I understand what you were trying to say, your metaphor was just very poor. It also contains the totally invalid inference that cheap equals low quality.
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Anybody Out There Who Hears a Difference Between 320 kbps MP3 and Red Book CD? What Differences Do You Hear?
- Thread starter VMAT4
- Start date
I understand what you were trying to say, your metaphor was just very poor. It also contains the totally invalid inference that cheap equals low quality.
Okay, I've finished my ABX DBT of Soniclife's song.
I listened to a loop of the part ranging from 0:42 to 0:52, paying close attention to the percussion sound (hi-hat?) at 0:48.
Here are my results:
If anyone else wants to try, they can download the file, along with my 320 MP3 encode, from here: https://mega.nz/#!hpYCzIAY!d-49RVbCZoRt0mj4VCg4XqTTq--n6kE0k6oV4jVeLmI
I listened to a loop of the part ranging from 0:42 to 0:52, paying close attention to the percussion sound (hi-hat?) at 0:48.
Here are my results:
foo_abx 2.0.6c report
foobar2000 v1.3.16
2019-11-25 20:15:10
File A: 05 - 1979 (FLAC).flac
SHA1: 580c09232a2bf114ae37cbd8da5eb418d707b0b3
File B: 05 - 1979 (MP3).mp3
SHA1: b08d4695ba6ec32cf230debbddb0cb3a98ada382
Output:
DS : Primær lyddriver
Crossfading: NO
20:15:10 : Test started.
20:16:19 : 01/01
20:17:04 : 01/02
20:18:11 : 01/03
20:19:13 : 01/04
20:20:19 : 01/05
20:20:57 : 02/06
20:21:28 : 03/07
20:21:58 : 03/08
20:22:30 : 03/09
20:23:28 : 03/10
20:23:28 : Test finished.
----------
Total: 3/10
p-value: 0.9453 (94.53%)
-- signature --
abddcb6d7b35435b9e4221b2a2ffca33c64dba55
foobar2000 v1.3.16
2019-11-25 20:15:10
File A: 05 - 1979 (FLAC).flac
SHA1: 580c09232a2bf114ae37cbd8da5eb418d707b0b3
File B: 05 - 1979 (MP3).mp3
SHA1: b08d4695ba6ec32cf230debbddb0cb3a98ada382
Output:
DS : Primær lyddriver
Crossfading: NO
20:15:10 : Test started.
20:16:19 : 01/01
20:17:04 : 01/02
20:18:11 : 01/03
20:19:13 : 01/04
20:20:19 : 01/05
20:20:57 : 02/06
20:21:28 : 03/07
20:21:58 : 03/08
20:22:30 : 03/09
20:23:28 : 03/10
20:23:28 : Test finished.
----------
Total: 3/10
p-value: 0.9453 (94.53%)
-- signature --
abddcb6d7b35435b9e4221b2a2ffca33c64dba55
ernestcarl
Major Contributor
You’re equipment chain probably was too cheap and low quality to resolve the details.
You’re equipment chain probably was too cheap and low quality to resolve the details.
MRC01
Major Contributor
For my ABX app, I created some clips to help people get started with testing. Anyone can download them here and use them in any ABX app:
http://mclements.net/clementslogic/abxAudio.html
This page links to two different high quality recordings, each having the original, plus others MP3 encoded at 64, 96, 128, 192 and VBR. These are 30-60 second snippets distributed for educational use in ABX testing, so I believe posting them conforms with US copyright law.
http://mclements.net/clementslogic/abxAudio.html
This page links to two different high quality recordings, each having the original, plus others MP3 encoded at 64, 96, 128, 192 and VBR. These are 30-60 second snippets distributed for educational use in ABX testing, so I believe posting them conforms with US copyright law.
Last edited:
Sgt. Ear Ache
Major Contributor
I'm not really sure what the point of low br mp3s are. I mean I don't think anyone is claiming the difference isn't noticeable under 192kb fairly easily...
There is also the decoder not just the encoder, if you use a computer you’re probably good .
But many devices with built in decoders may not be so good , I only have one datapoint on my own sorry .
The built in MP3 decoder in some of my Squeezeboxes are not top notch, probably from missing floating piont math and no tolerance for >odB peaks . Result is clear artefacts on some tracks others are just fine . Since then I do serverside decoding on my very few MP3 tracks
But many devices with built in decoders may not be so good , I only have one datapoint on my own sorry .
The built in MP3 decoder in some of my Squeezeboxes are not top notch, probably from missing floating piont math and no tolerance for >odB peaks . Result is clear artefacts on some tracks others are just fine . Since then I do serverside decoding on my very few MP3 tracks
Robin L
Master Contributor
I hear the same sorts of smears, sonic smudges and loss of resolution as with lo-bit rate mp3. However, there's a whole lot less of these artifacts, so that sometimes I hear no difference. However, as I listen off of the ripped files of my CDs, and storage is cheap, I go Apple lossless with everything. Most of the time, I'm listening over pretty good headphones. If I listen over my speakers, I don't hear a difference.
Robin L
Master Contributor
I think it has more to do with the resolution of the playback gear. By way of example, if there is some background noise from some other source, even if it's low-level, it can override the difference in sound.
You'll be surprised how many Johnny come latelys will argue the point with you on this thread...
Robin L
Master Contributor
When I was using a 1gb IPod Shuffle I squeezed as much music as I could into the gizmo. Lots of swishing and whooshing noises. Guess it was 128 kbps.
Sgt. Ear Ache
Major Contributor
Most of the conversation in this thread (as well as the title of the thread) has been in reference to high br lossy. Usually 320kb...which I completely agree is very very hard to distinguish from lossless in most circumstances.
Here's a track submitted a day or two ago, in a high frequency discussion:
The greatest difference appears in the applause at the very end, but that's also the loudest (and clipping) part.
You don't save much space over lossless FLAC
Waste o' time.
The greatest difference appears in the applause at the very end, but that's also the loudest (and clipping) part.
You don't save much space over lossless FLAC
Waste o' time.
Last edited:
Good point. People who don't understand what @Mnyb is talking about should read this:
https://forum.cockos.com/showpost.php?s=7bc696edfff860768ea52af3ad840718&p=2001665&postcount=30
...and try the files here. Remember to read the replies from other members up to page 5 as well.
https://www.audiosciencereview.com/...y-of-software-volume-control.5922/post-172865
ernestcarl
Major Contributor
It has been said by many before that space is cheap... I agree. If one has the original ready at hand, sure, keep that if space means little. However, I’ve for the longest time bought my music compressed because that was always the easiest and cheapest way to buy music — e.g. iTunes etc. I still have a few 24/96 flac downloads, but not many. It makes zero sense to upgrade every single thing at this point. I still have a bunch of DVDs which I have no intention of repurchasing HD or UHD copies of — unless for some very special reason, why bother? Some people simply want the best copy currently available regardless of the inevitability of senescence, imperfect transducers and room acoustics etc. That’s fine too.
Another important point is for some encoders, like the recent version of LAME, 320kbps is not necessarily better than V0 (highest quality VBR) in all cases. Try eig.wv in this post:
https://hydrogenaud.io/index.php/topic,49601.0.html
I attached the 320k and V0 (294kbps) files here. The difference in the first beat is very obvious.
https://hydrogenaud.io/index.php/topic,49601.0.html
I attached the 320k and V0 (294kbps) files here. The difference in the first beat is very obvious.
ernestcarl
Major Contributor
Lossy encoders are not perfect — certainly not even V0 VBR over 320kbps CBR mp3 for all lossless audio files. Still, I always make it a point to download the latest version of Lame when planning to compress a lot/batch of files. I sometimes do borrow a bunch of discs from the library and “archive” stuff I want to listen to again locally — if it’s a video disc, more often than not, delete the files later. Interlibrary loan system here is great... I pay my taxes, don’t I? why not make thorough use of the broad library resources. LOL
The encoder matters a lot. A Xing-encoded MP3 is painfully obvious even at 320kbps, but LAME is extremely competent already at 128kbps.
I've been able to ABX 320kbps LAME MP3 against the lossless source, but only by repeatedly looping very short (1s or less) sections and quickly switching between A, B, X and Y in order to pick out a very isolated reverb tail or resonance or something similar.
I sure hope nobody actually listens to music like that for pleasure!
It was a while aho, but IIRC it was a solo harpsichord recording (which is a know problem for MP3, along with castanets) and it was more a subconscious thing than something I could actually put my finger on an identify. I could never even come close to hearing a difference when actually listening to the music, and not just isolated sounds.
All of the more modern codecs (Vorbis, Opus, AAC and so on) are absolutely 100% transparent to me, at much lower bitrates. For my phone I use Opus at 128kbps, and even that is overkill with a decent amount of "just in case" margin.
I've been able to ABX 320kbps LAME MP3 against the lossless source, but only by repeatedly looping very short (1s or less) sections and quickly switching between A, B, X and Y in order to pick out a very isolated reverb tail or resonance or something similar.
I sure hope nobody actually listens to music like that for pleasure!
It was a while aho, but IIRC it was a solo harpsichord recording (which is a know problem for MP3, along with castanets) and it was more a subconscious thing than something I could actually put my finger on an identify. I could never even come close to hearing a difference when actually listening to the music, and not just isolated sounds.
All of the more modern codecs (Vorbis, Opus, AAC and so on) are absolutely 100% transparent to me, at much lower bitrates. For my phone I use Opus at 128kbps, and even that is overkill with a decent amount of "just in case" margin.
Xing is crap. IIRC quite a number of people was using the good old Fhg or "Radium" encoder for better quality, but they are very slow, and I switched to LAME since 3.96.1 or so... at around 2005?
Similar threads
- Locked
