AVS Forum banner

Status
Not open for further replies.
1 - 20 of 25 Posts

·
Registered
Joined
·
27 Posts
Discussion Starter · #1 ·
FOr some reason, I'm having trouble getting anything other than 48khz audio out of my DVD player. I'm on my second player, and still can't get it to do 96khz on a DTS track.


I have a Denon 3803 receiver, which is more than capable of handling the 96khz audio, but the three players I've tried aren't giving it to me...


Panasonic RV-31, RV-65, and the new CP72 are the three players, and all have been connected via Optical cable and digital coax (except the 72, which is optical only).


I've tried setting things to PCM and Bitstream with the downconversion off, and can't get either to do anything but 48khz. I've messed with every setting possible on the Denon as well, and still...nothing but 48.


Anyone have any advice? (why do I sense the first comment is going to be "get a different brand of dvd player"?)
 

·
Registered
Joined
·
7,335 Posts
It's KiloHertz btw, not Hz.


If you don't have any of the handful of DTS 24/96 titles in existence, it doesn't matter anyway. All you'll get is 48K DTS.


The only titles I know of with DTS 24/96 are DVD-Audio discs from DTS entertainment. For movies it will be rare, as 24/96 requires full bit rate of 1.5Mb/sec which eats way into the available bandwidth.


Regards,
 

·
Registered
Joined
·
27 Posts
Discussion Starter · #3 ·
Yeah, I realized I forgot the "k" in "khz" in my title...oh well.


Anyways, so is my understanding of a DTS movie faulty then? I thought that full bitrate DTS DVD soundtracks were supposed to be 96khz?
 

·
Registered
Joined
·
2,841 Posts
On DVD, DTS is encoded at a 48kHz sample rate. I only know of two DTS titles on DVD that are recorded at 96kHz. One is "A Night at the Opera" by the group Queen and the other is (I believe) the Italian DVD release of the movie "Tomb Raider". If you go to the DTS website, you can get a list of the DVDs that have DTS 24/96 encoding (right side of the site).


"Full bitrate DTS" means the soundtrack has a bitrate of 1.54 Mbs -- like the first DTS releases from Universal, rather than the now common -- unfortunately -- 768kbs (where DTS and Dolby Digital are offered on the same disc).

www.dtsonline.com
 

·
Registered
Joined
·
27 Posts
Discussion Starter · #5 ·
Joe...good info


Then let me ask this...is 48khz in the best I can hope for on DD as well?


I just don't think I'm completely clear on this...should my dvd player be set to "bitstream" out for both DD and DTS?


This all isn't to say that my setup and such doesn't sound good, it most certainly does, I just think I need a little education here. I'm sitting here thinking that because of the way I have things setup, I should be getting 96khz all the time from movies on DVD, especially those with a DTS track. But this is apparently flawed thinking, as most things are only 48khz.


And through all of this, what about the 192khz? My receiver and dvd player both are rated for the 24/192 standard, but is this just a "marketing" spec, and there's basically nothing out there in that, or is it possible I have something hooked up wrong?
 

·
Registered
Joined
·
1,668 Posts
You need the bitstream out set for DD and dts for a digital connection to your receiver. 48kHz is normal sample rate for these, you might find an odd DVD that has a 96kHz dts soundtrack, sure not many. The 192kHz is occasionally used on DVD-A for stereo tracks, 96kHz is more commonly used for the DVD-A hi-res surround mixes, and often the hi-res stereo mixes.


You have hit a sore point with me: I have a DVD-A that advertises on the package "advanced resolution" when in fact it is only 48kHz, and there is absolutely nothing "advanced" about that, it is no better than standard DVD DD and dts, which they don't call "advanced" on their DD etc. mixes on the same DVD-A even though it's the same 48kHz. Still sounds good though, it's the IMO false labelling I don't like.


Note for the DVD-A hi-res surround and stereo mixes you can not normally use a digital connection (unless you have a very big buck machine) and must use analog connections to hear them.


Additional edit: Forgot to mention, so far I have not found a DVD(-A) that can use the 192/24 capability of a common receiver DAC via a digital input, so far it's been 96/24 best.
 

·
Registered
Joined
·
1,881 Posts
Quote:
Originally posted by lord_zeppelin
Then let me ask this...is 48khz in the best I can hope for on DD as well?
Dolby Digital takes 48-kHz/20-bit signals at the encoder input and spits out 48-kHz/20-bit at the decoder output. Same for standard DTS. DTS does have a 96-kHz/24-bit system, but given that it, like DD, is a system based on reducing data rate by omitting inaudible information, the whole idea is a bit silly. There really is absolutely no reason to be concerned about getting 96-kHz sampling rates; all that gives you over 48-kHz sampling are audio frequencies between 24 and 48 kHz, where you can't hear and there really isn't anything of interest to hear anyway.
 

·
Registered
Joined
·
1,881 Posts
Quote:
Originally posted by cfraser
YYou have hit a sore point with me: I have a DVD-A that advertises on the package "advanced resolution" when in fact it is only 48kHz, and there is absolutely nothing "advanced" about that, it is no better than standard DVD DD and dts, which they don't call "advanced" on their DD etc. mixes on the same DVD-A even though it's the same 48kHz. Still sounds good though, it's the IMO false labelling I don't like.
I don't think there's any reason to get worked up about this. The advantage of DVD-Audio tracks over DD or DTS tracks is that no lossy compression is used. Even that is not a huge benefit, but you can at least be sure that the signal reaching your speakers is the same as if the master tape were feeding the amplifiers and that there is no possibility of degradation owing to compression artifacts. You can never be absolutely certain of that with a lossy system. A 48-kHz sampling rate gives you a maximum recorded frequency of 24 kHz. Assuming you are an adult (or even an average 10-year-old), you can't hear 24 kHz, so bumping the sampling rate up to 96 kHz is not going to buy you anything.
 

·
Registered
Joined
·
1,668 Posts
I know MD, but a higher sampling rate gives me more smoothness, less time difference between samples that has to be "averaged" by a single sample. So what if I can't hear it, I know it's there. It's like driving a 911 turbo at 30mph. I want my bits, I want my samples, and I want to eat them too.:)
 

·
Registered
Joined
·
27 Posts
Discussion Starter · #10 ·
Well, now that I feel better about this...thanks to all!


Really, I was just misinformed from the start on this one.
 

·
Registered
Joined
·
2,841 Posts
LZ, I believe that cfraser has answered most of your questions. As far as the 192kHz material goes, there is very little material available and, as was mentioned, you don't have access to this sample rate on the digital end -- unless you can afford megabuck gear.


As far as higher sampling rates are concerned, here are my thoughts. Just remember that these are my opinions and some (*cough*cough*hearing impaired*) may disagree.


When 96kHz recordings were made available on DVD-V by Chesky and Classic Records, the reviews cited how much better they sounded than their CD cousins. Was this due to the new medium's ability to capture higher frequencies in the recording? In part, yes. But the naysayers began their rants about human hearing -- it is limited to 20 - 20kHz at best and the older you get, the worse high frequency hearing becomes. Well, that's not entirely true. Musical notes aren't solid frequency plots. If you've ever seen a frequency plot of a plucked or struck instrument, you will see that the note played is accompanied by sound to the left and right of the primary note. Jumping ahead, sound that has a primary frequency of 30kHz (remember, this is beyond human hearing according to the naysayers) will also have output -- though lower in level -- down past 15kHz. I believe most people can hear 15kHz, and if it's high enough in level when played back, its inclusion or exclusion will affect the perception of the music.


If, however, you filter what is above 20kHz, then you have basically eliminated part of your perception of the music that could have been recorded. While this is a very simplified explanation, it should make it clear that those who say "humans can't hear above 20kHz so you don't need to record higher than that" are wrong. Period. There is no debate -- their case is based on a faulty premise.


Most people who have compared 44.1kHz recordings to 96kHz recordings of the same material discover another added benefit of a higher sampling rate -- tighter and more defined bass. Yes. Bass. Low frequency response. Kick -- call it what you want, it is perceived as somehow better at the higher sampling rate.


Another advantage to a higher sampling rate is that the digital filter is moved farther up in frequency. It has been shown on several occassions that filters can have a detrimental affect on our perception. Moving them farther up in frequency lessens their grip on our perception of the music.


Unfortunately, when it comes to lossy codecs (Dolby Digital, DTS, etc), increasing the sampling rate doesn't affect our perception as much as a lossless system does. Since there is so much data deleted in the first place, you can't expect a substantial quality improvement by increasing the sample rate.
 

·
Registered
Joined
·
1,668 Posts
Very good points Joe. And may I mention one more that may be controversial here? Using an external "better quality" DAC for stereo music listening. On occasion I have connected my DVD player to an external DAC. For listening to stereo DVD-A mixes with my stereo system, what I use for serious music listening. I have found that the *one* 48kHz stereo mix I have sounds better in both these instances: at 48kHz sampling with the external DAC, at 96kHz sampling (please note this is *not* upsampling) with the same external DAC. Likewise the 96kHz stereo mixes through the external DAC. Just some thoughts, your ears are connected to the brain very closely, and perception is all. My bottom line thought is that receiver DAC's tend to sound very dry and non-spatial.


Edit: This is not intended for you Joe, I know your thoughts from your post and elsewhere. For others who may have a DAC and haven't tried it.
 

·
Registered
Joined
·
2,841 Posts
You know what's crazy? There were some listeners that felt the bass reproduction at 96kHz was more solid/tight than 192kHz. I don't really know what was going on here. The only thing I could think of was that the 96 and 192 rates were either sent through different DACs (dual feed setup?) or the comparisons sent the 96kHz signal through a multibit chip and the 192kHz signal fed a delta-sigma chip. Who knows. Wish there was more detailed information listed when these early comparisons were made...
 

·
Registered
Joined
·
1,881 Posts
Quote:
Originally posted by cfraser
I know MD, but a higher sampling rate gives me more smoothness, less time difference between samples that has to be "averaged" by a single sample. So what if I can't hear it, I know it's there. It's like driving a 911 turbo at 30mph. I want my bits, I want my samples, and I want to eat them too.:)
It doesn't matter how much time difference there is between samples so long as there are at least two per cycle. As long as the sampling rate is at least twice the highest frequency to be recorded, you can reproduce the waveform perfectly. PCM is not a connect-the-dots system; adding more than the minimum number of samples per cycle does not improve the accuracy of the reproduction.
 

·
Registered
Joined
·
1,881 Posts
Quote:
When 96kHz recordings were made available on DVD-V by Chesky and Classic Records, the reviews cited how much better they sounded than their CD cousins. Was this due to the new medium's ability to capture higher frequencies in the recording? In part, yes. But the naysayers began their rants about human hearing -- it is limited to 20 - 20kHz at best and the older you get, the worse high frequency hearing becomes. Well, that's not entirely true. Musical notes aren't solid frequency plots. If you've ever seen a frequency plot of a plucked or struck instrument, you will see that the note played is accompanied by sound to the left and right of the primary note. Jumping ahead, sound that has a primary frequency of 30kHz (remember, this is beyond human hearing according to the naysayers) will also have output -- though lower in level -- down past 15kHz. I believe most people can hear 15kHz, and if it's high enough in level when played back, its inclusion or exclusion will affect the perception of the music.
First, there are no musical fundamental frequencies above a few kilohertz. (One of the things you discover listening to test tones is that 1 kHz is actually treble and that 15 kHz sounds like a sort of pitchless, disembodied squeal!) A few instruments (cymbals, most notably) can produce harmonics beyond 20 kHz, but people don't seem to be able to tell the difference when those are removed. Second, acoustical instruments produce harmonics of fundamental tones but not "subharmonics." In other words, you get output at multiples of the fundamental frequency but not at fractions of it. The only way to get output below the fundamental is by some nonlinear process (i.e., distortion). So if you actually had an instrument with a fundamental at 30 kHz, you would get output at 30 kHz and 60 kHz but not at 15 kHz. This is just the physics of harmonic oscillators.


People are going to believe what they want to believe about the audible significance of extremely high sampling rates and reproduction of ultrasonic frequencies. However, this is a subject that has been studied fairly extensively, and the hard evidence is that what goes on way up there is of no audible consequence to humans. (And I say this as someone who had exceptional high-frequency hearing acuity in his youth and not bad even now.) Not too surprising a result, since, as a psychoacoustician I know once said, we start running out of physiology at about 18 kHz; the ear's frequency response (sensitivity versus frequency) pretty much falls off a cliff starting at about 14 kHz. We just don't hear well way up yonder.
 

·
Registered
Joined
·
1,668 Posts
Sorry, I disagree MDRiggs. 2 samples per cycle is the abolute minimum number of samples required to reproduce without aliasing. More samples (of anything, under any circumstance) will always give more accurate results. We are dealing in the statistical domain with digital stuff... Yes, more samples can increase the accuracy if it's done properly. For DSP filters, the number of samples that can be handled is limited by the processing power. It is effectively connect-the-dots because the digital filters are interpolating between the samples.
 

·
Registered
Joined
·
2,841 Posts
As MDRIGGS pointed out, you don't get frequencies lower than the primary. I guess I need to proof my comments or not provide input so late at night. What I should have said was that all of the harmonics would not be captured correctly (15kHz to 30kHz, not the 30kHz to 15kHz that I typed). A 20kHz cutoff will basically "brick wall" the natural progression of the notes and their harmonics. The smooth decrease in the power level of the multiples will come to a rather abrupt halt, rather than continuing on to zero level (ie, a cliff instead of a ramp).
 

·
Registered
Joined
·
1,881 Posts
Quote:
Originally posted by cfraser
Sorry, I disagree MDRiggs. 2 samples per cycle is the abolute minimum number of samples required to reproduce without aliasing. More samples (of anything, under any circumstance) will always give more accurate results. We are dealing in the statistical domain with digital stuff... Yes, more samples can increase the accuracy if it's done properly. For DSP filters, the number of samples that can be handled is limited by the processing power. It is effectively connect-the-dots because the digital filters are interpolating between the samples.
If this were true, you would automatically get a distortion curve that sloped upward with frequency, which doesn't happen. For a really, really obvious refutation, hook a CD player up to an oscilloscope and play a 20-kHz tone off a test CD. It will be perfect.


Forget about digital filters for a moment and go back to the way D/A conversion had to be done in the old days. There was just a steep analog low-pass filter at the output of the D/A converter. This worked fine, and we would still be doing it that way today except that it is cheaper to get good results with digital filters than it is with analog filters, which are hard to design and build with both the desired slope and negligible response ripple. Anyhow, what's the difference between the output of the D/A converter before and after filtering? Before filtering it contains a lot of ultrasonic energy whereas after it doesn't. No interpolation. Yet you will get an exact replica of the original waveform before A/D and D/A conversion (save for some linear phase shift that can be undone if desired). And I do mean exact!


I agree this is counterintuitive, but let's go back and look at what happens on the A/D side. Sampling is an analog process that facilitates quantization. The sample-and-hold amplifier in an A/D converter essentially freezes the signal voltage momentarily so that the quantizer can measure the level. If you look at the output of a sample-and-hold amplifier fed a sine wave, what you see is a stepped version of the sine wave. But sharp transitions, such as you see in the stairstepped waveform, are an indication of very-high-frequency content, such as you would have in a square wave. And in fact, if you low-pass filter the output of the S&H you will get back the input signal exactly, so long as that signal did not contain any components at frequencies greater than half the sampling frequency. Because the output from the S&H is the original signal with a lot of high-frequency energy added to it. Remove that out-of-band energy and you're back where you started. That's the whole point of the sampling theorem--that as long as the frequency criterion is met, the process is lossless.


Now, you quantize all those voltages and put them on some medium, and then eventually you run them through a D/A converter. What the D/A converter does is to reconstruct that stairstepped waveform that came out of the S&H circuit into the A/D. Filter it, and you've got back the original signal. It really does work, and you really don't gain anything by adding extra samples.
 

·
Registered
Joined
·
1,881 Posts
Quote:
Originally posted by Joe Murphy Jr
As MDRIGGS pointed out, you don't get frequencies lower than the primary. I guess I need to proof my comments or not provide input so late at night. What I should have said was that all of the harmonics would not be captured correctly (15kHz to 30kHz, not the 30kHz to 15kHz that I typed). A 20kHz cutoff will basically "brick wall" the natural progression of the notes and their harmonics. The smooth decrease in the power level of the multiples will come to a rather abrupt halt, rather than continuing on to zero level (ie, a cliff instead of a ramp).
But it doesn't matter if you lose harmonics at frequencies you can't hear, and it doesn't matter how rapidly they are attenuated.
 
1 - 20 of 25 Posts
Status
Not open for further replies.
Top