High Definition/Lossless Formats?

B

Booger

Weaksauce
Joined
Dec 29, 2005
Messages
90
I have a relatively large collection (150-160gb) of MP3's that have come from various sources (at various levels of quality) and I've decided to finally bite the bullet and re-build my collection by ripping my own physical media.

My initial budget is about $2k for media and I'll build from there. Is there a high definition format out there worth buying? Is there one coming? From my searches, it seems like SACD CD's cannot be ripped. As for DVD-Audio, it can be ripped, but it's complex and time consuming, correct? Are those my only options for lossless, multi-channel sound?

If you had to start from scratch, what type of media would you buy? Right now I'm leaning toward standard CD's ripped to FLAC or WMA lossless, but that's been the standard for quite a few years now. I find it hard to believe that something better isn't coming along soon.
 
There are no SACD-ROM drives currently, but all DVD-ROM drives can rip DVD-Audio - unfortunately the tools for doing so are hard to find and also hard to use, and as you say, they're time consuming.

However, most DVD-Audio discs have a DVD-ROM zone which contains a 5.1 DD or DTS track, but most importantly most have a 2.0 PCM Track also, which is higher quality than CD (48kHz, 16-Bit rather than 44.1kHz, 16-Bit) and still lossless. So if you want to guarantee your purchases for the future, you could buy DVD-Audio discs that you know have PCM sections and then you can rip them (this would involve using a standard DVD-ROM audio ripper which are plentiful), and then perhaps rip the DVD-Audio layer when it becomes easier to do so.

I'm not actually sure how common 2.0 PCM sections are on DVD-Audio discs though, I only have one lol, so maybe have a look on Play.com, Amazon.co.uk etc. to see how common it is.
 
I believe that, typically, 2.0 mixes on DVD-Audio discs are still packed via MLP (losslessly, of course). The process of ripping is then pretty similar as the 5.1 mix.

The format issue with 5.1 is a little tricky, but WMA is a good route to go with that (though certainly not the best). For stereo music, I don't think WMA lossless is a great route to take. FLAC is better, being that its longevity should be the longest (and even if it fades away, you still have the source), but there are other options depending on what your needs are.

FLAC strikes a good balance between encode/decode time and file size. Other formats offer better encode/decode time (and larger files) or very slow encode/decode time (and smaller files). FLAC still has the best device support of any open-source lossless format, though.
 
No, the 2.0 mixes on the DVD-ROM (or DVD-Video) layer/section aren't MLP (otherwise they wouldn't play on standard DVD Players!), so are rippable using normal DVD Rippers.
 
Some DVD-A discs have LPCM tracks on their DVD-Video compatible session, some even have a DVD-ROM session with MP3s, however most of them have only DD/DTS. A stereo MLP track is far more common.
 
Booger said:
If you had to start from scratch, what type of media would you buy? Right now I'm leaning toward standard CD's ripped to FLAC or WMA lossless, but that's been the standard for quite a few years now. I find it hard to believe that something better isn't coming along soon.
Click to expand...

My personal belief for no new formats is that the spectrum we can hear is covered in the CD format.

Id go the same route too. Use EAC to rip them (make sure you configure it right), then encode to FLAC.

One thing to remember is that while FLAC does compress, its not MP3 level compression. A typical disc will use about 400MB of space compared to the 100MB VBR MP3.

What I did for my music collection was:

- All recordings made in analog converted to FLAC (Rolling stones, AC/DC, Metallica..)
- All digital recordings went to VBR MP3 (Techno/cRap).
 
phildabill said:
My personal belief for no new formats is that the spectrum we can hear is covered in the CD format.
Click to expand...
While you're essentially right, you're also essentially a bit wrong.

Yes, the sample rate of linear PCM determines the maximum attainable frequency. According to the Nyquist thereom, this frequency is precisely half the sample rate (corresponding to a maximum frequency of 22050 Hz), though this would entail the use of a theoretical "brick wall" high cut filter. In reality, the filter has a curve determined by the components involved in the sample-and-hold quantization process.

A very virgin human ear can hear frequencies slightly above 20 kHz, though these signals become highly attenuated at a near exponential rate after 20 kHz. Most adults will have difficulty hearing above 17 kHz. But, because most tweeters don't instantly fall off the radar at 20 kHz, these frequencies can be reproduced by a typical reasonable-quality set of speakers, and these frequencies will interact with the environment in variable ways. Some have described these interactions as being able to more closely define the original room that content was recorded in. 20 kHz may not be directly audible, but it still contains "data" about the audio signal, and audiophiles deem this to be fairly important.

From a very matter-of-fact perspective, you're essentially very right. There are numerous other advantages to higher sampling rates (and indeed advantages for higher word lengths), but these advantages would end up being lost on the typical consumer who finds ~128kbps MP3 acceptable. Ergo, record companies have had a very lackluster interest in pursuing the "finer" formats (and that includes vinyl, which is typically cut from an 88.2 kHz source).

Sorry to go off on a tangent there, but I always enjoy having the opportunity to enlighten a bit.
 
phildabill said:
My personal belief for no new formats is that the spectrum we can hear is covered in the CD format.
Click to expand...

It is a fact, not a belief, that CDs cover the frequency range that a human can hear.

On the other hand, 16 bit don't cover the dynamic range of our auditory system. If you listen to pop/electronic music 24bit won't make a big difference, but with accoustical instruments and overall classical music (not to mention Tchaikovsky's 1812 Overture with cannons) does benefit from an increased dynamic range. It allows to pick subtle details in the recording without sacrificing transient loud sounds.

Moreover, if you take a look at a digital signal processing textbook, you'll see that ideal sampling (convoluting the signal with a delta train) requires a perfectly square filter for reconstructing the signal, but that would need a non-causal filter with an infinite number of taps, in the real world we can only construct causal systems with finite memory. Sampling at higher rates allows to move out of the audible region the cutoff ripple of the imperfect filter.

The above only holds when you are sampling the analog waveform, upsampling a signal that was already digital does nothing.
 
DefineByte said:
Try not to laugh at the 'enlighten' part.
Click to expand...
If you have something to challenge, then challenge it. Please don't be a trolling dick.

The depth of my knowledge intimidates some people. If you can't handle it, don't read it.

HSE said:
On the other hand, 16 bit don't cover the dynamic range of our auditory system.
Click to expand...
But does cover the dynamic range (mostly) of most albums made in the last ten years or so, unfortunately. It's still nice to have the additional depth of 24 bit words, though, even if the effective dynamic range isn't really being utilized.
 
But does cover the dynamic range (mostly) of most albums made in the last ten years or so, unfortunately. It's still nice to have the additional depth of 24 bit words, though, even if the effective dynamic range isn't really being utilized.
Click to expand...

It is true, most recordings don't need even 20 bit strictly for dynamic rante.

But how much do you think the better overall result from 24/96 recordings has to do with better 'mixing headroom' compared to utilized dynamic range? The way I see it may be similar to the need of floating point colors for 3D rendering, for a normal picture is superflous but when several layers are being blended 8bit per primary color results in a less than optimal image.

Or is it the result of better equipment and a more careful process?
 
It's true. I haven't added anything to the topic at hand. For that I apologise. I stand by my statement though. I guess I should bow out.
 
HSE said:
Or is it the result of better equipment and a more careful process?
Click to expand...
For the most part, the process tends to be the exact same as the process of creating a Redbook CD with different output settings at the end of the process. Most gold, platinum and multi- albums today are still tracked to 2" tape and dumped into Pro Tools at the highest bit depth and sample rate that platform allows and/or the sample rate that the engineer prefers (some may prefer to track tape in at 176.4 kHz for that nice-and-easy divide-by-four final downsampling to 44.1). The fork occurs at the end of the mastering stage, and mastering engineers typically receive the a stereo bounce of the session at a high bit depth/sample rate or the digital session dumped to 1/4" tape (some engineers/producers might also include vocal stems). A good mastering engineer is typically armed with equipment superior to that of mixing engineers/recording studios, so that's certainly an important component.

While mixing, you absolutely want the additional depth. For every effect you add to an audio file, a 24-bit audio file will retain more clarity over its 16-bit counterpart because the 24-bit file has inherently more information, and for direct time-based manipulation (time compression and time expansion), you want the highest sample rate allowable. The headroom doesn't really equate to anything digitally because the amplitude level is full-scale, and when you're mixing digitally, you're really mixing at some multiple of the bit depth (Pro Tools, for example, always mixes internally at 48-bit precision, meaning you can maintain perfect mixing precision by "saturating" the bussing network and pulling down one or more master faders such that they don't clip the output buss - this is actually a little-known capability of Pro Tools). For that reason, headroom is something of a non-factor. In the end, you basically have the same degree of signal you're allowed to fill as you have with 16 bits. And when outputting at 24 bits, you're forcefully removing less of the original signal, while opening up expanded options for introducing dither. Coupled with the fact that a 96 kHz signal exhibits less aliasing than a 44.1 or 48 kHz signal, and you have another minor improvement to add to list.

Personally, I don't think there's that big of a reason to offer 24/96 mixes and beyond to consumers. It's handy for manipulation, but it doesn't end up making such a night-and-day difference for those who are just listening to it. Uncompressed 16/44.1 PCM still does an incredible job of maintaining the original mix.

24/96 recordings sound marginally better pretty much strictly because less is removed from the original recording. The process of A/D conversion is always destructive no matter how you slice it, so maintaining the quality of the recording essentially requires the best equipment and the most methodical process, but the equipment and process is typically the same for both Redbook and DVD-Audio -- the higher fidelity audio is really just a fork at the end of that process.
 
Mister X said:
Rather then subject you to my thoughts on why I feel the beyond 20Khz thing is a myth I will provide you a link
Click to expand...
Interesting article. Of course, given the components most engineers use, such as Rupert Neve mic pres, Sennheiser and Neumann mics and SSL analog/hybrid consoles (with G or E series components), how much of these ultrasonic signals are preserved during tracking, mixing and mastering?

A good mixer will inevitably notch instruments such that their fundamentals and early harmonics are maintained, while strongly attenuating unnecessary frequencies so that they sit better in the mix. I know mixers that are extremely methodical about this process, because they know that positioning instruments in the available frequency spectrum is critical to a clean mix.

So, in the end, it's going to be dependent on an array of 1,438 factors that determine whether or not frequencies above 20 kHz on a particular album are really prevalent enough to make any real difference for SACD/DVD-Audio. Albums that are remixed and remastered from the original source specifically for these formats probably fare better when it comes to higher frequency content.
 
You must log in or register to reply here.
Back
Top