Converting Quality

[sai]

Re: Converting Quality

You may find this useful. Many users save files at a higher rate for home use and at a lower rate for portable use. Some prefer to make backup copies in lossless format (which you can do using using dB CD Writer if you don't want to use up hard drive space but are ok with the layout for cd-rom discs) to maintain highest quality backup.

Will the new converted lossless file maintain lossless properties with the exception of converting the file into a lossy again?

Your advice is practically a problem relief for me. Thanks.

[GSV3MiaC]

Re: Converting Quality

Will the new converted lossless file maintain lossless properties with the exception of converting the file into a lossy again?

Yes, lossless files do exactly what it says on the box. You can convert from CD (.cda) .wav to/from .flac, .ape, .WMA 9.1 lossless, zipped .wav, or whatever you like, and you'll keep getting the same answer (plus or minus the data tags, which are not present in some formats). If you don't then the format was not lossless.

When you convert to lossy format (MP3, AAC, .OGG Vorbis, whatever) you lose information. Mostly you won't hear what you lost, unless you convert to something with extreme compression (<=128kbps MP3 for instance), but you lost it anyway. Every time you convert to a lossy format you lose some =more= (even 128kbps MP3 to 128kbps MP3 may lose more information). So avoid doing it if you can.

Having said that, I cheerfully archive .Ogg Vorbis Q7 (less than half the size of lossless) and downconvert that to .Ogg Vorbis Q0, .WMA two pass VBR 48kbps, or (spit) 128 kbps MP3, for use on portable players. These are all =so much= worse than the Q7 that it isn't really a problem.

[xoas]

Re: Converting Quality

Will the new converted lossless file maintain lossless properties with the exception of converting the file into a lossy again?

A lossless copy will be a perfect COPY of whatever audio it is a copy of. So a lossless copy of a CD will be a perfect quality copy of the CD. If you have converted the cd to mp3 (or other lossy format) and if you make a lossless copy of the resulting mp3 files you will have a perfect copy of those mp3 files but you will not have a perfect copy of the cd (you will still be missing whatever was lost in the conversion to mp3).

Serious archivists, at least those who post here, tend to prefer to convert first to a lossless format for archiving and then to a lossy format.

Personally, my habits are more like GSV3MiaC. But please feel free to follow your own path.

Best wishes,
Bill

[sai]

Re: Converting Quality

I see...

Hmmm. Is there any noticeable drawback when converting a mp3 file into its original mp3 format (with one lower level bitrate) once?

ie. 192kbps mp3 => 128kbps mp3

If so, what type of flaws does the converted mp3 file sustain based on your observations?

[xoas]

Re: Converting Quality

Hmmm. Is there any noticeable drawback when converting a mp3 file into its original mp3 format (with one lower level bitrate) once?

I would encourage you to try it yourself and see. It has been a long time since I did a serious listening test on bitrates and lossy format comparisons. I definitely recall noticing that I could no longer differentiate between quality once the bitrate hit around 220 to 240 kbs for mp3, Ogg and Musepack. I do not recall the precise nature of those differences.

I also recall that these tests (I was comparing formats as well as bitrates) got to a point where they were interfering with my enjoyment of the music itself. It was definitely a worthwhile experience but it is not something I would want to do on an everyday basis.

I would also tell you that I do have files that are mp3 encoded at 128 kbs and that for everyday listening they are fine. I do not normally hear such a file and wish it was encoded at a higher bitrate. Because normal listening, you enjoy the music. Just like if you hear a song on the radio, you wil usually enjoy the song rather than note to yourself how much better it would sound straight from CD.

Best wishes,
Bill

[sai]

Re: Converting Quality

Yes, I suppose you're right. The entire incentive for obtaining higher quality songs is for enjoyment. I guess a song is a song no matter what the bitrate is. As long as it sounds decent, I suppose it's good enough to be heard. Hopefully, soon I will loose my significant interest in maintaining high quality songs and start enjoying the rewards. :smile: Thank you all for your comments and replies.

My respect,

Sai

[xoas]

Re: Converting Quality

Hopefully, soon I will loose my significant interest in maintaining high quality songs and start enjoying the rewards.

I hope you don't lose your interest in maintaining high quality songs.
I hope you have been guided to the point where you can do the legwork you feel is necessary to make your own decisions on these matters. I think this is important because you will, if you talk music and formats and stuff with others, you are going to run into all types of advice and information from people whose interests, tastes, equipment, and needs are different from your own. Their advice is not necessarily going to be right for you but you won't know that for sure unless you do the work. So I hope you have gotten here some good information about how to determine what is right for you and your needs.
And don't be scared from doing a little hard listening to help you decide. Just be forewarned that it may turn out to be harder than it seems and be prepared to step away when it is time.

Good luck and best wishes,
Bill

[elsabio]

Re: Converting Quality

It ought to be possible, I would have thought, to predict what sort of quality damage would be done to the signal when converting between (lossy) formats or between different bitrates. Has any work been done by analysis of the encoding and commpression algorithms concerned, or by information theoretical approaches that would give any clues here?

I have a small vested interest: I have some RealAudio files that I would like to migrate to a portable player. I use Ogg Vorbis most everywhere, but converting the RealAudio to Ogg Vorbis at my usual quality setting turns a 30M RealAudio file to well over 100M. The question is: How far can I reduce the quality before I begin to notice? What would I notice? Would this be per sample, thus making batch conversion unreasonable?

However, my interest is as much in the academic as it is in solving thie one little problem!

[xoas]

Re: Converting Quality

Interesting questions. I am sure you can find some work on the technical aspects of signal degradation. I would advise you to start with hydrogenaudio.org. They have a vast repository of information about different formats and their performance.
Although you would imagine that this might involve relatively straight-forward analysis of waveforms and such, from a practical standpoint you are better off going with what your ears tell you. This is because waveform patterns do not necessarily conform in any simple way to your hearing experience any more than a CT scan or eeg would be a good measure of what you are thinking or feeling.

The issue is more complex in the case of converting between lossy formats, in part because different lossy formats have somewhat different models of what information to toss away. However, to further compound the issue, these differences are likely to be more apparent with certain types of files than with others. A recorded lecture could easily be saved at far lower frequency and bitrates than a piece of classical music (which has a lot more signal complexity to try to maintain).

Likewise, with your problem, your ears alone will be the very best judge of your best conversion setting if you convert your Realaudio files to Ogg. I would urge you to look at one of your Realaudio files properties in terms of bitrate, frequency, channels. Try making some copies of this file to Ogg at different bitrates starting with the same bitrate as your RealAudio files and then converting to the one or two settings just above and just below your RealAudio files. Compare these. If necessary, go a bit further in either direction (although I am fairly certain that you won't need to go up).
The good news (if it is such) is that your RealAudio files are not likely to respond much to higher Ogg bitrates because it is a lossier codec than is Ogg (and much like mp3pro, it is designed to be that way-its focus is on compression and being able to provide streaming content rather than on preserving optimum audio quality). An exception might be the case of Real Lossless although then I doubt you'd have any greater file size converting to Ogg.

Those arer my thoughts. Hope they help. I'd be interested in hearing from you or others about their thoughts and findings. BTW, if you aren't already using the Ogg aoTuV b3 or b3 SSE codecs, you should rerally give them a try.

Best wishes,
Bill

[elsabio]

Re: Converting Quality

Thanks for the pointer - looks like I have lots of reading to do.

The good news on my RealAudio -> ogg requirement is that it is only for a portable player. (I capture radio concerts/programmes from rtsp:// streams that I've missed so that I can listen to them on my all-too-frequent-flights across the Atlantic.) Therefore, the quality is not that important, for the most part. The bad news is that these are (broadly speaking) classical music. However, the idea just got me thinking about the broader issues of format conversion.