HydrogenAudio

Hydrogenaudio Forum => Uploads => Topic started by: /mnt on 2008-04-02 14:47:23

Title: A sample that Helix Mp3 sounds really bad on.
Post by: /mnt on 2008-04-02 14:47:23
I have uploaded a zip file with a 30 min sample, that was edited on Audacity as a FLAC encode, and a Helix VBR Mp3 encode. For some reason it sounds really bad on Helix, it sounds almost like it was encoded on Blade, there is warbling and precho all over the place, while on iTunes at 128kbps VBR it sounds fine but not transparent but alot better.

Helix (Rarewares build) -X2 -U2 -V65

Code: [Select]
foo_abx 1.3.1 report
foobar2000 v0.9.5.1
2008/04/02 13:40:09

File A: F:\Replica(Sample).flac
File B: F:\Replica(Sample-HelixV65).mp3

13:40:09 : Test started.
13:40:26 : 01/01  50.0%
13:40:30 : 02/02  25.0%
13:40:33 : 03/03  12.5%
13:40:36 : 04/04  6.3%
13:40:39 : 05/05  3.1%
13:40:42 : 06/06  1.6%
13:40:48 : 07/07  0.8%
13:40:52 : 08/08  0.4%
13:40:59 : 09/09  0.2%
13:41:11 : 10/10  0.1%
13:41:15 : 11/11  0.0%
13:41:18 : 12/12  0.0%
13:41:20 : Test finished.

----------
Total: 12/12 (0.0%)


Very easy to ABX it sounds like a classic Blade encoded mp3.
Title: A sample that Helix Mp3 sounds really bad on.
Post by: shadowking on 2008-04-02 15:08:01
I think helix will have problems like lame with ringing / flanging only much worse. I encountered it at 160~180k with relaxed listening to normal music. It happened very fast with helix and heavy music, but with lame it was over a few years that I discovered 2 ugly mistakes . it seemed easier to find helix flaws at 160k than lame 130k. That made me weary of helix but I haven't tested much.

I heard the sample and it sounds like ringing and preecho. Lame vbr is generally good at 130k with loud stuff, but less so with quite music (ringing).. I wonder how helix would do solo intros etc.. I suspect it will be even worse.
Title: A sample that Helix Mp3 sounds really bad on.
Post by: /mnt on 2008-04-02 15:17:19
I think helix will have problems like lame with ringing / flanging only much worse. I encountered it at 160~180k with relaxed listening to normal music. It happened very fast with helix , but with lame it was over a few years that I discovered 2 ugly mistakes . it seemed easier to find helix flaws at 160k than lame 130k. That made me weary of helix but I haven't tested much.

Am not suprised that Helix performs worse then most encoders, but I did not know it performs that bad on most of the tracks I tired like this sample, it even sounds worse then a average 128 CBR file even if it is on VBR. I have heard a couple Xing encoded files in the past which sound better then this. ATM I am woundering if it is the build  have used, or my headphones might be faulty and beeing sensitive on artifacts, or somthing like that.