I don't think you'll have to wait too long for that. I do still find myself tracking down CD copies for most non-vinyl things I want (and kinda like the physical copy anyway), but I'm seeing more and more [legal] FLAC downloads out there. Make sure you check the artist and label websites while you're waiting on iTunes and Amazon.
Test Your Digital Format Perception
- Thread starter dumptruck
- Start date
BadassBob
Digitally rendered
This. I purchase a lot of my music direct from the labels, and many do offer both lossy and lossless files. Purchasing direct from the label was my main avenue of purchasing FLAC files until Bandcamp expanded their offerings.
It's an interesting problem. I'm 57, and have too much stuff, despite winnowing. That includes CDs. I'd like to continue to get new music, but I'm really not interested in the care and feeding of it - of course, that cuts both ways. If I don't have a physical copy, then I have to obsess about back ups, migrating formats, etc.
What I think about now is - if I live to be 83 (Social Security's very rough guess of my death age) that gives me an absolute maximum of 365 days times 25 years to listen to music I love.
Assuming that I can get in an album a day, that would mean that if I own only 365 albums, I can only listen to each one another 25 times. That doesn't account for, say, the big boxes from Mosaic, where I have five or six cds in the set.
In light of that, how many cds should I own? I guess another way of putting it is, my attention is now a finite resource, one I can see the end of. How do I best "spend" it?
s.
KOWHeigel
High Gear of Your Soul
Although I am not super familiar with Windows audio (linux guy) ... I would warn that some care should be taken in the setup of foobar to make sure that all audio is not simply being resampled by windows mixer thereby possibly negating any potential differences in the files to be ABX'd.
Yes, if you're comparing high quality formats using an older version of Windows, that may be a concern (lots of info on the web for getting around the old windows mixer). I'd suggest anyone giving this a shot start with a humble 192k CBR MP3 vs. WAV/FLAC, though.
Wigwam Jones
Caesar non supra grammati
Keep waiting.
robgmn
Super Member
Looking forward to giving this a shot this weekend.
I know for a fact that I can hear the difference on SOME music between 192K (fixed rate) and uncompressed, but bet it becomes very difficult at the higher encoding rates.
As for FLAC versus WAV, since there is technically no difference between the two I am hard-pressed to believe anyone could hear a difference unless their codecs were causing some problems with playback
I know for a fact that I can hear the difference on SOME music between 192K (fixed rate) and uncompressed, but bet it becomes very difficult at the higher encoding rates.
As for FLAC versus WAV, since there is technically no difference between the two I am hard-pressed to believe anyone could hear a difference unless their codecs were causing some problems with playback
JimmyNeutron
Super Member
Several people have asked me off-board about my own results. I'm not trying to keep secrets here, I just didn't state them because my goals with this thread are:
1. have a thing I can link to that explains how to do perform one of the VERY few potentially 100% valid objective tests in hifi that can be done with what people have at home.
2. maybe instigate a few experiences that cause people to change their tone when discussing digital formats.
This isn't a challenge to hear better than me. I'm sure many of you can, just as I'm sure I could 20 years ago when HF noise all over the place tortured me. I did a handful of tests years ago, which were a lot of work because I didn't have this handy software. My results were:
1. I can identify almost any 128k MP3 vs. redbook example. At the time of posting this thread, I tried a recording I think sounds a bit odd and looks a bit odd in analysis, Cate Le Bon's Mug Museum. I failed! I may try it again.
2. At 160k, I got some recordings 10/10, but others I had no idea.
3. At 192k, it seemed like I was doing well, but I kept getting some wrong answers, and didn't have the patience to do enough trials to sort it out.
4. At 256k, I couldn't tell. I didn't think I could tell subjectively, and I certainly couldn't tell objectively, with any of the recordings I chose.
This was all a long time ago, and I have much better equipment now. If I hear a lot of people in this thread claiming better results, I might try it again with the tracks they used, just for curiosity's sake. On the other hand, I just don't care that much, I want lossless tracks, and I'm laser-focused on speakers for the forseeable future. I just want people that DO care about this stuff to at least know their own hearing a bit.
1. have a thing I can link to that explains how to do perform one of the VERY few potentially 100% valid objective tests in hifi that can be done with what people have at home.
2. maybe instigate a few experiences that cause people to change their tone when discussing digital formats.
This isn't a challenge to hear better than me. I'm sure many of you can, just as I'm sure I could 20 years ago when HF noise all over the place tortured me. I did a handful of tests years ago, which were a lot of work because I didn't have this handy software. My results were:
1. I can identify almost any 128k MP3 vs. redbook example. At the time of posting this thread, I tried a recording I think sounds a bit odd and looks a bit odd in analysis, Cate Le Bon's Mug Museum. I failed! I may try it again.
2. At 160k, I got some recordings 10/10, but others I had no idea.
3. At 192k, it seemed like I was doing well, but I kept getting some wrong answers, and didn't have the patience to do enough trials to sort it out.
4. At 256k, I couldn't tell. I didn't think I could tell subjectively, and I certainly couldn't tell objectively, with any of the recordings I chose.
This was all a long time ago, and I have much better equipment now. If I hear a lot of people in this thread claiming better results, I might try it again with the tracks they used, just for curiosity's sake. On the other hand, I just don't care that much, I want lossless tracks, and I'm laser-focused on speakers for the forseeable future. I just want people that DO care about this stuff to at least know their own hearing a bit.
ashok
Active Member
In this other thread on 24 bit / 96kHz audio, there was a small sidebar about the ABX plugin in Foobar, and how it internally downsamples all higher than 44.1kHz signals to 44.1kHz.
I wanted to point out that it depends on the operating system you use, and the method you use to get the data out of your computer to the external DAC.
For instance, when I did my ABX tests, I used the following setup:
If the computer and/or Foobar were not configured properly, one could end up with the ABX tester indeed downsampling to 44.1kHz, but that was not the case with my tests.
Of course, before my tests, I restarted the ABX plugin, and covered up the indicator lights on the DAC.
I wanted to point out that it depends on the operating system you use, and the method you use to get the data out of your computer to the external DAC.
For instance, when I did my ABX tests, I used the following setup:
- Computer based on Intel dn2800 motherboard, running Windows 8.
- Foobar set to communicate to a Meridian Audio Explorer DAC via Wasapi
If the computer and/or Foobar were not configured properly, one could end up with the ABX tester indeed downsampling to 44.1kHz, but that was not the case with my tests.
Of course, before my tests, I restarted the ABX plugin, and covered up the indicator lights on the DAC.
I was assuming that anyone who would be doing this with higher-than-CD quality files would already be familiar with getting them to play without downsampling.
It was suggested that the ABX plugin itself downsamples, which does not appear to be true to myself or ashok, but I am certainly open to editing the instructions if someone can point to any real issues/problems with the software suggested. Not trying to trick anybody here. I'm about to add some notes on interpreting results now.
It was suggested that the ABX plugin itself downsamples, which does not appear to be true to myself or ashok, but I am certainly open to editing the instructions if someone can point to any real issues/problems with the software suggested. Not trying to trick anybody here. I'm about to add some notes on interpreting results now.
robgmn
Super Member
Having done this test, I can say "WOW"!
I KNOW I can tell the difference between 192K and uncompressed, but never tested it doing my own compression. It was always perceived in listening to files encoded by others (in the early days, I downloaded under Fair Use instead of going through the hassle of ripping all of my CDs myself).
Compressing a file myslef through Audacity/Lame (David Huntsinger - The Christmas Song) has opened my eyes.
Through speakers (Carver), headphones (Skullcandy) and earbuds (Etymotic), I found the differences a LOT harder to discern than I expected.
Once I got to 192K, it was just barely over a guess statistically.
At 128K I was over 98.5% accurate, but was impressed as hell at how good a job the compression did.
If online radio could do it as well as Audacity/Lame does it, a true 128K stream would blow FM into the weeds.
This leads me to believe the following possibilities: The files I previously listened to were encoded poorly (by operator fault or by software fault). If by software fault, then encoder software has improved.
A lot.
It'd be great to hear other results of this test, but I'd bet a nickel that those who are most vehement in their anti-MP3 diatribes will be the least likely to actually try it themself...
I KNOW I can tell the difference between 192K and uncompressed, but never tested it doing my own compression. It was always perceived in listening to files encoded by others (in the early days, I downloaded under Fair Use instead of going through the hassle of ripping all of my CDs myself).
Compressing a file myslef through Audacity/Lame (David Huntsinger - The Christmas Song) has opened my eyes.
Through speakers (Carver), headphones (Skullcandy) and earbuds (Etymotic), I found the differences a LOT harder to discern than I expected.
Once I got to 192K, it was just barely over a guess statistically.
At 128K I was over 98.5% accurate, but was impressed as hell at how good a job the compression did.
If online radio could do it as well as Audacity/Lame does it, a true 128K stream would blow FM into the weeds.
This leads me to believe the following possibilities: The files I previously listened to were encoded poorly (by operator fault or by software fault). If by software fault, then encoder software has improved.
A lot.
It'd be great to hear other results of this test, but I'd bet a nickel that those who are most vehement in their anti-MP3 diatribes will be the least likely to actually try it themself...
Last edited:
This is the sort of thing I was mostly hoping people would try, and find. I got sucked into discussing beyond-CD stuff in that other thread, but I really want people to try this with MP3s (and other lossy formats). When you compare, non-blind, a 128k MP3 vs original, it often seems like a joke, but this is a psychological enhancement of the difference that would not apply with normal, blissfully ignorant listening to the bad quality file.
Don't get me wrong, there is little reason to create a 128k MP3 in 2014, but I think it would be nice if more people did this sort of test and perhaps dialed down their scoffing at, say, people evaluating speakers with 320k MP3s. I'm pretty sure they can still tell if the speakers are any good.
As for your comments on online radio, it could be the same psychological effect, but the streaming services have unknown libraries of sources, and who knows how any given simulcast from a real radio station is being processed and encoded. That will probably have to remain a mystery.
P.S.
You're obviously not leaping to any conclusions, but just to be clear, being "barely over a guess statistically" would mean a 5% or less "chance of guessing" as displayed by the test, with at least 10 trials. If it says, for example, 45% chance that you are guessing, that means, statistically, nobody knows if you are guessing or not. You could keep doing more trials, and if you keep getting better than 50% right it will eventually mean something, but your odds aren't good if it's near 50% chance of guessing after 10 trials. Probably better to try a different file if you want to keep going.
Last edited:
charles 1973
Super Member
When I got my computer 11 years ago, I had alway's dismissed MP3 Music as a poor choice for Audiophiles. One day, I decided to try this thing they call MP3. I Tried it at Various Bit rates from 48 - 320k, both VBR and Fixed. I was pleasantly surprised with the results through my computer speakers, which for the Freq range they cover aren't bad with respect to resolution.
I found by 192 k VBR the sound was pretty close to Cd. I concluded that a 256 k VBR MP3 File to be a good choice for most listening on a Real stereo system. As a joke, I Even recorded a song on a 3.5" Floppy Disc at 48 k vbr, think of it as a Digital 45. While hardly audiophile grade sound, It didn't sound Bad either.
I have played my 192 k MP3 disc on my main stereo system and must say, it sounded pretty darn good. Of course with today's digital storage devices I'd probably choose a lossless format. But certianly for the car, Even today, I have no problem with a MP3 disc at 192k vbr.
I found by 192 k VBR the sound was pretty close to Cd. I concluded that a 256 k VBR MP3 File to be a good choice for most listening on a Real stereo system. As a joke, I Even recorded a song on a 3.5" Floppy Disc at 48 k vbr, think of it as a Digital 45. While hardly audiophile grade sound, It didn't sound Bad either.
I have played my 192 k MP3 disc on my main stereo system and must say, it sounded pretty darn good. Of course with today's digital storage devices I'd probably choose a lossless format. But certianly for the car, Even today, I have no problem with a MP3 disc at 192k vbr.
Last edited:
robgmn
Super Member
Online "radio"- Spotify is a great example. Some of the music available sounds like a Youtube feed that was run through a flanger. So yes, I'd want provenance before paying for a "high quality" stream from any provider.
No further testing needed. I'm suitably impressed with the trials I did and am comfortable with the results.
They were more for my own edification and enjoyment anyway, as I am running almost 100% lossless on my digital listening at home/work (portable sound quality is a non-issue as long as there's music
)...
Wigwam Jones
Caesar non supra grammati
Well, after all is said and done, I guess I will stand by my original statement - human nature. No one wants to find out the truth (whatever that may be); they're afraid of it. They have too much invested, emotionally as well as financially, to be 'wrong' about what they can and cannot hear.
They will either ignore the challenge, explain in minute detail all the things that are wrong with the testing procedure, or obfuscate, ignore the request, and insist that they can still hear the difference.
Not a shocker.
The best answer I heard was the one about choosing higher-rez regardless of whether or not it sounds better, just because it makes them happier. Cannot argue with that.
They will either ignore the challenge, explain in minute detail all the things that are wrong with the testing procedure, or obfuscate, ignore the request, and insist that they can still hear the difference.
Not a shocker.
The best answer I heard was the one about choosing higher-rez regardless of whether or not it sounds better, just because it makes them happier. Cannot argue with that.
Wigwam Jones
Caesar non supra grammati
You are a very optimistic person. :thmbsp:
Ahh, Wiggy....what I mistook years ago for rabble-rousing in your posts, is simply hardcore devil's advocacy.
You called it, brother. :thmbsp:
My ears have gotten so bad with my ever-increasing tinnitus and hypertension (which aggravates the tinnitus) that these type tests are pointless for me.
I'm pretty sure I cannot tell the difference between 320 and FLAC. Anything above 256 is a challenge, and only with cymbals and acoustic guitars can I perceive a difference.
With that said, I rip to FLAC. WHY? Because one day, perhaps stem cell research might fix my ears. Maybe Oscar Goldman will rebuild my ears. Hell, aliens might give me supersonic hearing.
All jokes aside, it makes me happy. And, one day my gear might get to the point where it reveals it.
There's my 2 pence.