|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
Someone in an earlier thread alluded to a flaw in MLP compression. Since it is lossless, I do not understand how it can be flawed compared to PCM, since it decodes to the exact same data as the original PCM stream. Perhaps it does not always compress well, but it should not affect the stream sonically, unless it's not truly lossless.
Follow Ups:
.
.
With DVD-A, if measuring the outcomes against their objectives, its easy to see they failed. Also, I'll include a dictionary definition of flawed, so we can measure whether my use of the word is fairFlawed : An imperfection, often concealed, that impairs soundness
1) MLP : Couldn't achieve the desired bit-rate, so a low-pass filter was needed to roll off high frequencies.
An imperfection : YES
Concealed : They tried to keep it from the public
Impairs Soundness : YESMLP = Flawed
QED.
2) Verance Watermarking : Was promised to be inaudible, and then proven not to be. Made not mandatory since it failed to meet its promises.
An imperfection : YES, in more ways than one
Concealed : Warner still won't label their recordings.
Impairs Soundness : YES
Does it work as intended : NO, it was intended to be inaubileVerance Watermarking = Flawed & Broken
QED.
3)CD Hybrid Compatibility :
Stuffed : It doesn't work
Poor planning : It wasn't planned for, so when they tried to implement it, it didn't work.CD Hybrid Compatibility = Stuffed & Poorly planned for
QED.
4) Encryption : I will accept this is currently unproven and is conjecture.
5) Stereo System Compatibility : Not Mandated, but included ATM
A simple statement of fact.
6) Copy Protection Integrity of 5.1 :
We know its hacked
We know the general public don't mind lossy compressionHow can it be argued this isn't a serious problem ? Do you think people will care to buy an original DVD-A if they can get a copy for a $1 with everything on it but MLP Hirez ? No they won't care.
7) Sales : Low
If selling 700 odd copies of Elvis #1 greatest hits isn't LOW, what is ?
8) Catalogue : Small
Small is being generous!! The number of Actually released titles is tiny.
9) Labels : Few
Compare the number of labels on SACD with those on DVD-A... and many of those on the DVD-A list have released "1" title only.
10) Exclusive Equipment Manufcatures : Leaving the format in droves
This is a fact, most of the previous DVD-A only manufactures have or are moving to Universal players.
Royalty Costs : Too High
Ok, so this is subjective, However, posted on this board is that DVD-A royalties are 6 times those of DVD-A. Surely that makes it high compared to SACD.
I suggest doubters read this excerpt from "The MLP lossless compression system" by Gerzon, Craven, Stuart — published at the 17th AES (sorry -- I cannot include illustrations):-/ . . . . . . . . . . . . . . . . . . . . . .
4.7 Buffering
We have explained that while normal audio signals can
be well predicted, there will be occasional fragments
like sibilants, synthesised noise or percussive events that
have high entropy.MLP uses a particular form of stream buffering that can
reduce the variations in transmitted data rate, absorbing
transients that are hard to compress.FIFO memory buffers are used in the encoder and
decoder as shown in Figure 13. These buffers are
configured to give a constant notional delay across
encode and decode. This overall delay is small –
typically of the order of 75ms. To allow rapid start-up or
cueing, the FIFO management minimises the part of the
delay due to the decoder buffer. So, this buffer is
normally empty and fills only ahead of sections with
high instantaneous data rate.During these sections, the decoder’s buffer empties and
is thus able to deliver data to the decoder core at a
higher rate than the transmission channel is able to
provide. In the context of a disc, this strategy has the
effect of moving excess data away from the stress peaks,
to a preceding quieter passage.The encoder can use the buffering for a number of
purposes, e.g.:~ Keeping the data-rate below a preset (format) limit.
~ Minimising the peak data rate over an encoded
section.Figure 14 shows an example of the latter. The entropy-coded
data rate from the encoder core is shown along
with the buffered result. The buffered data has a
characteristic flat-topped curve. This is not due to
clipping or overload, but to rate absorption in the
encoder/decoder FIFOs.Another illustration of data-rate minimisation is shown
in Figures 15 and 16. Again the encoded data rate is
plotted through a 30-second 96kHz 24-bit 6-channel
excerpt featuring a close recording of a jazz saxophone.
Figure 15 indicates the underlying compression with the
encoder set to limit above 9·5Mbps. The minimum-rate
encode shown in Figure 16 makes long-term (but low
occupancy) use the decoder buffer.It should be obvious that the situation in Figure 16 is
preferable if the transmission channel (maybe DVD
disc) has other calls on the bandwidth – for example
bandwidth to transmit associated picture or text.
Figure 17 shows how hard-to-compress signals can be
squeezed below a preset format limit. This 30-second
96kHz 24-bit recording features closely recorded
cymbals in 6 channels. At the crescendo this signal is
virtually random and the underlying compressed data
rate is 12·03Mbps. Buffering allows the MLP encoder to
hold the transmitted data rate below 9·2Mbps by filling
the decoder buffer to a short-term maximum of 86kbyte
(bottom curve).Figure 18 shows the potential for peak data-rate
reduction on this item with different amounts of
available FIFO memory.
It is flawed, since MLP did not meet their own objectives, and had to be "patched up" to make it work.This is undeniable, the amount of importance you place on this hack however is subjective. What we can say with 100% certainty is that your not getting the "master tape" sound, with watermarking and HF roll off applied.
No matter how good your DA's get, you can never get back what has been ruined.
> What we can say with 100% certainty is that your not getting the "master tape" sound, with watermarking and HF roll off applied. <Watermarking can and is applied to SACD, but so is a mandatory 50kHz filter which removes all content above that point because of the huge amount of noise present in the DSD stream, unlike the flat response of MLP out to 96kHz.
If you think you're getting "master tape" sound with SACD, then you've got another thing coming buster! ;-)
or at least the use of some sort of buffer, in any sort of streaming where the rate varies.The fact that MLP uses a FIFO is not surprising. What Dave P. seems to be implying that they have not used a large enough FIFO, so that they still have problems.
If the offending signal is sustained the problem will occur slightly later.If a signal is random and has high level high frequency content the information cannot be packed below the maximum allowable range.
With normal music it will practically never occur because music is by nature rather 'predictable'. And if it happens the measures are simple and have no significant or detectable impact on sonics.
I similar problem exist with dsd where they strongly suggest to keep the signal input -6dB below the maximum range of the ad converter.
With good recording (for dsd) and mastering practises these 'overload' problems are non issues.
As tunenut mentioned it's not a fixed bitrate encoding system.When signals in all channels are not correlated to each other AND have high amplitude high frequency content. Then the resulting bitrate will exceed the maximum set for the DVD Audio spec.
With normal music this will virtually never happen. High level noise and rock music with high compression are likely candidates.
The problem can be 'fixed' by using a slight rollof above 24kHz.
This is the easy way out.It's also possible to lower the bitrate for the rear channels by lowering the bit depth or even the sample rate.
It's also a problem that only will occur for short instances of time if at all.
So measures to prevent encoder 'overload' can be temporary to prevent ill effects.A good mastering engineer should know is business anyway.
It's also important to realize that the original pcm recording and the master are still linear pcm and not affected at all.
Frank
Since my old post went missing here's a copy..It turns out that MLP doesn't handle high frequency content very well, so as a patch it is sometimes necessary to apply a low pass filter to keep the data rate within DVD spec.
This trick can also be used to squeeze more content onto a disc. For example, the DVD-A of "Swing Live" has the surround channels rolled off to make room for two surround mixes.
Bottom line, you can't rely on the displayed sample rates as an indicator of the actual bandwidth. Even at 24/192 you may not be getting the full resolution of the master recording.
James Guthrie, the mastering engineer for DSOTM, cited this problem as one of the reasons that SACD was chosen for the release rather than DVD-A.
Cheers,
Dave.
but this is what people were talking about.A compression algorithm like mp3 is designed to put out data at a constant rate. A lossless compression will not be like this. Depending on the predictability of the incoming data, it will put out more bits or less bits per unit time. It has to put out enough bits to completely represent the incoming data, or it will not be lossless.
So there are times, according to what was written here earlier, when MLP puts out too much data for the available bandwidth. To stop this from happening, you can remove some information from the original signal, so you put in some filtering perhaps.
Not a big deal, probably not of audible consequence. But it would be a factor limiting the bandwidth below theoretical maximum of the samplling rate.
It isn't all Nyquist. You can have a smoother rolloff if you limit before encode.But, don't listen to me. I've been known to sleep with the enemy and shoot them in the morning.
After the bistream has been unpacked & decoded there is a constant bitrate to the DAC. Buffer control is done on the DSP side which takes care of overruns and underruns. A properly designed system will operate in real-time and avoid all these problems.
What we can say with 100% certainty is that you will never get the "master tape" sound, with watermarking and HF roll off applied.No matter how good your DA's get, you can never get back what has been ruined.
(a) Master tapes:
SnaggS, it may seem incredible to you, but I try to AVOID titles from analog master tapes (unless it’s for legacy music). But in the end, it’s LPCM digital masters at 24bit @192kHz (or 32bit @ 384kHz) which are the ideal.(b) Your alleged MLP "roll-off":
As countered elsewhere by people with FIRST HAND experience, the "HF roll-off" optional technique you mention with MLP, is virtually NEVER NEEDED with real music on DVD-A. (SnaggS, can’t you get it???) Sure, if you blast white noise continuously into all six channels at the highest frequencies and intensities, then you’ll be pushing the limits of the DVD playback system, although MLP will still make a pretty decent job of it (but anyway, is that real music?).Moreover, this situation is indeed mentioned in the technical paper (and no SnaggS, they’re not trying to "hide" it from the public domain), because as a technical paper, it rightly mentions all signal compression optimisation tools available to the user. That built-in flexibility is a "good thing". In fact, what it is doing is getting the very most out of the DVD playback media.
Furthermore, the actual studio masters have no such practical limitations, so any such concern becomes completely irrelevant for recording engineers. Indeed, with LPCM / MLP, the studios have the means to create a high-definition lossless master which is totally portable to ANY consumer playback media (past, present, or future) — whatever the limitations or otherwise which that particular media possesses.
(c) Optional watermarking:
As for watermarking, this is certainly open to debate (for SACD and CD as well!), but I have numerous classical DVD-As (among other styles) which do not have any watermarking. Anyway, classical music is not really the target of mass-pirates.(d) The "best" DVD-A can be (and frequently IS) =
1. For classical/acoustic:
good artist => good location => good microphones => 24bit/192kHz A-to-D converters => 96kHz (5ch + LFE) or 192kHz (stereo) / 24 bit LPCM master => MLP encoder => robust multibit encryption => dual-layered DVD-9 media => long play lengths well in excess of 70min (St Teldec’s St. Matthaus Passion approaches 3hours at 96kHz/24bit in 5.1 surround).2. For rock/pop:
popular artist => good studio => good microphones & direct feeds from instruments => 24bit/192kHz A-to-D converters => up to and exceeding 48 tracks realtime mixing & floating point DSP => 96kHz (5ch + LFE) or 192kHz (stereo) / 24 bit LPCM master => MLP encoder => robust multibit encryption => dual- layered DVD-9 media => long play lengths well in excess of 70min.Also to be mentioned should be these added-value items:- videos; artist bios; interviews; realtime song lyrics; other album trailers; fan-club / concert info; web links etc. etc. not to mention Dolby & DTS mixes, and now the possibility of including a redbook CD on the other side.
Meanwhile, the proportion of DVD players with DVD-A (MLP) decoders is increasing all the time. Soon they’ll practically ALL have MLP decoders in them.
Your patience is admirableI hope you're aware that SnaggS has not yet listened to a single DVD-A on a DVD-A player... I think he just doesn't like the way the name "DVD-A" sounds, and decided to go on a crusade of his own.
Best
.
Go listen to stuff on good systems and decide for yourself. There is too much BS here anyway (need those high boots to wade through it). Bottom line is if you hear a flaw or not. Everyone has opinions. One can go back and forth between the 2 boards and figure out easily who is more or less credible if you need some feedback. I have all formats since I can't control who releases music on a particular one. I know people perfectly happy with DD and if that is what makes them happy that is OK.
I have no way to compare like material, I don't know of any material that is available in PCM 96/24 or 192/24 and also available in MLP 96/24 or 192/24. If the compression is lossless however, and since the Hi resolution audio is only output through analog interconnects on most systems, I would think MLP is flawless, unless those mastering the material use some sort of unnecessary filter in making the disc.
"since the Hi resolution audio is only output through analog interconnects on most systems, I would think MLP is flawless, unless those mastering the material use some sort of unnecessary filter in making the disc."Doesn't matter whether you are outputting analog or not. We are talking about the internal digital bus rates, before the DAC.
Therefore, if a mastering engineer uses a filter in making the disk, this will be done to keep the MLP output within acceptable bounds, thus it is "necessary", not "unnecessary".
MLP by design will be flawless in its compression and decompression, but if the rate is too big to fit on the available pipe. this flawless operation cannot work.
I don't see why this is an MLP flaw. It's a limitation of the DVD spec.I also doubt it could be 'repaired' while remaining lossless.
is that MLP is designed to work within the existing DVD spec, not the other way around. So it if overflows its allotted capacity, then it is at fault. Although of course you are right, you could also attack this on the capacity side (DVD spec). It could quite likely be "repaired" by tweaking the predictive algorithm, although this would necessitate a firmware change to existing implementations.
The purpose of MLP is to take six channels of 24/96 LPCM (about 14Mbps) and pack it down so it fits within the DVD transfer rate of 10.5Mbps. If the algorithm can't do that without filtering the audio first then it's not really lossless, is it?
One little piece of information that might clear up things a bit: the MLP encoder will never do any changes of bandwidth, frequency response or whatever on its own. In fact, during the encoding process the resulting MLP data stream is decoded and verified against the original PCM input, just to make sure that the player/decoder output is an exact, bit-by-bit equivalent to the input into the encoder. If this verification step fails, the encoding process is aborted, and you get an error message.The Minnetonka version of the encoder has a feature (called ReBit) that allows you to reduce the word size from 24 down to 22, 20, 18 or 16) bits on a per-channel basis in case the encoder should fail with your source material. Until now, I've never needed to use that.
Mathias
You said:> > The purpose of MLP is to take six channels of 24/96 LPCM (about 14Mbps) and pack it down so it fits within the DVD transfer rate of 10.5Mbps < <
Meridian says otherwise. Have you read their MLP paper? See page 1.
Here is what they claim: "MLP performs lossless compression of up to 63 audio channels including 24-bit material sampled at rates as high as 192kHz." So, why would Meridian design a 63-channel algorithm for DVD-A? The alg. can be used for DVD-A, but it's purpose or use is of a much wider scope.Also, the DVD-A max. rate according to Meridian is 9.6Mbps. See page 9 which states, "In the case of DVD-Audio peak rate is a key parameter because the encoded stream must always operate below the audio buffer datarate limit of 9.6Mbps"
-Joe Friday ("just the facts")
The maximum DVD transfer rate is 10.5Mbps, but due to error correction codes the maximum useable bitrate is 9.8Mbps.And yes, I'm aware that MLP has greater capabilites than six channel 24/96. My point still stands that the primary purpose of MLP on DVD-A is to compress 6-ch 24/96 so it fits within the current DVD spec and if it can't do this without overflowing its FIFO buffer then it isn't doing lossless compression of the master.
Personally I don't see the filtering as a big issue. Whether it's a 24kHz filter to get MLP to work (as Frank suggested) or a 40kHz filter to eliminate the ultrasonic noise on DSD, they're both above the audible band so should have minimal effect on the sound.
Cheers,
Dave.
This post is made possible by the generous support of people like you and our sponsors: