|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
In Reply to: You probably missed it...(link) posted by Frank.. on June 16, 2004 at 05:39:56:
*** Only the dsd source has a missing range of sample sizes missing in the histogram. (Although there are counts near -85dB!)This really needs further investigation. ***
It has been investigated, and I've since posted about it (can't remember when, though). The "missing samples" phenomenon is quite common for recordings that are "noisy". i've replicated it in recordings of both LPs and DVD-As so it's not unique to SACD.
*** I do too, but that turn out to be impulses from scratches mostly. :-( ***
Well, maybe you do need to upgrade your turntable. as i've said, i am getting full dynamic range of conservatively 80-90dB all the way to 20kHz and beyond. i'll post some graphs in my next article.
Thanks for the MP3. can you tell me the exact timings of your recordings - i would like to compare against my system on the weekend.
Follow Ups:
***Well, maybe you do need to upgrade your turntable. as i've said, i am getting full dynamic range of conservatively 80-90dB all the way to 20kHz and beyond. i'll post some graphs in my next article.***80..90 dB is practically 'impossible'. Dynamic range is reduced with a compressor for the disc cutter. It's also a possibility that your element arm combination/setting overshoots on transients.
You need a test disc to check if that's the case.***Thanks for the MP3. can you tell me the exact timings of your recordings - i would like to compare against my system on the weekend.
***I'll see what I can do. I have placed markers at the 'candidate silence' locatations in the track but not at the exact spots that are used in the mp3 sample.
The first and the last locations are fairly easy to spot though. :)
Note that I sampled the left and right front channels of the multichannel mix. I did this intentionally because usually there is ambiance added in a stereo mix picked up with an additional set of mics further out in the venue. The stereo mix could also be made from a pair of spaced omni's.(AB)
Ambient surround recording often use more 'discrete' mic setups with cardiods pointing in 5.1 directions.
There is a good chance the front mics pick up less hall ambiance and deliver better defined sonics from the direct sound field.
This way it's easier to seperate the noise from the ambient sounds.
*** 80..90 dB is practically 'impossible'. Dynamic range is reduced with a compressor for the disc cutter. It's also a possibility that your element arm combination/setting overshoots on transients. ***I must admit - I admire your ability to "spin" just about anything to suit a conclusion that you want to make. I particularly like the way that if you measure a limitation then it's proof that the medium is at fault, but if someone offers a counter example then their equipment is at fault :-) Have you considered a career in PR? The iraq situation needs people like you right now. :-)
"Impossible"? Evidence, please. Your repeated mention of disc cutter limitations just doesn't "cut" it - given that Discrete Quad LPs produced in the 70s require frequency response to 50kHz, so the technology was available even then.
And can you elaborate on exactly how a "passive" stylus tracking grooves can somehow cause "overshoots on transients"?
Frank
***I must admit - I admire your ability to "spin" just about anything to suit a conclusion that you want to make. I particularly like the way that if you measure a limitation then it's proof that the medium is at fault, but if someone offers a counter example then their equipment is at fault :-) Have you considered a career in PR? The iraq situation needs people like you right now. :-)***I was only pointing out a *possible* flaw in your arm/element setup.
There are testrecords with a test signal approaching a 'squarewave' as best as it could be reproduced by a mechanical analog system.
With an oscilloscope the overshoot/resonance can be checked.You have sampled vinyl. Just view the effect of a loud scratch by lookin in your wave editor to see how arm/cantilever resonance come into play.
As in the case of the missing samples. You didn't present evidence of missing samples with pcm or cd at the time.
It's clearly because of noise and the method used by the software to calculate the graph.
It isn't counting sample size occurances it's averaging them.
Due to the noise the the 'missing samples region' shows little because it averages out to near zero or below a treshold set in the software's algorithm.However it's still a valid indication that low level resolution might be poor.
***"Impossible"? Evidence, please. Your repeated mention of disc cutter limitations just doesn't "cut" it - given that Discrete Quad LPs produced in the 70s require frequency response to 50kHz, so the technology was available even then.***
I have had a guided tour in a vinyl pressing plant with it's own cutting facility.
It was clearly explained why this dynamic range compression and hf frequency roll off is needed.It was also explained at an informal lecture I attended by veterans in this business who have cut thousends of disc.
Do some research and you'll find out that it's the simple truth.
I simply refuse to buy into the current audiophile trend to blowup vinyl's capabilties into mythical proportions.
It's just a nice sounding format that served it's purpose in it's day.This dynamic range limiting isn't such a bad thing as it's contributing a lot to the particular charm of vinyl playback.
***"Impossible"? Evidence, please. Your repeated mention of disc cutter limitations just doesn't "cut" it - given that Discrete Quad LPs produced in the 70s require frequency response to 50kHz, so the technology was available even then.***
Now you are confusing extended bandwidth with dynamic range capabilities.
'Discrete' Quad LP's. A joke I presume. :) Quad LP's used some kind of matrix processing or carrier wave technology to encode the mc in the analog domain.
The technology was available but in the end hardly practical and difficult to implement.***And can you elaborate on exactly how a "passive" stylus tracking grooves can somehow cause "overshoots on transients"? ***
It's a mechanical device. The mechanics are comparable with that of a car's suspension hitting a pothole or bump in the road.
A more sporty trim and you feel every dent in the road (and ruin your back) or a more limo type setting where you can become seasick.Sample a loud scratch and look how your setup reacts.
*** However it's still a valid indication that low level resolution might be poor. ***If that is so, it is poor on ALL formats equally: LP, CD, DVD-A and SACD, since the behaviour can be replicated on recordings from all three formats.
*** Do some research and you'll find out that it's the simple truth. ***
Actually, i have done some research and what you say doesn't appear to be the simple truth. it is true that there are many contraints and limitations in pressing vinyl and it is possible to make a bad LP - there are many examples in the 70s and 80s due to commercial constraints of putting as much music as possible onto a side that filtering and compression was necessary.
However, there are also ways to avoid or mitigate those limitations. And given that LP is a "dead" format these days, LPs pressed for audiophiles are done by who do take the extra care.
*** Now you are confusing extended bandwidth with dynamic range capabilities. ***
No I wasn't. You were the one who said it was "impossible" to reach 20kHz at full dynamic range. but clearly the technology to do so was there in the 70s.
*** The technology was available but in the end hardly practical and difficult to implement. ***
So then it's not impossible to have extended frequency response and dynamic range, then, as you originally claim?
As for practical/difficult, my father had a Quad capable system, supporting both matrixed and discrete Quad LPs. you do need a special stylus for the discrete quad lps, but they were no more expensive than a reasonable stereo only stylus. we played quad for many years. it was quite practical and no more difficult than stereo. And even when quad died, a lot of the technology (such as the stylus shape and half speed mastering) was carried over to improve stereo LPs.
*** Sample a loud scratch and look how your setup reacts. ***
I have. It looks exactly like a loud scratch.
***No I wasn't. You were the one who said it was "impossible" to reach 20kHz at full dynamic range. but clearly the technology to do so was there in the 70s.***No you didn't get my point or are twisting my words.
I said it is impossible to reach 20kHz at the full dynamic range *you* claimed. (80..90dB)Fancy audiophile pressings are limited by the cutting heads limitations just as much as normal pressings.
If hf content is pushed to hard then the cutting head gets into trouble.***So then it's not impossible to have extended frequency response and dynamic range, then, as you originally claim?
***Yep, flat response up to 50kHz? Simply not possible.
The encoded quad info in the disc beyond 20Khz had a lower level.
*** I said it is impossible to reach 20kHz at the full dynamic range *you* claimed. (80..90dB) ***And i supplied a counter example to your "impossibility"
you may not realise this, but not all cutting heads are equal. there are different types, and they have different limitations. also, 20kHz at full dynamic range is easy to obtain if you slow down to half speed mastering. widening the spacing between grooves will also help.
just because you have visited one pressing plant and attended one seminar means you know everything there is to know.
by the way, DVD-A has limitations that require filtering as well. 96/24 5.1 channels require MLP encoding to fit into maximum bitrate of 9.6 Mb/s. however, MLP assumes it is able to achieve a certain level of compression.
Some program material cannot be compressed easily and generate "illegal" files. For example, by definition white noise cannot be compressed, so 96/24 5.1 of white noise will generate an illegal MLP file. The solution is - guess what: filtering the high frequencies.
SACD avoids this issue by specifying a higher maximum transfer rate for the player: 16 Mb/s. so if the material cannot be compressed by DST, it is still legal and able to be transferred to SACD without filtering. you do reduce the maximum playing time though.
***by the way, DVD-A has limitations that require filtering as well. 96/24 5.1 channels require MLP encoding to fit into maximum bitrate of 9.6 Mb/s. however, MLP assumes it is able to achieve a certain level of compression.***'require filtering'? Apart from very rare occurring events in normal music content other filtering requirements than that to fulfill the nyguest criteria are not neccesary.
***Some program material cannot be compressed easily and generate "illegal" files. For example, by definition white noise cannot be compressed, so 96/24 5.1 of white noise will generate an illegal MLP file. The solution is - guess what: filtering the high frequencies.***
White noise at full level is very usefull in a music format.
Filtering at 30kHz instead of 40kHz probably solves the issue.
No big deal.***SACD avoids this issue by specifying a higher maximum transfer rate for the player: 16 Mb/s. so if the material cannot be compressed by DST, it is still legal and able to be transferred to SACD without filtering. you do reduce the maximum playing time though.
Obviously dst is less efficient in reducing the bitrate so a higher bitrate from the player is needed.
The maximum playing time is less than that of a dvda with mlp.
DVDA with mlp is a more flexible format. If the bitrate is too high than mild filtering can be used to remedy the problem.
Filtering hf is only needed during the difficult passage.Maximum playing time isn't compromised by a fancy ability to handle vary rare occurring musical signals or non musical signals like wideband noise at high levels.
Frank
you were very quick to criticize limitations in vinyl technology that require certain signals to be filtered or compressed in rare circumstances (and even then the problem can be solved by using half speed mastering or widening the groove spacing).but a very similar limitation is in DVD-A where certain signals require filtering before they can be mastered onto DVD-A.
This post is made possible by the generous support of people like you and our sponsors: