Hi all.
I'm working on a project about Ogg (no, I'm not doing my own AoTuV, don't expect a new encoder ), I modified "something" and I need to know if I did something wrong

Would you please give a shot to these pairs of files and tell me if you hear any differences, if you hear some artifacts and so on?
You don't need to listen to all of them if you don't want to, just compare some random parts of some files

Here are the files (if they don't work, copy and paste in the address bar):

1a
1b

2a
2b

3a
3b

4a
4b

TIA, and sorry for my bad english

You may not get many responses for a while; I think many people interested in listening tests will be busy doing the public multiformat 64 kbps one at present - I know I for one will be! After that, if time permits I might be able to have a listen, but spare time is quite rare for me now we have a baby (sorry).

Without the original to compare to, how can we decide which sounds "better"?

@rockcake: No problem, I understand that there are surely more important list. tests than mine, I'll wait

@pdq: one of the two is the original one, guess which one

Welcome to the forum psychowood.
What pdq meant was that you also need to provide original wav samples which you encoded using Vorbis. That is because it is expected that both files that you provided have some distortions. For a listener to determine where there are less distortions, original without any distortions is needed.
Example situation: some high frequency noise exists only in one file. How a listener would know if that noise also existed in original or not?

please select what best describes what i see

blue or reed

how would people know what is best without knowing what they are trying to describe ?

The only thing coming out of this is what colors people like the most.


let me try again with a music example

which of the music files its encoder best

encoding a: bestie boys intergalactic
encoding b: fur elise


again without know if you did encode fur elise og bestie boys intergalatic.
we would not know which encoded file is best.


and btw if you encode bestie boy and got fur elise you got a pretty bad encoder :-)

He stated that one of the files is the reference one, but hidden. Ah sorry, thought he meant reference with "original".

Hello muaddib, and thanks for the welcome



I understand what you mean but, in this specific case, I'm not doing a comparison between two different methods of encoding: I'm working on a steganography project, and I'm trying to figure out if (slightly) editing a vorbis stream introduces hearable distortions or not, because I was rather surprised not to found any hearable one...

I mean, if there are some high frequency noise, they should be present in both encoding because I used the same encoder, and if they are not, then just say "the version X does not have high frequency noise and I think it sounds better"

@sven_Bent: you have a point there, but I'm not doing an encoder comparison
Let's say we have two copies of Pablo Picasso's Guernica, one untouched and one on which I did some (very very small) scratches: what I'm asking is if you can guess which is the one I scratched. I didn't tell it in advance because I didn't want people going around looking for scratches instead o simply comparing the two pictures ^^

Then maybe it would be better to rephrase your question from "which sounds better" to "do you hear any differences".

Yep, better choice. Edited, thanks

May ask you out of curiosity: Is the steganography applied before encoding or during encoding (i.e. with a modified encoder)?

After encoding

I tried working on a lossless steganography like in mp3stego (which uses packet lenght parity to store bits), but I couldn't figure out how to increase a vorbis packet lenght of a single bit... so I worked on changing the parity of the vorbis packet itself (instead of its size), and it seems to work quite well: up to now I haven't found a single person which could hear a difference between the two versions. Then I posted here on HA looking for some golden ears