1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Possible to record 24/96 with a PC DVD burner?

Discussion in 'High resolution audio' started by The_Mule, Oct 17, 2003.

  1. tigre

    tigre Moderator Staff Member

    Joined:
    Feb 18, 2003
    Messages:
    790
    Likes Received:
    0
    Trophy Points:
    26
    Why? How exactly did you create them? What did you use for playback?

    Have you tried downsampling these files to CD resolution using high quality algorithms like SSRC or foobar2000's diskwriter (noiseshaped dither enabled)?

    BTW: for soundblaster soundcards you can improve playback quality noticably in many cases, if you upsample 44.1kHz output to 48Khz, using decent software foobar2000, winamp + SSRC output plugin (because otherwise soundblasters resample themselves before D/A conversion using crappy algorithms)

    AFAIK there's no free software for this task ATM; see here: http://forum.doom9.org/showthread.php?threadid=45815& (<- anybody interested: spam this thread with "I want this too!" messages ;) )
    Some expensive DVD-V authoring software can do this. It should be quite easy to code, though because doing it the other way round (ripping 24/96 LPCM stream to .wav) is already possible easily.

    I don't remember which, but some of audigy soundcards (at least audigy 2 models) support full resolution DVD-A (analog output only, of course). You'll find more information about that at Creative and/or WinDVD homepage.__X_X_X_X_X_[small]AFTERDAWN FORUM RULES: http://forums.afterdawn.com/thread_view.cfm/2487[/small]
     
    Last edited: Dec 16, 2003
  2. wilkes

    wilkes Regular member

    Joined:
    May 31, 2003
    Messages:
    954
    Likes Received:
    0
    Trophy Points:
    26
    Beware of using 24/96 files in the DVD-Video section, as most DVD-Video players truncate to 48KHz, check your manual to be sure.
    Most entry level authoring apps do not support 24 bit.
    You will also have to include some form of visual content in the DVD-Video zone too, although just one still image will do the job.
    Freeware - Forget it. Be prepared to spend at least $500, and if you are going to spend that much, you may as well get discwelder steel and write proper DVD-Audio discs instead, as long as your player supports this. Current version of WinDVD does, as does Soundblaster's Audigy 2 and a terratec card also does DVD-Audio nowadays, although I do not remember which one off the top of my head.
    From what you say, your soundcard DOES support DVD-Audio. Seriously, forget doing it through DVD-Video, as it's a pain in the backside compared with discwelder.
    Hope this helps, and for what it's worth I honestly think you will easily spot the better quality of DVDA.
    Enjoy..
     
  3. maxg

    maxg Member

    Joined:
    Dec 16, 2003
    Messages:
    10
    Likes Received:
    0
    Trophy Points:
    11
    Well thanks for the input - looks like I am boldly going...so I will let you all know if I achieve any success.

    As regards my previous forays into CD recording there was nothing wrong with the CD's produced - in comparison with commercially available product, but it was just too far below the standard of the vinyl.

    FWIW - the recorders tried included a dedicated unit from Sony (i.e. not on a computer at all), a Plextor drive with Nero (old version), unknown drive on a laptop with Nero software again I think and a variety of others. All much of a muchness.

    I can already produce 48KHz / 16 bit DVD video's so the idea of upsampling / downsampling onto CD has little appeal. I cant say the sound-quality varies greatly from CD - it maybe a bit better - but I wouldnt want to have to pick one out in a double blind test.

    Oh yes - playback systems include: Marantz CD6000 CD player, Denon 3000 DVD player and a locally supplied no-name player that does just about everything imaginable except DVDa and SACD.
     
  4. wilkes

    wilkes Regular member

    Joined:
    May 31, 2003
    Messages:
    954
    Likes Received:
    0
    Trophy Points:
    26
    Where you will really notice the improvement is in the 24 bit word, rather than a few Khz on the sampling rate. - At least, that's what I think really makes an audible difference. To my mind, the big difference that bigger samplerates maje is in doubling the bandwidth you are using for any processing, with half of that bandwidth being above 20KHz. This shoves most of the errors way out of audible range, as well as helping with aliasing errors too.
     
  5. maxg

    maxg Member

    Joined:
    Dec 16, 2003
    Messages:
    10
    Likes Received:
    0
    Trophy Points:
    11
    Wilkes that is funny because it is the complete opposite of the conclusion I am leaning towards.

    In other words I think the bigger difference will be due to the increase in the sampling rate to 96 Khz from 44.1. The higher the sampling rate - the closer you are to the original sound waves.

    Of course this is all theory. right now I have no means of copying either 96 KHz or 24 bit to a DVD. I am stuck with 48 Khz / 16 bit!
     
  6. A_Klingon

    A_Klingon Moderator Staff Member

    Joined:
    Dec 29, 2001
    Messages:
    529
    Likes Received:
    0
    Trophy Points:
    26
    It's nice to have both but I tend to agree with maxg.

    If the sampling rate is high enough, I imagine the wordlength (accuracy) of those samples wouldn't be so critical.

    If there are two *few* samples to begin with, does it really matter how well those (too few) samples are rendered?

    (We need both), and I admit this is just a 100% subjective opinion.
     
  7. wilkes

    wilkes Regular member

    Joined:
    May 31, 2003
    Messages:
    954
    Likes Received:
    0
    Trophy Points:
    26
    But it's the wider word that increases your dynamic range from 96dB @ 16 bit to 144dB @ 24 bit. Greater accuracy gives a better result than merely more measurements, plus you then get to shift any noise at all remaining well below the audible threshold.
    A 16/96 signal is not even close to as good as a 24/48 signal. PCM encoding works a little differently to how most people percieve it to. It is an oft quoted myth that higher sample rates more closely approximate the analogue waveform. A higher sample rate allows a wider bandwidth, that is you can encode a signal with a greater frequency range thanks to our old friend Nyquist. The advantage of doing this is you avoid a lot of the aliasing problems and other artifacts introduced by the (necessarily) steep filters used. A wider word, on the other hand, vastly improves the quality of the measurement being taken. The higher the rate, the less noise is audible.
    Another thing to consider, going back to the measurements, is that a 16 bit system has a total of 65,536 values (Bit depth is the amplitude part of the process, Sampling rate is the frequency component), and a 24 bit system has a total of 16,777,216 values for each sample. To my mind, this clearly shows the value of the wider word.
     
  8. maxg

    maxg Member

    Joined:
    Dec 16, 2003
    Messages:
    10
    Likes Received:
    0
    Trophy Points:
    11
    "It is an oft quoted myth that higher sample rates more closely approximate the analogue waveform"

    Yup - shared with Creative for one - their documentation of the benefits of 96 Khz is full of it. I am not as sure as you are that it is a million miles from the truth..
     
  9. tigre

    tigre Moderator Staff Member

    Joined:
    Feb 18, 2003
    Messages:
    790
    Likes Received:
    0
    Trophy Points:
    26
    Using noise shaped dither you can convert 24/96 to 16/96 with noise floor in the audible range lower than -120dB. Every DAC will add a louder noise floor, so talking about signal-to-noise ratio or dynamic range, 16/96 and 24/48 are equal in practice. If noise shaped dither is used (with high energy noise in > 20kHz range) filters and DACs must be designed properly to avoid aliasing issues - similar to 48kHz sampling rate for other reasons.
     
  10. The_Mule

    The_Mule Member

    Joined:
    Mar 20, 2002
    Messages:
    32
    Likes Received:
    0
    Trophy Points:
    16
    I also agree that sampling freq has a lot more benefit than bit length. Why?

    1) If you look at the measurements that Sound & Vision does for even the best DVD-A and SACD players available today, current *hardware* technology can only give 18 - 20 bit resolution at best.

    2) Rock/pop/rap music? With all the compression, limiting, and processing involved in mastering these days, I bet most CDs only take advantage of *maybe* 8 - 12 bits.

    3) Ever hear someone say that CDs sound harsh, edgy, bright, or glareful compared to lp? That is sampling freq my friend. The higher the freq, the smaller the time chunks are, and hence the better approximation to the original analog waveform. Affects high freqs the most. Not rumor, fact.
     
    Last edited: Dec 17, 2003
  11. tigre

    tigre Moderator Staff Member

    Joined:
    Feb 18, 2003
    Messages:
    790
    Likes Received:
    0
    Trophy Points:
    26
    Nope. That is high frequency attenuation and distortion added by *LP* (during recording + playback). It's similar to tube amps. They add distortion to the signal perceived as "warmth" etc., so one might have the impression it sounds better, while CD / good transistor amps reproduce the sound closer to original.
    If you want to reproduce inaudible 22-48kHz sounds to make listening more enjoyable for your cat and dog, that's true. Approximation of sounds < 22kHz can't be improved using higher sampling rates.
     
  12. wilkes

    wilkes Regular member

    Joined:
    May 31, 2003
    Messages:
    954
    Likes Received:
    0
    Trophy Points:
    26
    You obviously do not understand how PCM encoding works. Tigre has it right when he says that
    Your understanding of PCM is , simply, wrong.
    The oft-quoted myth that "the closer you follow the original waveform, the more accurate the reproduction" is just not true.
    If you increase the sampling rate, the benefit is that you double the available bandwidth, and thereby shift a large amount of the quantization errors into the ultrasonic range, thus reducing audible aliasing. The accuracy of measurement comes from the wider wordlength. It's a basic tenet of digital recording, my friend. I strongly suggest you go take a quick refresher course in how it all works. Quick quiz - what is better? 8 bit at 44.1 KHz or 24 bit at 44.1 KHz?
    What will make more difference? 8 bit at 96KHz or 24 bit at 44.1 KHz?
     
  13. maxg

    maxg Member

    Joined:
    Dec 16, 2003
    Messages:
    10
    Likes Received:
    0
    Trophy Points:
    11
    I think we are drifting off topic a bit - I suppose it happens in threads like this. For me the only really important questions are:

    Which is better 96/24 or 44.1/16?
    Is the difference audible?
    Can it be produced on a DVD+RW or equivalent for replay in a normal DVD player?

    As it happens I have now checked both of my DVD player manuals and, purely out of luck, have discovered that both support 96/24. This was something of a surprise for my old Denon 3000 (a generation 1 player that is still going) but it does mean I am kean to persue this issue into the ground.

    I have just learned that the latest version of PowerDVD player supports 96/24. As they also produce recording software I would hope that the ability to record using consumer software is just around the corner.

    Otherwise I am still working through a list of DVD recording software titles that I got elsewhere on the net. Maybe one of them does it...

    List so far includes:

    Arcsoft ShowBiz DVD
    Cyberlink PowerDirector
    Cyberlink PowerProducer
    Dazzle DVD Complete
    DVD Junior
    DVD Menu Studio
    DVD X Maker
    DVD-lab
    dvdauthor 0.6
    Easy CD & DVD Creator 6
    IfoEdit
    Magix Movies
    neoDVD
    neoDVDPlus / neoSTUDIO
    NeroVision Express
    Pinnacle Impression DVD-Pro
    Pinnacle Instant Video Album
    Pinnacle Studio version 8
    Sonic Drag'n Drop CD+DVD
    Sonic MyDVD
    Sonic MyDVD Deluxe
    SpruceUP
    SVCD2DVD
    TMPGEnc DVD Author
    Ulead DVD MovieFactory
    VideoPack
    WinDVD Creator
    WinOnCD
    WinProducer 3 DVD


     
  14. wilkes

    wilkes Regular member

    Joined:
    May 31, 2003
    Messages:
    954
    Likes Received:
    0
    Trophy Points:
    26
    Yes, you can put 24/96 stereo PCM audio onto a DVD-Video disc for playback in a standard DVD player. Unfortunately, it's not quite that simple as you need an authoring application that also supports 24/96 audio. Most of the "budget" apps just don't do this, and a lot of them will only give you 16/48 audio.
    You also have to include visual content, even if only a single still image or videoblack or you will not conform with the DVD Specs.
    Check & see if your player supports DVD-Audio, as that is your best bet. Armed with a copy of discwelder (www.discwelder.com) you will be fine.
    For DVD-Video, the app you need is either Sonic Scenarist or Adobe Encore DVD. Encore is much cheaper, but is far simpler to use.
    To begin with, go to www.dvddemystified.com, and start to learn the specs, then you will have a better understanding of all the various DVD specs. Honestly, it's the only way to do this as DVD authoring is ridiculously complex and nothing like as easy as CD.
    It's a fun journey, though, and well worth the effort. I would also go and look at these 2 pages:
    http://emusician.com/microsites/mag...releaseid=&srid=11343&magazineid=33&siteid=15
    and
    http://emusician.com/microsites/mag...releaseid=&srid=11343&magazineid=33&siteid=15
    where you will get an introduction to Audio on DVD-Video,which is a different animal altogether from DVD-Audio.
    Hope this helps. If not, ask away and we will do our best for you.
    PS - if you do get a proper authoring app, I would remove all the junk burning apps you mention, as you will get serious conflicts with DVD burning & drivers being corrupted.
    Keep a DVD machine as clean as possible.
     
  15. maxg

    maxg Member

    Joined:
    Dec 16, 2003
    Messages:
    10
    Likes Received:
    0
    Trophy Points:
    11
    Thanks Wilks,

    I am investigating Adobe encore now. I realise I will need a slide show for this to work but that is not really a problem for me - I will be running this with the TV off anyway.

    Thanks also for the heads-up on the conflicts. I have already experienced this to a degree and am deleting apps as fast as I am picking them up.

    Hmmm - just looking at the product PDF and it talks about automatically encoding audio to Dolby digital - this is not what I want.

    Ah well - more searching...
     
  16. wilkes

    wilkes Regular member

    Joined:
    May 31, 2003
    Messages:
    954
    Likes Received:
    0
    Trophy Points:
    26
    I assume you are talking about Encore here?
    The auto Dolby Digital transcoding is not compulsory, just an option. You can leave it as LPCM if you choose to.
    Also, Encore will happily import Dolby Digital, in both Stereo & surround, and you can have both forms of Audio on a single timeline, giving the end user the option to choose.
    There is also the superb, and very cheap, TMPGEnc DVD Authoring app, at www.pegasys.com. They are the same guys that do TMPGEnc for encoding to M2V files. Again, well worth the money for the Pro version.
    Enjoy.
     
  17. The_Mule

    The_Mule Member

    Joined:
    Mar 20, 2002
    Messages:
    32
    Likes Received:
    0
    Trophy Points:
    16
    If sampling freq doesn't matter, then why do SACD/DVD-A versions of recordings (with proper mastering) get universally reviewed as better sounding than CDs?

    Can't be bit depth, because again, most modern SACD and DVD-A players only get to 18 to 20 bit resolution anyway, and that isn't enough improvement for what people hear.

    And if most modern masters don't take advantage of the larger bit depth available anyway because of compression, limiting, processing, then where is that improvement coming from?
     
  18. The_Mule

    The_Mule Member

    Joined:
    Mar 20, 2002
    Messages:
    32
    Likes Received:
    0
    Trophy Points:
    16
    >> and that isn't enough improvement for what people hear.

    Never mind. Realized 15 min after I wrote that, that it wasn't correct. For every bit you add, you double the resolution.

    But that still doesn't account for recordings that don't take advantage of the full bit length like I said. And why SACD/DVD-A's still are typically preferred to CD.

    And, this is a heresay quote, a dude over on HTF says that most people can't hear better than 18 bits anyway.

    Well, I still know that I want to get to 24/96, and now I know a little bit (sic!) more why. :)
     
    Last edited: Dec 18, 2003
  19. wilkes

    wilkes Regular member

    Joined:
    May 31, 2003
    Messages:
    954
    Likes Received:
    0
    Trophy Points:
    26
    I think you'll find that the 18 bit thing is referring to the actual performance of a lot of so-called 24 bit DAC's out there. Tomlinson Holman has a lot to say about this too - apparently a lot of stuff out there only has 18-20 bit performance once you factor in the noise and real world performance.
    Going back to the sample rate thing, it's interesting how the AES has gone on record as saying you only need an upper rate of around 60KHz to capture everything you can hear. Having said that, things get a little more complex as there is good evidence to show direct perception of frequencies as high as 90000Hz from "vibration transducers in direct contact with the mastoid bone behind the ear", but those frequencies are pitched by listeners as being in the 8 to 16KHz range. It also appears that the absolute upper limit for air conduction hearimg is no higher than 24KHz.
    From my understanding, the main difference between a sampling rate of 96 or 88.2 KHz and 44.1 is the filters used in the converters. Think about this - in a system based on, say, a 48KHz rate, a bell filter at 20KHz must go to zero at 24KHz, or you will get aliasing. This, as I said earlier, is the main argument for high sample rates. You double your bandwidth, and half of the new range is in the ultrasonic band so your errors - both sampling & quantization, if also using 24 bit, are inaudible. The aliasing is inaudible as it is all up around 30-40KHz range, and quantization errors are inaudible as the least significant bit is also well below perceptible limits. Also, all editing & processing such as EQ is performed at the higher rates, thus doubling the precision, before dropping back down to the rates required by the release medium. So even if you then drop down to 16/44.1 for CD-A, with proper dithering & noise shaping, you are getting a product that is sonically superior to one that was produced entirely at the lower rates. If, as for DVD-A, you can keep the wordlength & samplerates high, you are doing still better as there can be no aliasing or audible errors.
    SACD is flawed from the outset because of the enormous noise in the ultrasonic bands. Very high, I have a graph of it if anyone is interested. If I knew how to do it, I'd post it here for you. There are other arguments - from the AES - on why 1 bit DSD is totally unsuitable for high resolution work, not the least of which is the apparent anomalous nature of 1 bit Delta Sigma conversion. BTW, all your SACD discs were made, mixed & EQ'd etc. on a PCM system as you cannot do that in DSD.
    The arguments rage, as usual, and I'm losing my thread here so will stop for now.
    Back later, I've no doubt.
    Happy thoughts
     
  20. maxg

    maxg Member

    Joined:
    Dec 16, 2003
    Messages:
    10
    Likes Received:
    0
    Trophy Points:
    11

Share This Page