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: Re-encode from 128 kb/s to 64 kb/s (Read 15156 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Re-encode from 128 kb/s to 64 kb/s

Maybe it's pretty obvious just I do not get it?

I used the usual method: downloaded the WEBM from YouTube, demuxed the OGG which si 128 kb/s as usual.

I used fre:ac with default settings. OGG quality set to 0, that gives you 64 kb/s output. (Enough for a speech)

On a side note I always wanted to ask if you re-encode with the same encoder but lower bitrate how is it different from re-encoding with a different encoder? So it's not like making a WAV first, then simply re-encode that, but simply 'taking out some bits', so to speak, so the result can be better? As in my example; an overall better quality 64 kb/s file for spoken word audio.

Re-encode from 128 kb/s to 64 kb/s

Reply #1 –
Hi Encoder
Quote from: encoder on
I used fre:ac with default settings. OGG quality set to 0, that gives you 64 kb/s output. (Enough for a speech)

Well Ogg Vorbis is always VBR (variable bitrate) and adjusts the bitrate to the complexity of the audio material. So if you encode something with q0 it might get a higher (or even lower) bitrate that 64 kbit/s. In fact 64 kbit/s is only the approximately average bitrate if you'd encode a large number of different music.

Quote
On a side note I always wanted to ask if you re-encode with the same encoder but lower bitrate how is it different from re-encoding with a different encoder? So it's not like making a WAV first, then simply re-encode that, but simply 'taking out some bits', so to speak, so the result can be better? As in my example; an overall better quality 64 kb/s file for spoken word audio.

No, with most codecs (MP3, AAC, Vorbis etc.) there's no way to simply reduce the bitrate by 'taking out some bits'. The result is the same like of you'd make a WAV first. About quality: I think some people even say that it's better to use different codecs because it creates less notable artifacts. But I don't know if this has been proved by ABX tests.

Re-encode from 128 kb/s to 64 kb/s

Reply #2 –
Ogg Vorbis has a feature called "bitrate peeling" which allows creating a lower size variant of the file without re-encoding and dropping insignificant bits away, but it was never implemented in any relevant encoder. Also as far as i remember initial tests has shown that eg. a 128kbps Ogg file peeled to 96kbps is far not identical quality-wise with a file which was encoded directly to 96kbps from the source PCM.

Re-encode from 128 kb/s to 64 kb/s

Reply #3 –
OP: Re-encode from 128 kb/s to 64 kb/s, Size went not to 1/2 but to 2/3?

I still not get this one. OK, I ask it in another way:

I had some 64 kbps, mono OGG files, I normalized them with Audacity, created a WAV, now want to re-encode them to 64 kbps OGG again with fre:ac, that is, quality=0 (http://wiki.audacityteam.org/wiki/OGG; there is no mono/stereo switch in fre:ac), the file size will shrink from the original, the bitrate will show as 48 kbps mono. Why?

Re-encode from 128 kb/s to 64 kb/s

Reply #4 –
It's my intuition from experience that every time a file is encoded in a lossy format, a certain amount of artifacts/distortion/quantization noise(?) is added to the resulting lossy file.  It's not normally audible except sometimes is audible as cymbals and other sibilant sounds sounding wrong and such.  So when you re-encode a lossy file, it's now got these artifacts to encode as well, and they are usually high frequency artifacts, so the secondary encoding can't likely get the results down to 50% filesize.  I hope this is understandable.  Guys, correct me if I'm wrong.
Be a false negative of yourself!

Re-encode from 128 kb/s to 64 kb/s

Reply #5 –
You should be able to normalize ogg vorbis without any quality loss with Vorbisgain(!). This tool will not decode and then re-encode, but it will "simply" change the volume "gain" for each frame.

If you want to reduce bitrate, then i would recommend converting to mono first (you don't need stereo for speech) and then just try with -q setting is still acceptable for you. In my opinion you should not go any lower than q0. q-1 (yes, minus one) sounds quite bad. It seems you are already using q0 and mono.

Personally i like aac and even mp3 more than ogg at lower bitrates. I use lame version 3.97 with -V9 (mono) for my audiobooks. Newer versions of lame seems to be worse at lower bitrates.

Re-encode from 128 kb/s to 64 kb/s

Reply #6 –
I also want to throw in there that if your audio player support opus, it has a mode specifically for speech. You can likely maintain your expected quality level by pushing that bitrate much lower using opus instead of AAC, Vorbis, or mp3.

Re-encode from 128 kb/s to 64 kb/s

Reply #7 –
@Mark7

Thank for getting VorbisGain to my attention!

Quote from: encoder on
OP: Re-encode from 128 kb/s to 64 kb/s, Size went not to 1/2 but to 2/3?

I still not get this one. OK, I ask it in another way:

I had some 64 kbps, mono OGG files, I normalized them with Audacity, created a WAV, now want to re-encode them to 64 kbps OGG again with fre:ac, that is, quality=0 (http://wiki.audacityteam.org/wiki/OGG; there is no mono/stereo switch in fre:ac), the file size will shrink from the original, the bitrate will show as 48 kbps mono. Why?


OK, the above might not be 100% correct, because some of the files I normalized were actually MP3s, but still; 64 kbps should be around the same file size, regardless of file format, shouldn't?

The bottom line is, I still not get it why my OGG with quality=0 became 48 kbps, not 64. Thanks.

Re-encode from 128 kb/s to 64 kb/s

Reply #8 –
Quote from: encoder on
The bottom line is, I still not get it why my OGG with quality=0 became 48 kbps, not 64. Thanks.


(You meant "84" where you wrote "48"?)

The explanation is that this quality isn't "64". It is "likely around 64 on average". That means that some material will be below 64, and some will be above 64. You happened to have the latter.

Why isn't it precisely 64? Because it targets quality, not quantity. It makes a well-educated guess of what requires more and what requires less. And that is why it has a tag as "0" rather than "64". It is just a grade, not a bitrate.

Re-encode from 128 kb/s to 64 kb/s

Reply #9 –
Well, q0 will result in 64kbps for stereo files, but when encoding mono files, it looks like ogg vorbis q0 will encode to around 48kbps.

Just play around with the settings until you get 64kbps for these mono files, i guess q1 or q2 will do the job.

Re-encode from 128 kb/s to 64 kb/s

Reply #10 –
Ah, mono!

But the subject title should be "1/3" rather than "2/3", or did the OP try two different operations?