Transcoding source quality
Jul 5, 2009 at 10:35 PM Thread Starter Post #1 of 19

xicex

New Head-Fier
Joined
Feb 24, 2007
Posts
11
Likes
0
Will a 128kbps MP3 transcoded from a 320kbps mp3 will have better sound quality than a 128kbps MP3 transcoded from a 192kbps mp3?

Maybe this would sound better if I said, will 128kbps encoded from HD lossless super leet professional stereo quality sound source sound better than a 128 from a 128kbps mp3 transcoded from a mp3 file recorded analogolly from grandpa's radio.
 
Jul 6, 2009 at 2:57 AM Post #3 of 19
Quote:

Originally Posted by KONAKONA /img/forum/go_quote.gif
Well besides the fact that this is in the wrong forum,

Both the 128>320 and 128>192 will both sound worse than the original 128, so there isn't much of a point in it.



You should possibly try reading the post BEFORE replying.

I don't think it would make a difference what the source is. As long as the recording is the same, it still ends up as 128k.
 
Jul 6, 2009 at 4:10 AM Post #4 of 19
Quote:

Originally Posted by fraseyboy /img/forum/go_quote.gif
You should possibly try reading the post BEFORE replying.

I don't think it would make a difference what the source is. As long as the recording is the same, it still ends up as 128k.



I did read it, just not very well.
tongue.gif


There might be some differences in the encoder being used etc, but fraseyboy basically hit the nail on the head.

Or you can just use a higher quality.
 
Jul 6, 2009 at 4:36 AM Post #5 of 19
Higher source quality will generally give higher transcoded quality. Ideally going back to the original source is preferable. Most MP3 encoders have a high frequency cutoff at lower bitrates, so going back to the original CD is usually the best way to ensure the encoder is hearing the full frequency range of the recording (even if it then applies it's own high frequency filter subsequently) . The best encoder in the world cannot put data in a file if it cannot hear it in the first place. Personally I don't transcode except from lossless sources these days but I did try it once- the results were not encouraging to say the least.
 
Jul 6, 2009 at 5:33 AM Post #6 of 19
Any transcodering, except from lossless to lossless, will result in quality lost. Similarly, transcoding from a better lossy source will have better result then from a bad source.

Transcoding lossless is like cloning apple, you should expect everything to be the same as long as your cloning technique (software) is perfect.

Transcoding lossy, however, is more like squeezing apple for juice (essence). Even when two apple went through the first juicing process, the quality (sweetness, nutrition level, etc) of the apple will result in two glasses of different tasting apple juice. Any further (but identical) extraction process of these two glasses of apple juice will diminish the difference, as the more nutritious juice will lose more (simply because there is more to lose) while the less nutritious juice has nothing more to lose - however, both will probably taste bad after so many extraction processes. If you keep repeating the extraction process, eventually you will end up with two glasses of no-taste-whatsoever.
 
Jul 6, 2009 at 5:47 AM Post #7 of 19
Quote:

Originally Posted by ClieOS /img/forum/go_quote.gif
Any transcodering, except from lossless to lossless, will result in quality lost. Similarly, transcoding from a better lossy source will have better result then from a bad source.

Transcoding lossless is like cloning apple, you should expect everything to be the same as long as your cloning technique (software) is perfect.

Transcoding lossy, however, is more like squeezing apple for juice (essence). Even when two apple went through the first juicing process, the quality (sweetness, nutrition level, etc) of the apple will result in two glasses of different tasting apple juice. Any further (but identical) extraction process of these two glasses of apple juice will diminish the difference, as the more nutritious juice will lose more (simply because there is more to lose) while the less nutritious juice has nothing more to lose - however, both will probably taste bad after so many extraction processes. If you keep repeating the extraction process, eventually you will end up with two glasses of no-taste-whatsoever.



Might want to edit that to "Cloning an apple", as it might come off as cloning apple the company, which would turn this into a troll thread quite quickly.
tongue.gif
 
Jul 6, 2009 at 5:53 AM Post #8 of 19
I thought people will get more offended when they read the next line
biggrin.gif


Quote:

... squeezing apple for juice...


 
Jul 6, 2009 at 7:27 AM Post #9 of 19
Probably, since less audio data where thrown away during the first encoding (320kbps vs. 192kbps).
That said, transcoding lossy to lossy is not recommended. You are much better off starting fresh from a lossless source (ex. CD)
 
Jul 6, 2009 at 1:46 PM Post #11 of 19
i think we are just spouting a lot of hot air without real evidence other than theories. if you want a better idea of what happens to lossy to lossy trascoding, go to hydrogenaudio forums - that is their speciality. Our speciality is debating 1,3 or 6 driver IEMs.
 
Jul 6, 2009 at 2:00 PM Post #12 of 19
Quote:

Originally Posted by shigzeo /img/forum/go_quote.gif
i think we are just spouting a lot of hot air without real evidence other than theories. if you want a better idea of what happens to lossy to lossy trascoding, go to hydrogenaudio forums - that is their speciality. Our speciality is debating 1,3 or 6 driver IEMs.


Transcoding - Hydrogenaudio Knowledgebase

I happen to be a member of HA as well. As are quite a few others from this forum.

Edit: Either way. Transcoding is not recommended as it will further decrease the quality of the files. The best way is to transcode from lossless (the source) to lossy. I don't really see a good reason to transcode from lossy to lossy anyway.

Edit II: I don't know anything about whatever kind of IEM btw. I might just be the exception on this site though.
tongue.gif
 
Jul 6, 2009 at 3:39 PM Post #15 of 19
Quote:

Originally Posted by shigzeo /img/forum/go_quote.gif
i think we are just spouting a lot of hot air without real evidence other than theories.


Evidence as in posting results from a series of ABX test? Feel welcome! I am confident enough from looking at the bitrate of the source files..
 

Users who are viewing this thread

Back
Top