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: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101 (Read 5664 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Have I really freaked out, or is there a subtle sonic difference between the reference libFLAC 1.3.1 and 1.3.2 versions?

In many tests, version 1.3.1 to my ears sounds subtly more "open", "breezy", than 1.3.2, which I felt (subtly) more "warm", "closed".

Has anyone noticed this?

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #1 –
Quote from: Adil on
Has anyone noticed this?
Nobody has noticed this, because what you're describing is impossible. FLAC is a lossless format.

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #3 –
Quote from: doccolinni on
Yes.
Not really. The reason for blind testing is that this is not only normal - it is so hard to kick that we instead implement test protocols to circumvent it completely.

So experiencing this isn't weird. Spoiler (click to show/hide)

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #4 –
I've done conversions from a few files from different music genres (pop and metal) and in both cases I've found them to sound the way I initially described them. Today I did it again and the same happened.

I simply took any file compressed by version 1.3.2 and recompressed it in 1.3.1. With headphones the subtle difference becomes more evident.

Anyway, thanks everyone for the comments, I take the opinion of this community very seriously.

EDIT: and sorry for the topic, it won't happen again.

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #5 –
Quote from: Adil on
I've done conversions from a few files from different music genres (pop and metal) and in both cases I've found them to sound the way I initially described them. Today I did it again and the same happened.

I simply took any file compressed by version 1.3.2 and recompressed it in 1.3.1. With headphones the subtle difference becomes more evident.

Anyway, thanks everyone for the comments, I take the opinion of this community very seriously.

EDIT: and sorry for the topic, it won't happen again.
The same can happen all the time until you keep out the mental thing. You have more than 200 posts here and should have heard of proper blind testing.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #6 –
I still don't know how to do this, but I will try to supply myself with the necessary knowledge to do so (I only know how to use foobar2000's ABX test).

Thanks.

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #7 –
Quote from: Adil on
I only know how to use foobar2000's ABX test
That should be enough to compare two FLACs.

Also you can use foo_bitcompare to be sure that you haven't converted with any DSP settings. But if you just use the command-line flac --force filename.exe, it has nothing such to do.

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #8 –
I thought it was necessary to use some additional software that would generate graphics with objective results. The ABX test of foobar2000 is pretty cool, but I thought it might be tainted by an eminently subjective judgment laden with false impressions.

My first post on this topic reflects my judgment after using the ABX test. I hadn't thought of foo_bitcompare, though. This tip is worth gold.

The conversion setting I used was this: -s --ignore-chunk-sizes -3 - -o %d

I believe it is standard, with the exception of level 3, which I like to use.

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #9 –
The post of a 15 trial log would be nice.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #10 –
Which program generates such logs, foo_bitcompare?

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #11 –
Quote from: Adil on
The ABX test of foobar2000 is pretty cool, but I thought it might be tainted by an eminently subjective judgment laden with false impressions.
Why would you think that? It's purpose is exactly the opposite, to tell you if your impressions are false or not.

Quote from: Adil on
Which program generates such logs, foo_bitcompare?
No, the ABX comparator:
X

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #12 –
Thanks, good to know. I don't know how to interpret this log, but I'm willing to learn.

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #13 –
Quote from: Adil on
My first post on this topic reflects my judgment after using the ABX test.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #14 –
It was just a listening exercise and not a text (log) interpretation exercise.

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #15 –
Quote from: Adil on
I don't know how to interpret this log, but I'm willing to learn.
And have you clicked that link What is a p-value? :-)

It tells you the probability of getting the result that you got, or a better one, if you were simply guessing.

For example, if you get 6/10 answers right, it will show 37.7%. This means there is 37.7% chance of getting such result, or a better one, just by guessing. That's quite high chance, so such result will not convince anyone that you can distinguish the files.

If you get 9/10 answers right, it will show 1.1%. This means there is only 1.1% chance of getting such result, or a better one, just by guessing. If that chance is less than 5%, it is then commonly accepted that you probably were not guessing but actually heard a difference.

Also bear in mind, that this value is not "the probability that you were guessing", although it is a very common mistake to treat it like that. No, it is "the probability of getting the result, if you were guessing". A subtle difference :-) Calculating "the probability that you were guessing" is apparently hard, or maybe not possible, not sure, I'm not that well versed in statistics.

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #16 –
@danadam

Thanks. I still don't know how the component will tell me how right or wrong I am in my impressions, but your answer gives me guidelines for a good start to learning.

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #17 –
Over the years we had many extraordinary claims about lossless files sound different because of some magical behaviour.
Fuzzy impressions alone are not enough here on Hydrogenaudio and this thread is a good candidate for the recycler when it goes on in circles.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #18 –
Feel free to delete this topic and take the opportunity to banish me too.

Re: reference libFLAC 1.3.1 20141125 vs 1.3.2 20170101

Reply #19 –
Wombat keep it civil