Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Why is Xing crap? (Read 5271 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Why is Xing crap?

We all know Xing MP3 encoder sucks badly. But why does it suck so bad?
I mean, from a scientific point of view. The people that created Xing surely
didn't wake up one day and said "Hey, let's make the worst mp3 encoder
on the planet!". What's wrong with the encoder tech-wise?

P.S. I guess the same question could be asked for Windows Media Audio...
Wanna buy a monkey?

Why is Xing crap?

Reply #1
Somebody will invariably post a list of problems but two come to mind immediately:

* no short blocks - this is why Xing encodes can sound so swishy and watery (especially at 128kbps CBR and even at 256kbps).
* destroys signals above 16KHz

Why is Xing crap?

Reply #2
Well, the goal of Xing was speed ( up to 8x speed compared to others, they said ) so we can't expect, from the start, the best quality.

To achieve this, they didn't just use better code and optimizations. Rather, they removed the parts that took more time, and give less impact in quality.

As Mithrandir has said, the first thing was not to use short blocks. If they don't need to detect when it needs to use short blocs or long blocks (as well as the extra CPU usage of making these short blocks), they can spend that time on other things.

Second, the problem about the joint stereo. Xing uses alot of Joint stereo frames compared to LAME (or Fhg), without caring if the result is good or not (In fact, I don't know how does it decide to use a JS or a MS frame. Fhg used to use more JS. frames in low kbps, and less in high kbps).

Next, the problem above 16Khz. Xing was made when MP3 at 128kbps was the almost exclusively used bitrate, thus, and thinking that Fhg uses (or used to use) a filter at 16Kbps in this setting, and thinking about the sfb21 problem, they initially discaded everything over 16Khz. Less bits to care about.

With AudioCatalyst, they improved, made the VBR modes, and added the 20Khz cutoff,  (as an option) instead of the 16Khz cutoff, but for some reason, the analysis to an MP3 encoded with Xing, shows faults from 16Khz and upwards.
The origin of these faults is not clear, but since the codec was already made with the idea of discadind them, it's a posibility that the engine didn't support that, to start with.
The other posibility is that the engine is really failing at deciding which frequencies to discard or not, which would show that they failed in the Psychoacoustic analizer implementation.

I'm not sure if this is all, but most of it.


Edit : The same goal applies to WMA, and most of its artifacts is because they push the format to the limits (that is, higher frequency than what it can reproduce without faults. They don't give any margin, and fails quite easily).

Why is Xing crap?

Reply #3
i think simply because their aim was to make the fastest encoder on earth. it is actually very fast but has poor quality. many people don't hear a difference between xing and lame but they all notice that xing is faster than lame.


Why is Xing crap?

Reply #4
For what it's worth, It doesn't look like Real has put any work into Xing since they bought it a few
years ago.  Even if others have improved since then, Xing has been static.

I see reference to "new" and "old" xing, but remember that the newest is still 4 years or so old.

Why is Xing crap?

Reply #5
I use to use Xing quite a lot since I couldn't tell the difference.    Maybe its because I only used it at VBR (75 or above), so with the higher bitrate, it was a bit harder to listen for artifacts.

Now I encode using Vorbis

Why is Xing crap?

Reply #6
Quote
I use to use Xing quite a lot since I couldn't tell the difference.    Maybe its because I only used it at VBR (75 or above), so with the higher bitrate, it was a bit harder to listen for artifacts.

Now I encode using Vorbis

I agree with you. Lots of xing vbr and higher bitrate stuff sounds OK to me.

But when it fails, it does so miserably.

And at 128 CBR, it sounds terrible on even slightly hectic music.

Why is Xing crap?

Reply #7
Quote
We all know Xing MP3 encoder sucks badly. But why does it suck so bad?
I mean, from a scientific point of view. The people that created Xing {snip} wake up one day and said "Hey, let's make the worst mp3 encoder
on the planet!"

Think Windows 1.0. Think symphonic TVs. Think packard bell a few years ago. Think of AOL and MSN as ISPs.   

At 256k, you can really notice how much it sucks. I didn't even need to have the original, or put headphones on, to say to myself "the high tones on this piano music sucks".

I, also agree it was about speed. I don't care about the speed of the encoder, if it sounds like that[/u].

Why is Xing crap?

Reply #8
My top priority is quality, not speed.
I'm a patient man 
Wanna buy a monkey?

Why is Xing crap?

Reply #9
I recently downloaded a copy of dbpoweramp and see that it will display the encoder used for a  mp3 or ogg file.  Is there an actual attribute in the file, or does dbpoweramp guess by the style of encoding?

Anyhow, I decided to dig out what I remember as being some really bad files I downloaded maybe 4
or 5 years ago to see what coder was used.  Survey says: Fraunhoffer. 

One noticable artifact was percussion that sounds like there were foil streamers  tied to the drumsticks (pre and post- impact swirly sounds).  This (and all the others I checked) were  standard 128kb/s
cbr.

I suppose it's possible these tracks went through a couple of encode/decode cycles ( maybe originally
recorded on atrack minidisk?)

Why is Xing crap?

Reply #10
To have an idea about how much Xing is bad just think that it produce a lot of artifacts even at 256kbps (with a 16Khz cutoff)..
[ Commodore 64 Forever...! ]