Hi there ,

I am entering the world of digital mixing after years of vinyls and I have questions about how formats interacts against each others once edited , I hope you can help and the questions ain't too silly .

As I have mp3's of various quality as well as other formats I wonder how the conversion from one to the other sounds like

For instance , let's say I take an OGG track that I find too sound a bit thin , I process it within ableton or a wave editor to maximise the gain or add some eq but if the software can only export in another format than the original one , how much loss is there ?

Ideally one should always convert to the very same format as the original unless the formats are uncompressed (FLAC/WAV etc etc ) , am I right ?

That could also be the case for time stretching odd tracks ; not in a drastic way ( within a 5bpm range maximum, more likely 1 or 2 in order to tighten it )

I understand there's only a slight amount of processing that can be applied to ehnance poor quality files but any advice / recommandation would be usefull and appreciated . Cheers !

F

The wave editor will convert the ogg (or mp3) file to a normal waveform. This conversion won't cause any loss of quality (the loss of quality has already happened when the original wave was converted to ogg, and this loss can never be undone).

Once you have finished editing in the wave editor, you can then convert the wave into any format you like. There will be another loss of quality when you do this conversion (assuming the conversion is to a lossy format). But you don't have to convert it to the same format as before. You lose quality whatever lossy format you convert to, whether the same format as before or another one.

I suppose it might be said that if you convert to a different format than before then you will now have two types of artifact -- eg. slightly audible and unwanted effects of the lossy conversion -- in your audio, so it's better to convert to the same format as before because then you will just have the same sort of artifact that you had previously. Then again, it might be said that compressing to the same format twice will accentuate the artifacts that were there before, so better to convert to a different format so that even though there are more types of artifact, none are really loud enough to be problematic. Or will converting to the same format twice just give you the same level of artifact as the first time? I just don't know about this (Hugh might). All I can say is that you should use the highest quality files and conversion rates that you can - for example converting to and from a 256k mp3 will give vastly superior results to converting to and from a 128k mp3.

Edited by Scramble (04/02/12 12:10 PM)

Scamble has pretty much said all that needs to be said.

Any processing in a DAW will inherently be in a linear format (wav etc), but processing may not always behave the way you hope or expect. For example, if the original lossy encoding has thrown away subtle high-end harmonics (as it's likely to do) adding HF boost won't be able to bring them back, it will just emphasise any louder components in that frequency region that were retained.

If the modified file is then resaved to a lossy format more data will be thrown away and the quality will decrease again -- much like copying from one cassette to another! It's called 'concatenation loss' and it sounds pretty grim because the artefacts are so unnatural.

Staying within the same family of lossy codec is probably a good idea as the degradation tends to be slightly slower than if you mix and match different codecs, largely because the filter bands line up nicely!

But the best principle is to treat lossy codec files as end-user replay-only files. That's what they were designed to be used as. If you want to process the audio, go back to the source and start again!

hugh

--------------------
Technical Editor, Sound On Sound

Quote Puzzled:

---

For instance , let's say I take an OGG track that I find too sound a bit thin , I process it within ableton or a wave editor to maximise the gain or add some eq but if the software can only export in another format than the original one , how much loss is there ?

Ideally one should always convert to the very same format as the original unless the formats are uncompressed (FLAC/WAV etc etc ) , am I right ?

---

No. Quite the opposite!

Don't confuse "compressed" with "lossy". FLAC, ZIP, RAR etc. compress the data, but you get back exactly what you put in. MP3 compresses size much more, but you lose quality.

When you open (say) an MP3 in an editing program it will be converted to WAV. You don't get back the quality that was thrown away at the initial MP3 conversion. But, if after working on it, you then convert to MP3 AGAIN, you will lose MORE quality.

If you're going to work on a file, keep it as WAV (or one of the lossless compressed formats). When you're sure it's finished, if you need to send it to someone and the WAV file is too big, degrade it to MP3 if you must. (But keep the WAV version - hard drive space is cheap!)

Thanks for the advice that's pretty much what I found out but I thought I might just been missing something out .

"If you're going to work on a file, keep it as WAV "
That's what I do most times and I don't have to do it too often but sometimes you need audio to be as tight as the day is long and that's when it can be confusing . Well not anymore ! And you're somewhat right , I don't mean compressed as in "use of compression" but lossy . It just got lost in translastion . cheers

Thanks Hugh I know what you mean by "earable" artifacts . All advices above and below are confirming my thoughts on the matter but it's always good to hear from others to find out more . I must say, to give some sort of credits to lossy formats , that their overall quality has been greatly enhanced over the last few years but let's face it below a certain rate it's unbearable ! Which makes me wonder why Audacity (for instance )gives you the option to convert to a whooping 64kbps ! mind you it might produce some interesting noises!