maxg Posted October 4, 2004 Share Posted October 4, 2004 Not that there have been too many responses to the previous posts on this - but for completeness here goes: Background for those that missed the previous posts: I have been playing with ripping CD's to MP3 on my PC and then copying the resulting files onto DVD's for playback on my main system. A variety of strange discoveries have been made during this process, namely: 1. It works - I had been under the impression that a DVD player, even if it supports MP3 does not support it on the DVD format - only on the CD format. This, at least for my cheapo machine - does not appear to be the case. 2. The sonic quality of the result is remarkably good and for some arcane reason (which I am still not sure I have gotten to the bottom of) better when the files are on DVD than on CD. 3. Whilst trying to discover the reasons for 2, above, I discovered that my DVd player provides LPCM output options of 48 Khz - 16 bit, 96 KHz - 24 bit and, most surprizingly of all 192 KHz - 24 bit. The continuing story: On friday I was all excited at the possibility that this unit ($70 remember) actually played DVDa in native format. 192/24 is beyond the normal DVD video standard - and I thought that this might be there for DVDa. It does not appear to be so. When I bought the SoundBlaster Audigy 2 USB sound card it came with a DVDa player and a sample disk. This sample disk is virtually unique in that is has only a DVDa layer (no support for playback in a normal DVD player) and so appears to be an excellent test source. This disk loads - but wont play - implying that DVDa is in fact not supported on my unit. The question therefore becomes - what the hell is this 192/24 option. I now think that this is indeed an upsampler built into the unit. Further, it seems that this only applies to source material coming from a DVD disk - and not from a CD. This, it seems, is the reason MP3 on a DVD sounds so much better than MP3 on a CD (same MP3 file in each case). Yet further, and this is the eye-opener. For most disks the MP3 copy sounds better than the original CD (i.e. not the MP3 on a CD - the CD itself). Again I put this down to the upsampling. So where are we now? In simple terms I have now downloaded to my PC harddisk around 9 gig of MP3's - some 90 odd CD's and have burned the resulting files (well all but about 5 albums) to 2 DVD+RW disks within the directory structure as created in "My Music" - the folder the Windows media player uses. I estimate that by the time I have finished all 400 odd CD's will be on 9 DVD's - at the very least with no loss of quality overall (although for a few recordings they may not be as good). Another thing I am not sure about is how common it is for a DVD player that supports MP3 to be able to play them back from a DVD disk. I plan to take one of my disks off to an audio shop at some point at test out a bunch - just for interest. It would also be interesting to see how portable the quality is. I do not know how many of these cheap and cheerful units are doing upsampling on DVD source material - maybe they all do. If so - this is something I will be recommending for others to try. If not - well arent I the lucky one!! Someone please post a comment or 2 - just to let me know I am not talking to myself.... Quote Link to comment Share on other sites More sharing options...
doudou Posted October 4, 2004 Share Posted October 4, 2004 you re not talking all alone what you ve noticed is quite amazing: a compressed sound is better than the original source. don t you think it s quite strange? Quote Link to comment Share on other sites More sharing options...
maxg Posted October 4, 2004 Author Share Posted October 4, 2004 ---------------- On 10/4/2004 5:28:20 AM doudou wrote: you re not talking all alone what you ve noticed is quite amazing: a compressed sound is better than the original source. don t you think it s quite strange? ---------------- Strange!!! That doesnt cover it in my book - truely bizzare - which is why I am thrashing around trying to find an explanation for it. I really dont have a clue why this is happening - other than this possibility (which is entirely guesswork on my part) that there is some upsampling going on - but only for DVD. If that is wrong I am stumped. I a mdesperately hoping someone else will chip in with a why.... Quote Link to comment Share on other sites More sharing options...
hwatkins Posted October 4, 2004 Share Posted October 4, 2004 hello Max - I probably can't answer anything for you but I have a thought on the 192/24. If I get things correct, your last DAC used in the system chain has to have the 192/24 capability. For instance, with my system if I want 192/24 capability I would have to come off the analogue out of my transport/DAC - if I send the signal digitally (let's say Lpcm) to the amp DAC it is 'downsampled' (I may have made that word up, but it fits) to the highest capability of the DAC say 48 or 96khz. That said, there is probably a setup issue - regardless of sampling, your downstream DAC should be handling the signal. As to the MP3 sounding better on the DVD - I really can't say. Is it possible that in your setup the DAC solution for DVD is better than for CD? I say this because when I went to the DAC in my new CD player (granted the replaced player was very old) instead of the DAC in my amp I got a significant improvement in sound quality. The same wasn't true with my DVD arrangement. I have not had your experience with MP3 - quite the opposite, the sound quality is obviously different to the bad to my ear. Take all this at face value - i am just starting to get transport skinny to see what improvements I can make. Quote Link to comment Share on other sites More sharing options...
maxg Posted October 4, 2004 Author Share Posted October 4, 2004 "I have not had your experience with MP3 - quite the opposite, the sound quality is obviously different to the bad to my ear." OK - lets jsut clarify here. MP3 on a CD is NOT as good as the original CD. MP3 on the DVD is not only better than on the Cd is seems to be better than the CD original. In other words my findings with CD based MP3 concur with your own. Its on the DVD that strange things are happening. I guess I am proving that the DVD audio stream (whatever the type) is handled differently within the machine from the CD audio stream (again whatever the type). All I want to do is find out how - or why - or what. On the other hand - not being one to stare a gift horse in the mouth - I am furiously ripping all my CD's to MP3 as fast as I can. whatever the reason the observation on quality stands (at the moment) and the sheer inprovment in functionality is worth it on its own. Quote Link to comment Share on other sites More sharing options...
powdermnky007 Posted October 4, 2004 Share Posted October 4, 2004 You may want to look at this max. http://flac.sourceforge.net/ It's called flac.. its a lossless encoding method. so your file will be the same as the original cd. The files are larger though (around ~250mb per cd) Quote Link to comment Share on other sites More sharing options...
ben. Posted October 4, 2004 Share Posted October 4, 2004 Don't worry too much, Max. Tomorrow you'll think they sound worse than the original. Quote Link to comment Share on other sites More sharing options...
DrWho Posted October 4, 2004 Share Posted October 4, 2004 ---------------- On 10/4/2004 3:12:02 PM bclarke421 wrote: Don't worry too much, Max. Tomorrow you'll think they sound worse than the original. ---------------- haha, i was just gonna say the same thing! i don't see how upsampling increases the audio quality. let's see if i can provide a visual representation: Let's call this your 44.1kHz bit rate . . . . . . . . . . . . . When recorded, the DAC is recording a voltage for every dot. Each dot is an incremental segment of time (every 1/44,100th of a second). In an 88.2kHz bit rate, the number of dots doubles: ,,,,,,,,,,,,,,,,,,,,,,,,, (using commas to stress the point later on) When upsampling from 44.1 to 88.2, you really aren't gaining any extra information. your bit rate essentially looks like this: , , , , , , , , , , , , , even though your data looks like this: ,,,,,,,,,,,,,,,,,,,,,,,,, The reason for this is because every second comma is going to have the same value as the comma preceding it. In other words, every voltage recorded is going to be recorded twice. ok enough rambling, i hope that makes sense. if not, i can get more intense with real pictures. I've got an experiment for you before you get carried away putting MP3's of your CDs on your DVDs. Try recording your cd audio onto a dvd and compare the quality between the mp3 on a dvd versus the cd on a dvd. Btw, what is your playback system? You've been talking a lot lately about using headphones and never getting to listen to your real speakers at loud volumes (or am i totally off my rocker?) At quieter volumes, compressed music will tend to sound better because you can actually hear everything and compression makes it seem louder than it really is. Another factor, though this is a long shot is that MP3's start cutting out any information above 10kHz. I personally find that in the majority of recordings I've heard, that any information in the 8kHz range and up is just plain annoying sounding. It's like albums were recorded with some old guy behind the board who's got old ears and is cranking those frequencies when they don't need to be. So an mp3 is almost a natural EQ in that sense. Another factor is that the DAC on your dvd player is definetly much better. I'm not sure if there's solid scientific reason, but different DAC's sound different and some much better than others. I personally am a huge fan of all the 1 bit DACs. I find that a good DAC can be as drastic as the difference between CD and SACD (or maybe even mp3 and CD), though it's a different kind of distortion that's causing the difference. Quote Link to comment Share on other sites More sharing options...
scriven Posted October 5, 2004 Share Posted October 5, 2004 I agree with the good Doctor. Upsampling just means that the device is "making up" data to fill in the holes in the higher data rate until it gets to the next "real" data point. While it could be that you like the upsampling algorithm that player uses I dont think that can account for the kind of differences you are describing. Unfortunately, I can't come up with a reasonable theory as to what is causing the differences you are describing. (But rest assured, we are out here listing.) As Mr. Spock used to say when he didnt know what was going on - Fascinating! Quote Link to comment Share on other sites More sharing options...
maxg Posted October 5, 2004 Author Share Posted October 5, 2004 Well it is nice to know people are reading what I write. Thanks for the replies. Just to answer a couple of the points made: Powder, Flac may be the best thing since sliced bread - but I dont think my DVD player will understand it and so the point of the whole exercise would be lost. BC, Yup - that is what I was expecting last night - but it didnt happen - see below. Dr. Who / Scriven, I cant pretend to understand how upsampling works but my guess would have been that it takes 2 adjacent values (sums of bits rather than bits themselves) and then generates an inbetween value - smoothing out the resulting sound. Boy - is that a guess!! Anyway - as I understand it upsampling DACs are a fairly common approach to trying to improve the quality of digital audio - so I still think this is the most likely source of the sonic gains. Update I spent 3 hours listening last night to a range of albums and another couple of hours on the phone to Tony. Before I get to the Sonic results Tony came up with an interesting idea. MP3 is a lossy compression - it is supposed to remove all the sub-sonic crap you cant hear anyway. He is of the opinion that it also removes stuff you can hear - even at the higher bit rates. Now here is the interesting part - removal of some of the sonic information might actually produce a resulting sound that does indeed appear to be more pleasant - subject to the surrounding equipment - and not in all cases. This may be the effect I am observing here: Last night I listened to the following on the original CD and on the recorded MP3 files on DVD. The results are a shock!! 1. American 111 Solitary Man - Johny Cash 2. Here's to Ben - Jacintha 3. History - Michael Jackson (disk 2) 4. Immaculate Collection - Madonna 5. The Kick Inside - Kate Bush 6. The Legend of 1900 - Original Soundtrack 7. Pictures at an exhibition - Mussorski 8. 12 Spanish Dances - Z 9. Ultimate hits Collection - Ray Charles (disk 1) If asked before testing I would have predicted that all the good original recordings would be better than the MP3's - especially for the detailed voice tracks from Jacintha and Cash. Failing that I would have predicted some consistancy with either all poorer recordings coming out worse and better ones better or vice versa. None of this is the case: Assessing the music as follows: (I hope this table works as well on the forum as it looks in this window) (C = Copy prefered, O = Original prefered, "-" = No discernable difference, E = Excellent, G = good, A = average, P = poor) 1 2 3 4 5 6 7 8 9 Impact C C C C O C - C C Realism - - - C O C C C C Voice C C C C O C* - - C Instrumentation - O C C O C C C C Soundstage width C C C C - C C - C Soundstage depth - O - O - C O - C Image size C O - C O C O - C Clarity C C C C C C C C C Initial recording E E E G A G A E P Quality Overall Preference C C C C O C C C C As you can see from the above the only case thus far where the original sounds preferable over the copy is for Kate Bush - The kick inside. This album has a fairly unique sound - and the added clarity really worked against the copy here - her voice was thin and small - as was all of the instrumentation. Conversely - the Ray Charles recording, whilst still not sounding good even as a copy, was so resoundingly better than the original it was hard to believe they were from the same recording. This was the only CD that was primarily mono - I dont know what that means for the purposes of this test. The real shocks were the Jacintha and the Cash. Both of these are excellent recordings (the Jacintha is actually an SACD - I was listening to, and recorded from, the CD layer). The "improvement" in sound was far more dramatic for the copy than I remember (dangerous ground audio memory - but there you go) for the SACD layer Vs the CD layer on my old Sony NS900 player. Of course much of the "improvement" we should put down to personal taste - until such point as I can prevail on others to sit down and undergo the same test. In fact - the change for the Jacintha was very much in the opposite direction to going from CD to SACD. SACD seemed to sound lusher and warmer - but with some loss of detail compared to the CD. The MP3 track went the other way. Everything was cleaner, clearer and crisper. Soundstaging benefited (as a result?) with channel separation being quite breath-taking. Summary to date I am not confident enough to say, yet, that the copies are better than the originals - merely that they are preferable to my ears. The thought that removal of some sonic information may actually improve the result runs against the grain, but, as I prefer vinyl to CD, and according to the techies that contains less info than a CD, maybe it is not such a stretch. Could it really be that the efforts to improve CD by going into ever higher bit resolutions and sampling rates is actually the opposite of what is needed. I wonder if CD actually contains too much info as it is - rather than too little, and that that is the real problem with its sound. Just musing on this one - prepare for the usual about face from me in the near future - as my tag line suggests. Quote Link to comment Share on other sites More sharing options...
maxg Posted October 5, 2004 Author Share Posted October 5, 2004 Hmmm - table doesnt look so good - sorry about that - looked great in the rich text window - but I guess you can follow the snakes down. Quote Link to comment Share on other sites More sharing options...
gerbache Posted October 5, 2004 Share Posted October 5, 2004 A quick pedantic note on the way upsampling works: Even though you're right, no new information is added to the signal when you upsample, it doesn't just double each value. It performs an interpolation between the two values, such that rather than being the same as the value preceeding it, it'll be in between the previous and next value. I haven't looked much into the technical method of how they upsample in these players, but they're probably doing some relatively sophisticated methods of interpolation that may have some impact on the sound at the output. It's also worth pointing out that many upsamplers out there nowadays aren't increasing the frequency in integer multiples, so it has to provide a different method of mapping the new values, which I will almost guarantee is some form of interpolation. In any case, upsampling can actually have more of an effect on the sound than simply doubling the sampling frequency, even without any "real" extra information. Whether this is an improvement or a detriment to the sound would depend on how they do the interpolation, but you're right that upsampling doesn't really create any more data from the original source than was encoded on the CD. Quote Link to comment Share on other sites More sharing options...
gerbache Posted October 5, 2004 Share Posted October 5, 2004 I should have read Max's post before making that response..his explanation of how upsampling works is actually spot-on, and much simpler than what I said. In a nutshell, an upsampler smoothes out the spaces between each sample; if it does a particularly good job of it, it could potentially have a benefit on the sound output. Quote Link to comment Share on other sites More sharing options...
DrWho Posted October 5, 2004 Share Posted October 5, 2004 ---------------- i cant pretend to understand how upsampling works but my guess would have been that it takes 2 adjacent values (sums of bits rather than bits themselves) and then generates an inbetween value - smoothing out the resulting sound. Boy - is that a guess!! Anyway - as I understand it upsampling DACs are a fairly common approach to trying to improve the quality of digital audio - so I still think this is the most likely source of the sonic gains. ---------------- ---------------- On 10/5/2004 7:04:04 PM gerbache wrote: A quick pedantic note on the way upsampling works: Even though you're right, no new information is added to the signal when you upsample, it doesn't just double each value. It performs an interpolation between the two values, such that rather than being the same as the value preceeding it, it'll be in between the previous and next value. I haven't looked much into the technical method of how they upsample in these players, but they're probably doing some relatively sophisticated methods of interpolation that may have some impact on the sound at the output. It's also worth pointing out that many upsamplers out there nowadays aren't increasing the frequency in integer multiples, so it has to provide a different method of mapping the new values, which I will almost guarantee is some form of interpolation. In any case, upsampling can actually have more of an effect on the sound than simply doubling the sampling frequency, even without any "real" extra information. Whether this is an improvement or a detriment to the sound would depend on how they do the interpolation, but you're right that upsampling doesn't really create any more data from the original source than was encoded on the CD. ---------------- Correct me if I'm wrong, but I think the added interpolated value (if it exists) would cause the same resultant waveform on both the original and upsampled versions. To test this, you could record some music into a program that shows you the waveform and then take a screen shot. Then upsample the music and take another screen shot and then overlay the two and compare the differences. I'm 90% certain that you'll get the same waveform, but perhaps someone that understands the digital to analog conversion process better might be able to chime in and explain what's happening. I've got some better recording equipment available to my disposal so I will try recording the output of upsampled music and comparing that to the original waveform as well. It's all a matter of time (which I don't have much of nowadays). I wonder what kind of information on the web might be found on this topic...that might be a better approach. I think the idea behind the mp3 subtracting information (less is more?) and the quality of the DAC have more to do with the differences heard than any differences with the upsampling (and interpolation). What happens if you record the mp3 back to cd red book audio? Also, what happens if you record the mp3 onto a dvd at a 44.1kHz bitrate? Quote Link to comment Share on other sites More sharing options...
timbley Posted October 5, 2004 Share Posted October 5, 2004 ---------------- Yet further, and this is the eye-opener. For most disks the MP3 copy sounds better than the original CD (i.e. not the MP3 on a CD - the CD itself). Again I put this down to the upsampling. ---------------- That's fascinating. I've yet to hear an MP3 that I actually liked better than the original CD. My Panasonic recievers have some kind of upsampling mode that includes 4 different effects. My computer sound card will also upsample to 96/24 and send that to the recivers. Trying these things, I don't really hear anything happening that gets me exciting. The mp3s still sound like mp3s. I might have to try this DVD thing though. Quote Link to comment Share on other sites More sharing options...
maxg Posted October 6, 2004 Author Share Posted October 6, 2004 ---------------- On 10/5/2004 9:14:23 PM timbley wrote: ---------------- Yet further, and this is the eye-opener. For most disks the MP3 copy sounds better than the original CD (i.e. not the MP3 on a CD - the CD itself). Again I put this down to the upsampling. ---------------- That's fascinating. I've yet to hear an MP3 that I actually liked better than the original CD. My Panasonic recievers have some kind of upsampling mode that includes 4 different effects. My computer sound card will also upsample to 96/24 and send that to the recivers. Trying these things, I don't really hear anything happening that gets me exciting. The mp3s still sound like mp3s. I might have to try this DVD thing though. ---------------- Timberly, I am definitely only observing this effect on DVD disks - not on CD (played again last night). The improvement still does not match vinyl - but as I am doing this entirely with free software and a goon's "let the wizards make the decisions" approach there may be yet more scope for improvement as I get more into this. All, Just as a matter of interest there is a new Pioneer all in one player out (DV 578A) that not only plays back SACD, DVDa, DVD, DVD-R/RW, DVD+R/RW, CD, MP3 et al is also supports MP3 on a DVD disk just like my unit does (reviewed on CNET somewhere - and they specifically mention the MP3 on DVD thing). I do not know if it upsamples like my unit does. Actually I am still not sure my unit is upsampling the output through the RCA's - it is all just guesswork. Retail on the above is $199 - and you can find it on-line for less (around $150 I believe). T'aint in Europe yet so I cant play with it. If anyone is looking for a player at this price level it might be an interesting thing to look at - and then to try out what I have done here - MP3 @ 256 Kb onto a DVD+RW. Would love someone else to report on this - could confirm that I am not insane - in what I am reporting here at least... Quote Link to comment Share on other sites More sharing options...
sunnysal Posted October 6, 2004 Share Posted October 6, 2004 Max, At first blush it would seem to me you are just digging the type of distortions (don't take that wrong) created in your digitization process. In that you like what happens when you take the original, digitize it and burn it to DVD using what ever specific algorithms are used during that multi-step process. You obviously don't like the distortions produced when you do the same chain to CD. More power to ya, probably it will not result that others will have the same tastes and feel the same way. Some fun testing could be done by first listening to the files before burning them to any media, play them off the HD through a good sound card to headphones, I did this when trying to decide which system to use (MP3, WAV, etc.) and found BIG difference in what I liked (I ended up using windows new lossless). However, when I burned to CD or to my RAM player I could not "get scientific" I had the different technologies of the players to contend with. You may have similar problems trying to track down the source that is creating the3 sound you like, not to mention I have to use different burners and burn software between CD and DVD, do you?. So you burn the ame files to two different media and like one better? perhaps it is the media or the burner, or something else, what fun! I hope others will try this and chime in so we can see if there is a trend. warm regards, Tony Quote Link to comment Share on other sites More sharing options...
maxg Posted October 6, 2004 Author Share Posted October 6, 2004 Tony, Much the conclusions I am coming to on my own. I prefer this type of distortion over the type I get from a CD. As it happens all burning is done for me on the one unit. Further, in order to avoid any issues of possible re-coding going on at the burn process itself I am creating all the files on the hard disk and then using a shareware piece of software that just copies files onto the media - be it CD or DVD. It seems to my simple brain that the source of these different distortions is in the playback chain - even though the playback machine is, again, the same for both. I therefore concluded that is was upsampling - applicable to LPCM but not, presumably PCM, that was causing the different observations. I could be wrong. As I am currently on a less is more kick I am wondering whether the striping out of the - theoretically non audible information is actually cleaning up the sound making it appear cleaner and better - and dare I say more akin to vinyl to my ears. On the other hand we could forget all of this guff and just say - hey cool - 35-40 albums on a single disk - that is the way to get hold of your collection. By the by - I have tried output straight from my hand held (which seems to sound better than the portable computer itself) and that was terrible in comparison to the CD, let alone the DVD versions. the playback unit is definitely doing its thing - whatever it is... Quote Link to comment Share on other sites More sharing options...
sunnysal Posted October 6, 2004 Share Posted October 6, 2004 yes max absolutely it could just be the upsampling, or something else! lol! but i agree it is great to get great music, and so much of it, form just one disk! in the end, if that is what you end up with but are not sure why...who cares? enjoy! regards, tony Quote Link to comment Share on other sites More sharing options...
sfogg Posted October 6, 2004 Share Posted October 6, 2004 "When recorded, the DAC is recording a voltage for every dot. Each dot is an incremental segment of time (every 1/44,100th of a second). In an 88.2kHz bit rate, the number of dots doubles:" That isn't really what occurs. I'll write this more on the idea of changing the recording sampling rate whihc would allow more 'info' to be captured. The number of 'dots' per sample stays the same no matter if you record at 44.1kHz or 88.2 kHz. The number of 'dots' is the bit depth, not the sampling rate. IOW for a CD at 16 bits there are 16 dots per sample. If you increase the sampling rate to 88.2kHz there are still 16 dots per sample. The dots tell you the dynamic range available of the system. Each dot is 6dB basically. What there are more of is more samples per second. The number of samples doubles in the same amount of time. All this gives you is the ability to record music to twice the analog bandwidth as before. At 44.1kHz the top end limit of the system will be just under 1/2 the sampling rate. Kicking up the sampling rate to 88.2kHz means you can now capture to about 44kHz. Below 22kHz there is no difference in how the material is captured for 44.1kHz or 88.2kHz sampling rates. Shawn Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.