digital experiment

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi
A few weeks ago i start an experiment. I rip a cd with excellent recordings, I resample the file at 176.4 KHz with Wavelab and after that I burn the result on a DVD with Diskwelder crome, the result a DVD audio with 16bit 176.4 kHz LPCM information.
I play the result on a budget DVD Player like Panasonic S52 an the result was fabulous. A beautiful sound on budget equipment. The sound was different more clearly, the instruments more distinctly, more natural.
The system that i have at home and listen the experiments
Amp Onkyo A 8870, DVD player Panasonic S52, CD Player Sony CDP 515, Speakers JBL ATX 60, interconnect Ixos and speaker cable Ralcable Bimetal 4mm.

Question.
Is this a good way to listen music on budget equipment? These procedures make the same thing that expansive cd players do? The result is near to a good cd player?

I want to do something else now. I want to find a decent DAC and try to listen again the result of the experiment.

If some one tries the same experiment and have some good result or knowledge to share, please type on this post.

Have a nice weekend
 
Well, it is a good enough way to "uncompress" the over all headroom of the compressed CD files ... 44k uncompressed to 176k ... basically "undoing" the work of the publishers' mix down engineer to some extent.

Your results will of course be determined by the level of quality of the original ... sometimes the results will be very nice indeed, sometimes not ... I would speculate that vocals and acoustic instruments would show improvements, overly compressed electrified (and possibly distorted) instruments = not.

Try it on an album of acoustic music that is obviously degraded by the mix down process ... example: Eric Clapton's "Me and Mr. Johnson" is overly compressed acoustic guitar and vocals to the point of (possibly deliberate) introduction of distorted and clipped information === maybe this CD is beyond help ... maybe not.

:apathic:

" a decent DAC " ... IMOP would be a DAC capable of multi-channel 24bit / 192k (or 96k x 4 channels) ... something not normally found in USB connected devices, but there are some that can do it. A bi-directional (record & playback) device like this would do it ... http://www.m-audio.com/products/en_us/FireWireSolo-main.html ... or this: http://www.m-audio.com/products/en_us/FastTrackPro-main.html ... or this: http://rolandus.com/products/productdetails.aspx?ObjectId=731&ParentId=114 ... the idea here is to rip and play 24 bit audio capable of putting the DVD-Audio and DVD player scenario to its full potential.
:bigeyes:
 
What you actually do is to substitute the oversampling filter integrated in the D/A converter chip itself with a much better one processed offline on a computer. No wonder it sounds much better. Cheap players use cheap DACs with horrible digital filters, expensive players use, well, the best DACs market has to offer with mediocre digital filters. More info here.
 
You have introduces a lot of jitter by ripping, upsampling, burning the DVD, and playing "the original information" back on a cheap DVD Player - and you liked the sound: rounded, layed-back, "more analog-like".

A bit of jitter here and there (at correct spectrum) can do “wonders”...

Boky
 
Extreme_Boky said:
You have introduces a lot of jitter by ripping, upsampling, burning the DVD, and playing "the original information" back on a cheap DVD Player - and you liked the sound: rounded, layed-back, "more analog-like".

A bit of jitter here and there (at correct spectrum) can do “wonders”...

Boky

jitter is an analog phenomenon, it does not exist in the data itself.. ripping or copying from one digital media to another cannot introduce any jitter.. synchronous digital processing also cannot introduce any jitter.. asynchronous resampling using nowadays popular ASRC chips integrates the jitter into the resulting digital data, however there is no talk about ASRC in this thread, only synchronous processing is being applied.. burning data onto a CD or DVD also does not directly influence the conversion clock jitter.. only indirectly via coupling of servo action current spikes into another supplies of the player, including the DAC, oscillator and analog stage supplies.. it would be very easy to test this influence by burning the original 16/44.1 track next to the upsampled one on the same DVD-A.. I can guarantee you the result will be the same as with the original CD vs. upsampled DVD-A..

you are right though that the spectrum of the jitter is very important to the percieved sound, jitter amplitude in time domain doesn't tell much about it..
 
Extreme_Boky said:
You have introduces a lot of jitter by ripping, upsampling, burning the DVD, and playing "the original information" back on a cheap DVD Player - and you liked the sound: rounded, layed-back, "more analog-like".

A bit of jitter here and there (at correct spectrum) can do “wonders”...

Boky

Sorry if I seem ignorant, I'm quite new to this but I have a couple of questions about this.

First, I really really wonder how it comes that ripping a CD to wav introduces jitter. If i had a 650 MB textfile on a CD and copied that textfile onto my hard drive I expect all letters in that textfile to be 100% correct on the HD. Why is this different when ripping a Audio CD? In my knowledge there shouldn't be any jitter introduced in this procedure.

Then I wonder how upsampling would introduce jitter, what it actually does is what Glassman sugests, the jobb that the oversampler actually has. However since this isn't done in realtime the quality can be greatly improved.

As far as I understand, this is a great way to improve the experience of listening to your favourite CDs.
 
Hi folks.

Jitter effects on audio quality is a realtime issue only.
If you run a non realtime bit perfect application, such as ripping to file,
or transfering files, you won't have any issues, unless your Laser can't read a bit
properly. But that's not jitter.

Ripping to file doesn't introduce jitter. Writing to CD introduces jitter though.
The pits are written to CD with a certain Jitter by getting different distances to each other.
This of course causes jitter immediatly by the time the data are read-out,
converted and played in realtime from CD.

This is one of the reasons why the data should stay IMO on the harddisc.

If that is no option, there are just a few CDROMs out there, such as
Plextor Premium 2, which introduce very low jitter when writing CDs.
Many people report good results with these drives. With gold CDs and a premium drive you have a good chance to end up with better sound as before, since the pits are much better written to the gold surface then on the original. This is causing less jitter when reading-out again it realtime.

I would not call jitter an analog issue as pointed out by glassmann.
The digital data are getting inhomogenous before they become converted to analog. It is a purely digital cause for jitter. Analog jitter what's that supposed to be!
Jitter is getting appearant though, by the time the jittery stream is realtime converted to analog! You can also increase the jitter in the stream by doing
upsampling here and downsampling there, poor clocking, poor and noisy power supply of the crystal, converting from USB to SPDIF - you name it.


I do offline upsampling from 44,1 to 48khz with "Voxengo R8brain Pro" by myself.
The offline conversion is giving me great improvements, even when compared to Shibatch realtime SRC. Shibatch is most probably a great algorithm introducing no phaseshifts whatsoever. But doing it realtime will have certain effects on the jitter induced to the stream.


Cheers
 
Jitter effects on audio quality is a realtime issue only.
If you run a non realtime bit perfect application, such as ripping to file,
or transfering files, you won't have any issues, unless your Laser can't read a bit
properly. But that's not jitter.
agreed.. in fact laser quite often misreads a bit, but thats why there is EFM and two layers of CIRC so in the end we obtain correct data..

Ripping to file doesn't introduce jitter. Writing to CD introduces jitter though.
The pits are written to CD with a certain Jitter by getting different distances to each other.
This of course causes jitter immediatly by the time the data are read-out,
converted and played in realtime from CD.
not true! pit/land jitter is indeed present on every CD to lesser or greater extent, but this is not the reference signal by which the D/A converters are clocked.. signal from pickup has to undergo various processing and is buffered before being output to DAC.. the output as well as DAC is clocked by a free running usually crystal clock reference, which is not directly influenced by the pit/land jitter of the particular disk.. it's actually the other way around - the disk readout ie. spinning is adjusted to the free running crystal reference by the servo processor.. the only way how pit/land jitter or physical disk quality in general can influence the jitter at D/A convertion is via coupling of servoamp current spikes through the supplies of the player or even through RF radiation it generates.. so it is an indirect cause due to a faulty circuit design, not a direct cause!

This is one of the reasons why the data should stay IMO on the harddisc.
if the disk drive generates less interference to the rest of the audio system than it's better indeed..

If that is no option, there are just a few CDROMs out there, such as
Plextor Premium 2, which introduce very low jitter when writing CDs.
Many people report good results with these drives. With gold CDs and a premium drive you have a good chance to end up with better sound as before, since the pits are much better written to the gold surface then on the original. This is causing less jitter when reading-out again it realtime.
if you start with a physically perfect and regular disk, the servo will likely run smoother without abrupt actions and hence generate less interference, so yes, this can improve sound on badly designed CD players..

I would not call jitter an analog issue as pointed out by glassmann.
The digital data are getting inhomogenous before they become converted to analog. It is a purely digital cause for jitter. Analog jitter what's that supposed to be!
Jitter is getting appearant though, by the time the jittery stream is realtime converted to analog!
depends on your point of view.. i consider it an analog every time something doesn't work in a simple '1s are 1s and 0s are 0s' way, which is precisely the instant at which we convert data from digital to analog.. the clock is an analog signal and jitter is an analog issue, becouse otherwise one would say that digital clock is nothing but 1010101010101010 sequence and is always perfect in which case jitter does not exist..

You can also increase the jitter in the stream by doing upsampling here and downsampling there
tell me how..

poor clocking, poor and noisy power supply of the crystal, converting from USB to SPDIF - you name it.
of course, anything that makes the conversion clock worse is affecting the end result..

I do offline upsampling from 44,1 to 48khz with "Voxengo R8brain Pro" by myself.
The offline conversion is giving me great improvements, even when compared to Shibatch realtime SRC. Shibatch is most probably a great algorithm introducing no phaseshifts whatsoever. But doing it realtime will have certain effects on the jitter induced to the stream.
the Voxengo algorithm could just as well run in realtime with the same results, just like SSRC can work offline with the same results.. nearly all digital filters have completely linear phase response, with the exception of R8brain Pro in MinPhase mode, which is an IIR filter and introduces great amount of phaseshift..
again, you don't introduce any jitter via digital signal processing, only if the interference generated inside the computer would change dramatically under load caused by realtime processing, which to me is not quite probable..
 
Jitter has no effect when you are just in the digital domain.

1+1 = 2
2+2 = 4
3+3 = 6

It doesn't matter if you do those sums exactly 1 micro-second apart, 1 second apart, 1 year apart, or do 1 then get bored and go away for half an hour for a cup of tea and then come back to do the last 2. The answers will always come out exactly the same.

So long as the data is in the right order and bit correct the actual timing has no effect on the results.

Also something to note about pit/land jitter, remember the data on a CD is not stored in linear order. The data is interleaved.
 
I think that we are all talking about the same thing, only we might not understand each point of view.

Let's face it, digital signals is transfered in analog wires (well perhaps not in optical wires). But since digital data is coded into ONs and OFFs in these wires we need to code them using timescycles. To do that we use of high precision clocks that synchronises the sending and receiving chip, so that they work in the same speed. The clock, a particular wire or even a transmitting chip can introduce jitter in the pathway of digital data.

Now as long as the sending and the receiving chips is digital, a little jitter here or there is no problem because the chips generally understand each other anyway. If some chips become unsyncronised, and if something for this reason went horribly wrong somewhere, another layer ontop of the transport layer check the data and ask for a resend if it is corrupt.

However, a DAC chip is part digital and part analog. It stores up 24 bit (I am talking about a 24-bit DAC chip now, for the sake of simplicity) and when it got them all, it outputs the correct analog signal. So for a regular digital chip this doesn't matter because data arrives sooner or later anyway, and if somethings wrong, someone have the responsibility to check the validity and ask for a resend. But for the DAC chip it's critical that it steadily sums up and send out the correct analog signal. Therefore the digital stream that is sent to the DAC chip must be as jitterfree as possible.

There can be a number of reasons why the stream isn't jitter free when it ends up in the DAC. If the CD-player is stupid enough not to reclock the data properly before sending it to the DAC device, then ofcourse the quallity of the CD is important. But simple portable CD-players have shake protection memory and those would most certainly remove any jitter that might be written to the CD.

To sum up, jitter is a time phenomenon. In the digital world it can slow down data transferes, but in the end the data will turn up correct on the other side of the wire. When translated to analog signals this data stream must be very steady (jitter free). That means the digital pulses must come in as correct timing as they possibly can. So in a way, besides the analog out signal, the DAC chip works analogy internaly with time as the analog factor (strange centence).

However, there might be a second issue. If the digital pulses come correctly timed but with fluctuation of voltage or current, this might have an effect on the DAC chip aswell. But I'm not sure. I beleive that the DAC wait untill all 24 bit's have arrived, thus storing them. When it has them all, i think it utilises another power source to output the analog signal based on the value that the stored bit's sum up to. Could anyone shed some light on this? Fluctuation of this type might ofcourse affect the chip anyway.
 
fixerfrasse said:
ITo sum up, jitter is a time phenomenon. In the digital world it can slow down data transferes, but in the end the data will turn up correct on the other side of the wire.

VERY INTERESTING. Please define "correct on the other side" please! :D

------------------------------

Jitter accumulates over different stages on a realtime stream, if there is no stage involved being in charge to get rid of the jitter
as far as possible.
Take USB to SPDIF conversion just to give an example. Here are measurements clearly showing that the jitter almost doubles, when converting USB-SPDIF-I2S instead of doing it USB-I2S.
You'll hear that by the time the signal hits the DAC.


Regarding CD-Write Jitter:

You (glassmann) might want to read this.

http://www.cdrinfo.com/Sections/Reviews/Specific.aspx?ArticleId=6513&PageId=12

Cheers
 
soundcheck said:
Regarding CD-Write Jitter:

You (glassmann) might want to read this.

http://www.cdrinfo.com/Sections/Reviews/Specific.aspx?ArticleId=6513&PageId=12

Cheers

sure, thats the pit/land jitter I've been talking about.. however you have to realise that the self-clocking of the CD track is only used for synchronisation and raw data demodulation from the RF signal comming from pickup - it is not the master clock used for D/A conversion! I suggest reading through a few CD chipset datasheets, you'll understand..
 
Hi
I am glad. The subject seems to be interesting for the forum members but I think that it goes in the wrong directions.
Lets say that we have perfect rip,(for us the jitter is not important if we have 100% the same information as original). The result is 10 wav files , we process the data with a specialized program on a PC from 16bit 44.1kHz to 16bit 176.4 kHz after that we burn the result on a DVD R with another specialized program for DVD Audio. It is important to make a DVD Audio because the normal DVD Players can’t handle sample rates more than 96 kHz, only the DVD players that have DVD Audio futures can do this. So what do you think? Is the result better than a normal cd if we play on a budget DVD Audio player like Panasonic S52 or a low cost Pioneer? What to do? To pay 1000 euros or more for a decent cd player or process all the cd from our home to have a better sound on a decent audio setup.

A few weeks ago a guy sent me this link http://db.audioasylum.com/cgi/m.mpl?forum=digital&n=111123&highlight=Hallelujah&r=&session=

Anyone read this before?, any comments?
For most of us I think it’s important to have good sound quality in affordable stereo home system, not all the people have system that cost 30000 euros or more.
 
gabrielbecheanu said:

wav files , we process the data with a specialized program on a PC from 16bit 44.1kHz to 16bit 176.4 kHz after that we burn the result

Make it to 24 bit 176.4kHz, not 16b lest you lose accuracy of the filter results.

gabrielbecheanu said:

the result better than a normal cd if we play on a budget DVD Audio player like Panasonic S52 or a low cost Pioneer?

The only difference would be in the quality of the player's inbuilt reconstruction filter versus that of the off-line software filter. Either one can be better, although the software version stands more chance of factually being better.

Then it remains to be seen if this is audible.

What does not chance at all is the suspect quality of the budget player's DAC chip, power supply, and analogue output stages.
And that's where a lot of the sound gets lost.

gabrielbecheanu said:

A technical review in HFN by Keith Howard revealed that this program does some trickery to the signal that goes beyond mere high-quality sample rate conversion. This looked a lot like deliberately added distortion.
Again, some people may like it, some may not. Nothing wrong with that.
 
Werner said:

Make it to 24 bit 176.4kHz, not 16b lest you lose accuracy of the filter results..


Hi
Thanks for the advice.
I did some experiments also in 24 bit 176.4kHz.The software that i used for those experiments were : for ripping the cd : EAC, to resample the files : Wavelab5, for burning the result : Diskwelder chrome. Indeed, if you use a better DVD player the sound improves noticeably. For this test i also used a Denon DVD 3910 and AVR 4306.

Cheers
 
when versions abound ... subjective & off topic?

Of interest: 16bit v. 24bit v. vinyl, merchandising and marketing confusion prevail :

Consider the Beatles music album "Love" produced by George Martin, a remix from the original masters with some very interesting revisions and compilations:

"Love - Standard Edition" = "standard" audio CD (1 disc)
http://www.amazon.com/Love-Standard-Beatles-George-Martin/dp/B000JJSM4S

"Love - Special Edition" [EMI Import] = "standard" audio CD plus DVD-A version (2 discs)
http://www.amazon.com/Love-Special-Beatles-George-Martin/dp/B000JBXLOM

"Love - Special Edition" [Capital USA] = "standard" audio CD plus Audio DVD (2 discs)
http://www.amazon.com/gp/product/B000JJS8TM

"Love [Digipack With Bonus DVDA 5.1 Surround Sound - Apple Label!]" = standard CD plus DVDA (apparently with some video footage) (2 discs)
http://www.amazon.co.uk/Love-Digipack-Bonus-Surround-Sound/dp/B000JJS8TM

"Love" = "standard" CD (1 disc)
http://www.amazon.co.uk/Love-Beatles/dp/B000JK8OYU

"Love [Toshiba Import(?)]" = "standard" CD (1 disc)
http://www.amazon.co.uk/Love-Beatles/dp/B000JJSM4S

....
Typical Amazon review: " ... The mixing has, of course, been done on state of the art equipment and is beautifully crystal clear ..." You have to read deeply into these reviews to discover that, generally, those reviewers that panned the album listened to the "standard", 16bit CD and those reviewers that raved about the "crystal clear" sound and "impeccable" quality of the listening experience listened to the 24bit DVD-A version(s).

Even more interesting:

"Love Special Edition (CD + DVD) [Apple/Bea/EMI Doppel-CD]" = CD (SACD?) and DVD-A (no video apparently) (2 discs)
http://www.amazon.de/Love-Special-CD-+-DVD/dp/B000JJS8TM

!!
"Love [Vinyl]" = 12 inch LP Stereo recording (2 LP discs) - available February, 2007
http://www.amazon.co.uk/Love-VINYL-Beatles/dp/B000M06SU4

A plethora of benchmarks === so many formats, so little time. I have the 2nd from the top of the list (purchased from Amazon.com), and thoroughly enjoy this music. My wife plays it for special occasions and when we have visitors of our age group. I will be getting the vinyl for my collection, but may never play it except for the sake of comparison with the DVD-A. After hearing the DVD-A version, we do not play the CD version except in the car as it just does not sound right = no guts, obviously compressed, although not nearly as much as previous "standard" CD variations. (It is quite obvious that the folks at Amazon.com and Amazon.co.uk don't have a clue about what to do with this ... Amazon.de, may have or not.)

Recommendations: Get the Apple published version (check the labels before you buy) that fits your playback system, hopefully you have or eventually will get a DVD player with DACs capable of reproducing the full 24 bit data. (Either stereo or db5.1 sound just fine as long as its 24bit.)

I have a hunch that Sir George is as fed up with all the variations from all of the publishers as we are, releasing this album in as many formats as possible in order to get the reviewers to pay more attention to what's "real" and what's "imagined" about production quality verses costs verses hype and hoopala verses 16bit / 24bit comparisons. ... There is apparently a SACD version (1 disc) available in Europe. If anyone finds this, please advise.

I will eventually collect all versions and if nothing but for comparisons' sake. I will also expidite my searches for 24 bit DVD-A playback for my car.
:eek:
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.