copy from niche retailers like HDtracks

Spotify at 320kbps
https://open.spotify.com/track/0Fw6fTh1NZ6rzLGl26bGLr?si=x8GBD8_2TtuxR4PVLjuUqA
192kHz/24 Bit copy from HD tracks
http://www.hdtracks.com/to-the-sun-and-all-the-cities-in-between

My audio chain(s):
1) PC using jriver MC>XUSB>XDA-1>2808CI>XPA-100 x 2> ERT 8.3 towers +SW-12 subs x2.
2) PC using jriver MC>XUSB>XDA-1>Lyr/H10>T1/HE500
In any of the above scenarios I hear a marked difference in the decay of the bass notes, the clarity of the guitars, the depth of soundstage and the overall cleanliness of the image. If anyone has access to the same music or is highly musically curious like myself and cares to purchase said track, I would love to hear your thoughts EITHER WAY!
Request: Please be civil, let's try to break the mold on this type of discussion, there are a ton of factors at play and everyone's ears are different.
thumb_up
ChumbWumba, AK770, and 2 others
thumb_up4
36
remove_red_eye607


I have read many PCM~DSD evaluations; this is the most properly set-up and objectively executed test I have seen. You must, in such an evaluation, have hardware capable of the level of reproduction required. The integrity, for the method of obtaining true/proof comparison source files, is inarguable. Yes, I can so well hear the difference, DSD.
search
https://www.audiostream.com/content/dsd-v-pcm-file-comparison-16441-2496-24192-64x-dsd-128x-dsd
rastus
Thanks for the post, audiostream tends to put some serious thought into their comparisons and reviews.
I just built a new PC today, and I tried listening to music and knew something was very wrong... the output seemed as if it couldn't handle both bass and treble at the same time...
diving a lil deeper, I found the cause:
My optical output supports up to 16bit @ 96Khz Strange, as I always thought optical cables would give me the best results.
Then I linked my modi multibit via USB, and jacked up output to 32bit @ 192Khz
There was an immediate and HUGE difference.
(music was FLAC, mainboard is an ASUS ROG Strix Z370i)
Load 2 more comments
Yes, same DAC, eq and amp on each instance. The only difference was in the output by my pc to the DAC.
The DAC also has a switch on the fly, so it was easy for me to compare between the optical and USB.
The FLAC file is over 1,000kbps but not sure if it's 16/24bit. I used the exact same file for comparison.
Would how I have tested skewed the outcome somehow?
Again, the mainboard setting's max that I could hear something out of optical was 16/96. The usb had no issue jacking the settings up to max.(32/192)
Heavyboxer
OK, still my opinion is still valid, I develop audio product myself and there is more than just the resolution of the stream, USB because of it's very wide market as come a long way, in class 2 it's an asynchronous feed so it's all managed by the receiving end how accuratly clocked the samples arrive, with more error correction ressources, etc. I don't know exactly what Shiit do with the optical input and maybe they do it right, maybe it's also just a bonus side tought to offer an extra optionbut the amount of care they put on accuracy, i believe is more imprtant than just raw numbers. Same could be told about compression algorythms, they are not all created equal, for example for me a 256K AAC sound Better than a 256K MP3, and in most cases a 256 K MP3 download sounds better than Spotify streaming at 321 even if it's theorically better, more bit rate not necessarily better.
I am not going to buy stuff from HD tracks to compare, but I know quite a lot about digital audio, and I know that there's no way I could hear the difference between a 44/16 and a 192/24 properly downmixed from the same source. If the "high res" version sounds better, it's because it was mastered better. In many cases this is true. If original tapes, or pre-crushed sources are available, they can produce a much better track than something someone turned up to 11 and ran through a ton of dynamic compression to bring it to -3dBFS. That said, an equivalent remastering would sound just as good when resampled to 44/16 (and would take a lot less space)
IMO 320kbit lossy is just silly. If I want something max quality or for archival purposes, I'm going lossless (redbook is fine). If I just want it to sound good, and save some space so I can fit it all on my portable player, I'm fine with ogg q5 or q6, (160 or 192 ABR IIRC?) I generally keep all of my music as flac (rips from my CD collection, or ones I've purchased in lossless format), and I have an ogg copy of the entire tree, which I use on my portable player(s)
I _never_ convert lossy to lossless, or transcode lossy to another bitrate. (This sounds like hell, and wastes space)
When comparing online sources, you need to keep in mind that not all services follow my standards. I have spotify premium, and regardless of what bitrate they're feeding it to me at, there is "A LOT" of music that was clearly transcoded from trash-tier sources. Bubbly sounding under-water pre-echoes, and cymbals that sound like they're running through a chorus pedal, etc. This doesn't mean ogg q6 (or whatever spotify streams at) is bad. It means that they've transcoded low quality sources. It's like taking a high-res photo of a low res print.
Not going to lie, but FLAC is really good if you plan to transfer it, as it retains quality better than an MP3.
Otherwise sound quality lies entirely on the people who made the song. I’ve heard magic MP3 320 songs as well as FLAC that sound awful.
I tend to convert my MP3s to FLAC for the sake of ensuring nothing gets lost. Otherwise 2 well recorded FLAC and MP3 should sound the same.
Load 3 more comments
Hey I learned something new, thanks!
Whyweneed
Cool. The more people that understand this stuff, (especially in the music industry) the less bad sounding music we should all encounter.
It would be nice if companies like spotify would also treat things this way. only compressing from lossless sources, and treating any lossy tracks as "cooked" and always streaming them verbatim, without any re-encoding.
With my setup, I have yet to hear a significant difference between lossy MP3 or AAC at 320kbps and 256kbps respectively, and lossless FLAC and DSD at Redbook standards, much less Hi-Res. I might be able to pick up some minor differences in an ABX test if I'm really listening closely and using my highest end gear but I'd have to be really concentrating.
Load 2 more comments
Situation definitely has an impact on what formats I'm using. At work it's mostly Spotify streaming and bluetooth headphones as I'm working on other things and the increased fidelity isn't important/noticeable. At home I divide my time pretty evenly between my local files and spotify streaming. I use Spotify to locate new artists or listening to tracks where I don't want to pay extra to have my own lossless/hi-res copy. Musical enjoyment is most certainly the priority for me, the nicer gear just aides in that enjoyment. I would say my at home listening is also 50/50 casual vs. dedicated music listening.
So yeah... If I'm listening in the car, or my headphones on a noisy train, I likely wouldn't be able to tell the difference between flac and a q6 ogg that I typically use for portable use.
One of the main things that is important about having that lossless flac file is that it can be converted into whatever format you want, with only the 1 generation of losses. I can take my flac and encode it to a low bitrate MP3 for a crappy player in the car, a q6 ogg for portable listening on decent headphones, and listen to it at its maximum quality at home.
If I was stuck with 320kbit MP3, I either have to leave it alone, which takes a lot of space on my portable player, or I transcode it to a q6 ogg, which means I have the badness of 320mp3 and the ogg added together. (It sounds worse than either format would have if converted from an archival source like flac.)
Whether you can hear the difference in a single generation or not isn't really the issue, it's that you know that your music sounds as good as it can sound (or if it doesn't, you know it's because it was an intentional choice to compress it) If it sounds bad, you're not left wondering if it's a mediocre recording, or if it's just because it's been shuffled back and forth between formats a few times.
My experience has been the cheaper the gear, the easier to hear Hi Res, on a world class Rig, redbook is staggeringly good. so good, the last thing you think about is resolution or format. Same goes for tweaks.. The stronger the rig, the higher its integrity in not being affected by outside influences. mp3 is the absolute LAST way to listen to music if you respect music as well as your experience. Even .wav sounds better than flac in AB comparisons for me. Unfortunately the best audio repro costs. If a company claims that their $500 DAC sounds as good as the big boys, not in my experience. I actually set out to prove that the newest DAC technologies can and do sound as good as the $10K +, But all it takes is a really well designed $10K plus DAC to sit your ass back in the chair so you can think about it a little. Trust me it sucks, I'm not wealthy, but I do spend an unnaturally high amount for my gear because its the most important physical/material experience for me. Yep, I'm old enough that it passed up food and sex.. I am immersed in Audio most likely around 70% of my time.
Why would you even accept something with the name Lossy ? You dont even need to hear it.. they have told you up front ...
Load 3 more comments
But my point is, that if there were anything audibly worse about 16/44 audio than those other high-res formats, the most perfect audio system in the world would only make those differences more apparent. I believe that your system sounds very good with 16/44, what I'm disagreeing with is that a cheaper system will make the differences more easily heard.
It's not like you add up the "goodness" of all of your equipment and sources, and that defines how good the overall sound quality is. It's more like subtracting all the "badness" of the equipment and sources, and what you're left with is the overall quality.
In some cases if one part of the chain is bad enough, it can mask other problems, but there's never a case where one part can be good enough that it makes up for another deficiency elsewhere.
nwimpney
I see your point and agree. You cant take crap and make crap salad... or can you ? I'll leave that to the cooking forum. First off let me say I'm simply going from experience .. when I speak with other audiophiles that have high level rigs, most say the same thing.. format does not matter so much.. maybe digital or analog, but not between low and high bitrate PCM. Between DSD and PCM yes but youve got to be intently listening. strangely, many times DSD sounds worse. In other words Ive got a server chock full of 16/44, 24/96, 24/192, DSD, Double DSD, etc. When I am listening there is really no distinction.. Most of the time when I am amazed at how great the music sounds, its 16/44 with top shelf production. That being said, the only people/customers I hear talking about Hi Res or MQA are the non audiophiles or the fledgling audiophiles. When I ask them if it makes a difference or if they can hear it, its unanimous, "sounds better".
So again, I am going from experience not theory of what makes sense logically. Logic can be thrown out the window with high end audio. (see: audiophile fuses)
I think you can add up the goodness because the starting point is always shit. Mp3, radio, etc. Every tikme I add a higher level piece my system sounds better. So the goodness absolutely adds up, Subtracting the badness means we start from where ? "perfect sound" whats that ? But I agree with you that you cant add something thats not there to begin with..
Cheers
For me, it would be difficult to discuss discernible differences in audio quality without going through some history and my experiences with lossy compression.
When I was 14, the MP3 format had just started to emerge. I was experimenting with encoding my CDs in various compression rates whether it were fixed or variable. Back then, it was really easy to tell if something was encoded in MP3 even at higher bit rates. Anything encoded in 128kbps sounded like it was ran through a ”pluging-into-the-matrix” bitcrusher filter. Sounds encoded in 320kbps didn’t sound as bad, but still, the artifacts of compression were very apparent. This mainly had to with the quality of encoders.
Since then, different methods have been employed in encoding MP3s and M4As which nowadays, contain less of those artifacts. You can actually invert a modern 320kbps encoded MP3 and subtract it from a 44.1khz CD quality image and discover that very little is taken out. It’s at a point now that, unless you‘re doing an A/B comparison, there’s very little telltale signs that something has been compressed. However, there very much still exists a difference between lossy and lossless audio.
Although it’s not as aggregious as it was with older encoders, the difference is still noticeable; specially if you’ve spent the last 5 years listening to compressed, lossy music, and make the switch back to lossless. At 44.1khz / 16bit lossless, you start to hear a better representation of the soundstage. This is due to reverb trails being reproduced at a greater fidelity, allowing your ears to hear their full extent which perceptually opens up the space where the sound lives. Minute differences in amplitude become discernible. It’s microscopic, but you start to hear the trail end of snare hits last half a millisecond longer and the leading end of transients extending further out. Although minor, it makes quite a difference in how we perceive the music. The returning frequencies that were lost in compression also breathe a new life to songs. Then there’s diminishing returns with increasing to higher sample rates like 96khz / 24bit masters due to the limited physical ability of our biology to transduce differences in atmospheric pressures. There’s also the aftermath where the marginal difference in fidelity we experienced when first switching to lossless gets lost eventually due to our ears getting accustomed to the sound. At least we can rest in the fact that we’re getting the best quality if we have lossless.
As it stands, I would really love to listen to lossless since there are parts of songs completely lost in lossy compression. But, having no desire to buy and store physical copies of CDs and wanting the convenience of streaming content over the internet, I haven’t got much choice. Tidal offers lossless for around $25, but until they improve their music library and the god-awful app, I shall have to be content with Spotify’s 320kbps service.
Isn't it futile to ask anyone to see if they can tell the difference when you're showing us which one is which?
curleyfrei
Scientifically controlled this study is not... think of it more as an opinion poll on the value people attribute to different audio sources. I've found very little in life is all or nothing, the meat lies in enjoying the spectrum and finding ones own "happy place."
NPR has a pretty cool blind test with samples from 6 songs (3 different qualities). On that, I can say that I can tell the difference between 125kbps and 320kbps, but I can't pick out lossless. I was 5/6 on the 320kbps (with 1/6 being 125kbps).
I typically listen to 320kbps as that is the best quality my streaming service provides, but since I also can't tell the difference of quality for anything over 320kbps (at least on my current gear, with my ears) I am not sure looking for lossless would make much difference.
BetaWar
I did that test yesterday and it was a humbling experience. I will try it again in a few different configurations (I did it with my Sennheiser HD630VBs plugged directly into my MacBook Pro 2017), just to satisfy my own curiosity, but I only "scored" 2/6 in picking out "hi-res". I am now in my early 50s and I have noticed a degradation in my hearing over the past few years (same with my eyesight); so not overly surprising. A bit disappointing though. I did get it "right" with the classical selection, which I consider logical as it is with all acoustic instrumentation (not electronic, electric or otherwise amplified). But still, the margin was slight.
When my HD6XX arrive, I will try again--this time through my Schiit Stack (Modi2/Magni2--both Uber). I will also re-try my HD630VBs with that stack, to see if the amplification and/or DAC make a difference in my ability to discern the files.
One upside, if I still can't tell with a better average, is the savings for my wallet. I have been an Apple Music subscriber for convenience for about 18 months--and my disc buying has plummeted. I do still buy Blu-ray Audio, SACD and, more rarely (as they are hard to find), DVD-A. But I have always bought them for their MCH mixes first, and "hi-res" considerations a fairly distant second (CD has been "good enough" for sound, for me--making no judgements for others). I believe the "better" sound in two channel "hi-res" comes from the mastering with wider dynamic range than is usually done with mainstream CD releases, rather than the formats themselves (again, to my ears).
If I do find a difference with the new headphones and using the other gear, then I will likely step up my disc buying a bit (not a big buyer of downloads, though that might change), but not too much. I have nearly 2000 CDs, 300 SACDs and nearly 100 DVD-As already. I'm at the point where finding time to listen to what I have is hard enough. ;)
Edit to add: When I import CDs into iTunes (I'm lazy--iTunes is free and there--I know there are better media software players and, one day, when I'm not so swamped with working, parenting and doing a full-time PhD, I'll explore some of them), I do so in lossless ALAC. Probably couldn't tell the difference between that and lossy for most things, but I have the space and it's better to have a lossless base (I convert to 256 AAC for my iPhone/iPad to save space).
I remember the early days of mp3--digital artifacts on low-bitrate files. It seems those days are gone (well, mostly--my wife's previous car sounded a lot like those old crappy mp3 files with satellite radio, while her new one, and my current one, do not sound like that--I think that's owing to the head unit the car company chooses). For archiving my CDs, lossless seems like the best bet (much like many films are restored/archived at much higher resolution than they are presented).
In the end, the key to good sound is careful recording and mastering first, format second. The lossy stream of much of my Apple Music classical library sounds much better than some of my SACDs of music I like but suffers from "loudness wars" mastering compression.
I‘m currently undecided at the moment on the value proposition beyond Redbook standard due to murky source lineage of most commercially available recordings.
For me with current gear (Windows 10, FiiO E10k DAC, O2 amp, Fostex TH-X00), it’s hard to differentiate between 320 kbps and lossless. Highly dependent on content based on my experiences with tests such as https://www.npr.org/sections/therecord/2015/06/02/411473508/how-well-can-you-hear-audio-quality.
I‘ve not critically considered bit depth and sample rate. The 96khz/24 bit FLAC content I’ve listened to so far was well mastered and unfamiliar. I‘ll try the Spotify vs HD Tracks in A-B sighted testing though my chain is limited to 96khz/24 bit until my new DAC arrives. -------- Updated 5/5/2018 Replaced the E10K with a Topping D10 and added the MD Sennheiser 6XX. To my ears, the D10 seems to have more detail than the E10K. An hour long sighted A/B comparison of City of the Sun: To the Sun and All the Cities In Between, 96kHz/24 bit FLAC from HDTracks on Foobar compared with 44.1 kHz/16 bit @ 320 kbps streaming on Spotify. Still getting familiar with this material, so I focused on track 13: Everything. The guitar picking seemed slightly clearer at times within the decay and reverb. There also seemed to be a little more detail in the bass and percussion, tighter on attack in places. Both sources were enjoyable on the 6XX's and TH-X00's.
I myself have never been into the idea of HD audio because I am not really sure how higher res really adds anything? for me its been either 320 mp3 or flac files. I have tried some HD tracks for tests and never was wowed by it just don't think I listen to music critically enough. Lets compare it to the HDTV of today that looks so much better than the SD of the day but it seems the gap is nowhere near with the HD audio I think that says a lot about the quality of CD and our ability to hear that difference.
Load 1 more comment
I had the dilemma not long back about tidal HiFi and came to conclusion that I think the fact the light was on actually made me feel a difference but I personally didn't see the worth in it. Maybe like you say highly dynamic music or complex pieces will retain more detail but if I have to sit through a song listening to the tizz of a symbol I think I am doing something wrong lol.
Azamonde
I tried Tidal in its infancy and was fairly pleased but I ran into a series of "no results found" when looking for various artists. At this point if Spotify Hi-fi goes mainstream I plan to jump on that, beyond a handful of isolated cases their library has always taken care of my needs.
Feel free to discuss the topic in general... if you buy high resolution music how do you decide what to pickup and what to not? Are you a local file only kinda person or streaming guru/proponent? Again want to keep the thread civil but I would like to help those who are "high resolution curious" to get a spectrum of viewpoints on how it compares to the plethora of other options out there.