[peel] Re: techy. May I borrow your ears?

Alasdair Macdonald wewalkforonereason@...
Mon Mar 24 16:39:34 CET 2008


On 24/03/2008, Roger Carruthers <roger.carruthers@...> wrote:
>
>    I beg to differ: there is logic to my argument. The object of the exercise is to get the best quality sound , for the minimum file size.

And that is a subjective judgement.

My point was - and is - that there is a place for lossless and there
is a place for lossy. The considerations are different for each, but
there are no absolutes in lossy because everyone hears things
different. And if the issue with lossless is a consideration of hard
drive usage, then frankly that is also subjective.

>
>   Before I was a sound engineer I studied as a welding inspector, where we were taught that quality is defined as 'fit for purpose'; don't buy a Roller just to run the kids to school.
>   Ogg, like .mp3 uses perceptual encoding. Nobody is pretending that the signal remains unchanged, but if you use a decent quality encoder at the right settings, then there is no loss of quality  that you can hear.
>   As a sound engineer I was taught to trust my ears. If it sounds right, then it is right.
>   If you can store twice as many files, and you cannot hear the difference, then you are achieving quality - your recordings are fit for purpose.
>
>  This will always be a subjective issue – if you can hear the difference, then go with your ears. But the original poster said:
>  "I have personally found that by using the dbpoweramp prog with OGG at 350kbs giving the 1:4 that there doesn't appear to be any difference from the original."
>
>  As for the Betamax analogy, it just doesn't work like that in the digital world; even if it doesn't become a major player, Ogg support will never disappear entirely. It's true that no-one makes Betamax machines any more, but the code for playing Ogg Vorbis files will always be out there.

Well, we won't know if the analogy works until we wait a few years.
The point is valid, in that not all players that are capable of
playing mp3 files are also capable of playing ogg (or even mp2 files).
Right now mp3 is by far the most common format.

>
>  The purpose of this lboard is for the discussion and appreciation of Peel's craft, and lately it has become a place where people share recordings of his shows. If we keep the files to a sensible size, whilst keeping an eye on quality, we can share them so much the more equably.
>
>   It just grieves me to hear, as someone did here recently, that you should use 24/96 conversion for recording of cassettes, which will do nothing to improve the quality (think about it – the bandwidth & dynamic range are not there to record!), but the results will take many times longer to download, and thus reach less people.

Agreed - 44k or 48k probably are the most sensible starting points for
this kind of source.


>
>   If you have unlimited drive space/attic space and you only intend the files for your own use, go for it – use lossless compression. But if space is an issue, and you also want to share the wealth, then high bitrate .oggs are going to hurt no-one, and probably help a few.

If I had "unlimited" attic space I wouldn't be shitting myself about
the amount of cassettes, vinyl and CDs that I own. I would love to put
it all onto hard drive or solid state memory - I've calculated it out
to a block of hard drives (each drive mirrored) that would fit into a
single PC case. If I could do that easily I would shell out the ££ for
the drives today - but the problem is the conversion time, not the
number of hard drives, their cost, or how much space they occupy. And
I would no more dream of reducing the quality of the CDs I bought
yesterday, last week, or 20 years ago than I would dream of reducing
the quality of either my best or my worst FM masters.


>
>  On 23/03/2008 21:49, "Alasdair Macdonald" <wewalkforonereason@...> wrote:
>
>  > On 23/03/2008, Roger Carruthers <roger.carruthers@...m> wrote:
>  >>
>  >>    To be honest, I think that .ogg at 1:4 is preferable to lossless
>  >> compression at 1:2, because the object of the exercise is to get the best
>  >> quality sound into the smallest file size, and if you can't hear the
>  >> difference (and not many people of our age can – myself included, and I have
>  >> 'trained' ears) then the smaller file size wins.
>  >>   As I've said before, the vast majority of the material we're talking about
>  >> here was recorded to cassette, and you can't improve upon the quality of the
>  >> first generation; as you're starting from a fairly limited bandwidth and
>  >> dynamic range, you're pissing in the wind with lossless compression. In
>  >> short, go .ogg!
>  >
>  > There is no logic whatsoever in your argument. You seem to be saying
>  > the source is substandard, so making the closest possible digital
>  > version of that source is pointless.
>  >
>  > And your alternative is a knowingly degraded version.
>  >
>  > Digital compression slits into two camps - lossless, and lossy. The
>  > purposes of each type are in general rather different.
>  >
>  > Typically, an archivist will preserve a lossless version using a
>  > compression format that fits their needs - ie portability, speed of
>  > compression / decompression, and compression ratio.
>  >
>  > Those who prefer to make lossy copies - usually for personal use, not
>  > for public archiving - have one additional consideration - a trade-off
>  > - that of file size vs quality. And that's a very personal choice,
>  > which probably depends upon the abilities of the person's ears.
>  >
>  > It's my understanding that ogg [vorbis] *is* better a better quality
>  > encoder than mp3 - ie it produces a higher fidelity output for the
>  > same filesize, but that it suffers the same problem as Betamax - it's
>  > inferior relative is far more visible to the marketeers.
>  >
>  > It's worth bearing in mind the continuing fall in the price of
>  > storage. Hard drives of over 500GB can be bought for £60ish these
>  > days, and by the time you need to purchase the next drive the price
>  > will have dropped again.
>  >
>  > ------------------------------------




More information about the Peel mailing list