IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
Mp3 bitrate relating to time/quality along with an interesting challen, need advanced theory help on this one
blanknix
post Sep 3 2010, 19:21
Post #1





Group: Members
Posts: 10
Joined: 3-September 10
From: Victoria, BC
Member No.: 83579



Hey I was wondering if anyone could firstly clarify whether mp3 bitrate refers to either the total number of 'snapshots' (like frames per second) that make up an mp3 is or whether the 'fps' is constant and bitrate just effects the amount of data in each snapshot (thats my current guess, would seem to explain that low bit washout sound).

The question came up while I was inverting waves to make DIY acapellas and I was wondering purely for kicks, if you had a 320 bit mp3 and a 128 bit, couldn't you technically invert the 320 into the 128, and then combine that imperfect phase with the orignal 128 to essentially make it 320? This has no practical application I'm just curious whether I could do it.


Thanks a bunch hydrogenaudio, I hope some other music nerds find this interesting smile.gif


Quick addendum, if bitrate does turn out to be related only to the amount of data, does that mean that the frequency rate controls the 'fps'?
Go to the top of the page
+Quote Post
pdq
post Sep 3 2010, 19:30
Post #2





Group: Members
Posts: 3442
Joined: 1-September 05
From: SE Pennsylvania
Member No.: 24233



QUOTE (blanknix @ Sep 3 2010, 14:21) *
Hey I was wondering if anyone could firstly clarify whether mp3 bitrate refers to either the total number of 'snapshots' (like frames per second) that make up an mp3 is or whether the 'fps' is constant and bitrate just effects the amount of data in each snapshot (thats my current guess, would seem to explain that low bit washout sound).

The number of frames of data per second is constant throughout the file, but the number of bits per frame can vary.

What do you mean by "low bit washout sound"?
Go to the top of the page
+Quote Post
db1989
post Sep 3 2010, 19:31
Post #3





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



QUOTE (blanknix @ Sep 3 2010, 19:21) *
Hey I was wondering if anyone could firstly clarify whether mp3 bitrate refers to either the total number of 'snapshots' (like frames per second) that make up an mp3 is or whether the 'fps' is constant and bitrate just effects the amount of data in each snapshot (thats my current guess, would seem to explain that low bit washout sound).
The latter.

QUOTE
The question came up while I was inverting waves to make DIY acapellas and I was wondering purely for kicks, if you had a 320 bit mp3 and a 128 bit, couldn't you technically invert the 320 into the 128, and then combine that imperfect phase with the orignal 128 to essentially make it 320? This has no practical application I'm just curious whether I could do it.
You can't get something from nothing.

QUOTE
Quick addendum, if bitrate does turn out to be related only to the amount of data, does that mean that the frequency rate controls the 'fps'?
Yes, the sampling rate determines the number of samples per second.

Your usage of the term frame is incorrect and in fact confusing. Video frames have no analogue in audio. In fact, in the context of compressed audio formats, a frame is a 'packet' of data of a specific bitrate.
Go to the top of the page
+Quote Post
blanknix
post Sep 3 2010, 19:39
Post #4





Group: Members
Posts: 10
Joined: 3-September 10
From: Victoria, BC
Member No.: 83579



pdq:
Remember early, low bitrate audio streams? I listen to a lot of electronic music so it's the snare hits specifically that sound terrible (blurred, washed) in low quality.

DV1989:
Right I know of course you can't just convert 128 to 320 and expect anything to happen, but inverting the 320 against the 128 (along with the rest of the procedure) would do something right? The extra data that the 320 would have that the 128 didn't would be left behind when the two cancelled each other out (what I meant by the imperfect phase) which you could then merge with the original 128 to fill in the gaps so to speak. Or am I misunderstanding how sound is stored as data?

I'm assuming that in a sample (now I know not to call it a frame, thanks) the sound is reproduced with the right frequencies and respective amplitudes, is this correct? If I'm wrong about that then I absolutely agree it'd be impossible to succeed with my experiment.

Thank you both for the fast intuitive replies, I understand a lot more already.

edit: reclarified that first sentence of my experiment thesis, typing too fast lol.

This post has been edited by blanknix: Sep 3 2010, 19:47
Go to the top of the page
+Quote Post
db1989
post Sep 3 2010, 20:06
Post #5





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



QUOTE (blanknix @ Sep 3 2010, 19:39) *
DV1989:
Right I know of course you can't just convert 128 to 320 and expect anything to happen, but inverting the 320 against the 128 (along with the rest of the procedure) would do something right? The extra data that the 320 would have that the 128 didn't would be left behind when the two cancelled each other out (what I meant by the imperfect phase) which you could then merge with the original 128 to fill in the gaps so to speak. Or am I misunderstanding how sound is stored as data?
Yes to that last sentence.

QUOTE
I'm assuming that in a sample (now I know not to call it a frame, thanks)
Nope. A sample is not equal to a frame. A sample is one point on the audio wave, and any digital audio stream consists of a constant number of samples per second. Compressed audio encoders encode said stream as a succession of packets of data called frames, of which there are also a constant number per second (about 38 in the case of MP3). These frames may be of identical sizes, i.e. constant bit rate (CBR), or of different sizes appropriate to the complexity of the audio encoded therein, i.e. variable bitrate (VBR). They typically store audio that has been converted from the time domain (sampling points on a long wave) to the frequency domain (a crude analogy would be the kind of spectrum you can view in Audacity), but which in any case is equivalent to the same number of samples for every frame (typically at the same sampling rate as the input file).

QUOTE
is whatthe sound is reproduced with the right frequencies and respective amplitudes, is this correct? If I'm wrong about that then I absolutely agree it'd be impossible to succeed with my experiment.
I have no idea what you mean, and I don't think your 128 vs. 320 kbps idea is anything worth calling an "experiment" or a "thesis". I personally can't offer much more of a reason why than "It just doesn't work that way."

This post has been edited by dv1989: Sep 3 2010, 20:08
Go to the top of the page
+Quote Post
blanknix
post Sep 3 2010, 20:17
Post #6





Group: Members
Posts: 10
Joined: 3-September 10
From: Victoria, BC
Member No.: 83579



Thanks DV! I understand now. Raw audio in digital format is stored in samples, the frequency of which occur at whatever frequency rate. A sample is just a sample in a lossless audio format but it is stored in a frame if the format is compressed. I knew about the nature of CBR, VBR already but now I understand the context so thank you for elaborating on that.

Now I just need to find out exactly how music is stored in a sample and I'll have a decent basic foundation for knowledge I'd say!

I'm going to go try my experiment now, I will post the results back here in a few minutes!

thanks again hydrogenaudio, I heard good things about this forum and they were all true.

This post has been edited by blanknix: Sep 3 2010, 20:49
Go to the top of the page
+Quote Post
db1989
post Sep 3 2010, 20:21
Post #7





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



How music is stored in a sample? It's simpler than you appear to think.

Think of a sound wave. Then think of a digital device sampling this wave at regular points.

The dots are samples. That's it. The more samples per unit of time, the more faithful reproduction of the input analogue signal, and the higher the maximum frequency of the resulting digital signal.

This is very basic digital audio stuff; you should probably read a primer on the basics before putting forward "theses" or suggestions for experiments.

This post has been edited by dv1989: Sep 3 2010, 20:25
Go to the top of the page
+Quote Post
blanknix
post Sep 3 2010, 20:24
Post #8





Group: Members
Posts: 10
Joined: 3-September 10
From: Victoria, BC
Member No.: 83579



DV that is exactly what I tried to explain as my understanding of how music was stored and YOU told me I was wrong (I had said sound was recreated via stored frequency and amplitude).

This has been extremely helpful DV, don't get all internet on me now.

And its worth mentioning that I always google first, and I sort through all the unrelated bullshit as well before coming to a forum. I was just on a roll with a few of these questions and I wanted clarification from an expert, not just some crackpot article writer.

This post has been edited by blanknix: Sep 3 2010, 20:26
Go to the top of the page
+Quote Post
db1989
post Sep 3 2010, 20:29
Post #9





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



I'm by no means an expert; that's my point. The concept of sampling (rate) is so fundamental to any understanding of digital audio that (I'd hope) almost everyone else here could explain it better than I.

This post has been edited by dv1989: Sep 3 2010, 20:30
Go to the top of the page
+Quote Post
blanknix
post Sep 3 2010, 20:39
Post #10





Group: Members
Posts: 10
Joined: 3-September 10
From: Victoria, BC
Member No.: 83579



Holy shit I was right all along! My experiment totally worked. DV I'm gonna take a wild guess that you have never tried to extract acapellas by inverting waves in perfect phase to isolate the differences. Here's how to recreate what I was saying:

1. Export a 320 bit sample*
2. Export the exact same bit of sound in 64 or whatever bitrate you can audibly identify as being inferior
3. Import both into your DAW of choice**
4. Invert the waves of your 320bit track, you will hear ONLY the difference between the two songs, basically everything the 320 has that your low qual copy doesn't.
5. Introduce another instance of your 64 bit track, it will instantly sound like the 320 bit!

*I'm a producer so I just rendered a quick clip from a track
**if you attempt to save your imperfect phase as an mp3 in audacity, make sure your exported track is the EXACT same length as your two sources or this effect does not work

So basically I was right DV, I have a lot of your respect for your knowledge but don't insult my intelligence again okay? I'm just here to learn and explore.

If anyone wants I might be able to produce some physical evidence/samples of this if I can tackle that audacity trimming problem or recreate it in another host. Of course if you just follow these steps properly inside one instance of audacity you won't have any trouble.

This post has been edited by blanknix: Sep 3 2010, 20:42
Go to the top of the page
+Quote Post
db1989
post Sep 3 2010, 20:42
Post #11





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



Haha. I'm done here.
Go to the top of the page
+Quote Post
saratoga
post Sep 3 2010, 20:44
Post #12





Group: Members
Posts: 5116
Joined: 2-September 02
Member No.: 3264



So you added something, then removed it, thus proving that if you remove things they're not there anymore? Whats the point?
Go to the top of the page
+Quote Post
blanknix
post Sep 3 2010, 20:46
Post #13





Group: Members
Posts: 10
Joined: 3-September 10
From: Victoria, BC
Member No.: 83579



Saratoga, if you read my posts I said waaay back that this had NO practical implication whatsoever. I was just proving it could be done after being told it couldn't.

Guys please don't start flames!! This was never my intent!

This post has been edited by blanknix: Sep 3 2010, 20:46
Go to the top of the page
+Quote Post
db1989
post Sep 3 2010, 20:49
Post #14





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



What I said you couldn't do was to create quality out of nowhere, which I admit may have stemmed from a misreading on my part (not that you were particularly clear), i.e. interpreting your post(s) as meaning that you had thought of a way to 'upgrade' a 128/64/whatever kbps MP3 to a 320 kbps one.

In any case, all you have demonstrated is that 1+2-2=1.

This post has been edited by dv1989: Sep 3 2010, 20:52
Go to the top of the page
+Quote Post
Synthetic Soul
post Sep 3 2010, 21:30
Post #15





Group: Super Moderator
Posts: 4887
Joined: 12-August 04
From: Exeter, UK
Member No.: 16217



Less constructive posts moved to the recycle bin.

Please keep it civil, or the thread will be closed.

Thanks.


--------------------
I'm on a horse.
Go to the top of the page
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 24th November 2014 - 17:13