Home |
Search |
Today's Posts |
#1
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
How can I improve the quality of a clip that suffers from smearing/
ringing artifacts due to low-medium bitrates? The clips in question are 48 KHz 128 kbps CBR/VBR and a couple at 96. I don't understand why those assholes can't downsample to 32 KHz if they're gonna use such low bitrates. |
#2
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
Industrial One writes:
How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. -- % Randy Yates % "Remember the good old 1980's, when %% Fuquay-Varina, NC % things were so uncomplicated?" %%% 919-577-9882 % 'Ticket To The Moon' %%%% % *Time*, Electric Light Orchestra http://www.digitalsignallabs.com |
#3
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
On 15 Okt., 02:41, Randy Yates wrote:
Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. No ****. I said "improve quality," not "replace." Btw, you should know that machines are capable of whatever we can already do. If I can imagine in my mind a high quality version of the audio without the ringing from a low quality source, a computer can do the same. |
#4
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
In article
, Industrial One wrote: On 15 Okt., 02:41, Randy Yates wrote: Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. No ****. I said "improve quality," not "replace." Btw, you should know that machines are capable of whatever we can already do. If I can imagine in my mind a high quality version of the audio without the ringing from a low quality source, a computer can do the same. It's not enough that you can *imagine* it; you must know *precisely* how to *do* it. Then you can teach a computer how to do it, only a lot faster than you can. Or alternately, the computer can indeed "imagine" it, but can no more deliver that as a usable output than you can "output" what you imagined. Isaac |
#5
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
On Oct 15, 4:02 am, isw wrote:
In article , Industrial One wrote: On 15 Okt., 02:41, Randy Yates wrote: Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. No ****. I said "improve quality," not "replace." Btw, you should know that machines are capable of whatever we can already do. If I can imagine in my mind a high quality version of the audio without the ringing from a low quality source, a computer can do the same. It's not enough that you can *imagine* it; you must know *precisely* how to *do* it. Then you can teach a computer how to do it, only a lot faster than you can. Or alternately, the computer can indeed "imagine" it, but can no more deliver that as a usable output than you can "output" what you imagined. Isaac I know, which is why I'm asking this group for suggestions. There must be a way, just like there's a way to improve the quality of low- bitrate DivX clips by applying a deblocking algorithm -- the most advanced out there cannot completely restore the original quality but still looks WAY better than if you left it alone. So what's my best option? To leave my song as it is? |
#6
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Industrial One" wrote in message
... On Oct 15, 4:02 am, isw wrote: In article , Industrial One wrote: On 15 Okt., 02:41, Randy Yates wrote: Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. No ****. I said "improve quality," not "replace." Btw, you should know that machines are capable of whatever we can already do. If I can imagine in my mind a high quality version of the audio without the ringing from a low quality source, a computer can do the same. It's not enough that you can *imagine* it; you must know *precisely* how to *do* it. Then you can teach a computer how to do it, only a lot faster than you can. Or alternately, the computer can indeed "imagine" it, but can no more deliver that as a usable output than you can "output" what you imagined. Isaac I know, which is why I'm asking this group for suggestions. There must be a way, just like there's a way to improve the quality of low- bitrate DivX clips by applying a deblocking algorithm -- the most advanced out there cannot completely restore the original quality but still looks WAY better than if you left it alone. So what's my best option? To leave my song as it is? Your best option is to go find a better original. |
#7
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Randy Yates" wrote ...
Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. Hence the term "lossy compression". Does I1 have a list of troll questions that he posts regularly? |
#8
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
Industrial One wrote: On 15 Okt., 02:41, Randy Yates wrote: Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. No ****. I said "improve quality," not "replace." There are different software decoders I think. Or is it just encoders Try some anyway.. Graham |
#9
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
Industrial One writes:
On 15 Okt., 02:41, Randy Yates wrote: Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. No ****. I said "improve quality," not "replace." Which is a ridiculous request. Sorta like "improving the quality" of the output of an 8-bit A/D. The noise is (or in your case, artifacts are) there to stay. -- % Randy Yates % "So now it's getting late, %% Fuquay-Varina, NC % and those who hesitate %%% 919-577-9882 % got no one..." %%%% % 'Waterfall', *Face The Music*, ELO http://www.digitalsignallabs.com |
#10
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Randy Yates" wrote in message ... Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. No ****. I said "improve quality," not "replace." Which is a ridiculous request. Sorta like "improving the quality" of the output of an 8-bit A/D. The noise is (or in your case, artifacts are) there to stay. Sure, but since he doesn't define what HE means by "improve", maybe he *can* do it. IF silence is an "improvement" (sure is in many cases IMO) then it's actually very EASY! :-) MrT. |
#11
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Industrial One" wrote in message ... On 15 Okt., 02:41, Randy Yates wrote: Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. No ****. I said "improve quality," not "replace." It depends on what you mean by quality and what Mr. Yates means by quality. Information in the original waveform was discarded in the encoding process, so fidelity to the original sound is irretrievably lost. All you can do now is fiddle with it to see if you can find some further distortion that is more to your liking. Btw, you should know that machines are capable of whatever we can already do. If I can imagine in my mind a high quality version of the audio without the ringing from a low quality source, a computer can do the same. I can imagine all sorts of things that no computer is or ever will be capable of. |
#12
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Industrial One" wrote from Goooooooooogle Groups...
Btw, you should know that machines are capable of whatever we can already do. Nominated for silliest remark of the year. But then it is only mid-October. If "I-1" keeps up the good work, he will make 1st Class Troll and give Troll Emeritus "Radium" a run for his position. |
#13
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
On Oct 15, 3:39 pm, "Richard Crowley" wrote:
"Randy Yates" wrote ... Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. Hence the term "lossy compression". Does I1 have a list of troll questions that he posts regularly? "I1..." I kinda like it, despite how robotic it sounds. On Oct 15, 9:56 pm, Randy Yates wrote: Which is a ridiculous request. Sorta like "improving the quality" of the output of an 8-bit A/D. The noise is (or in your case, artifacts are) there to stay. Bull****, the noise can be removed via noise-removal techniques and re- saved as 16-bit. On Oct 16, 12:46 am, "Mr.T" MrT@home wrote: "Randy Yates" wrote in message ... Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. No ****. I said "improve quality," not "replace." Which is a ridiculous request. Sorta like "improving the quality" of the output of an 8-bit A/D. The noise is (or in your case, artifacts are) there to stay. Sure, but since he doesn't define what HE means by "improve", maybe he *can* do it. IF silence is an "improvement" (sure is in many cases IMO) then it's actually very EASY! :-) MrT. **** you Mr.T. Didn't I tell you to stay outta my threads? On Oct 16, 5:22 am, "Chronic Philharmonic" wrote: "Industrial One" wrote in message ... On 15 Okt., 02:41, Randy Yates wrote: Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. No ****. I said "improve quality," not "replace." It depends on what you mean by quality and what Mr. Yates means by quality. Information in the original waveform was discarded in the encoding process, so fidelity to the original sound is irretrievably lost. All you can do now is fiddle with it to see if you can find some further distortion that is more to your liking. By quality I mean presentability. By running a smart deblocking algo on a low-bitrate DivX clip, do I "restore information?" Not exactly, but I interpolate/extrapolate the information I already have to make the video much more presentable and perceivably higher quality. How do you think your own mind can simulate a higher quality image of the one you seen on your grainy TV? Some information is "gone" but the information already present makes it obvious what would be there if it wasn't gone. Neural networks just aren't at the stage yet where it can restore images automatically without heavy human guidance. I'm asking if the same can be done for sound. Can I "de-smear" and "de- ring" it? If you insist I can't, then ok. Btw, you should know that machines are capable of whatever we can already do. If I can imagine in my mind a high quality version of the audio without the ringing from a low quality source, a computer can do the same. I can imagine all sorts of things that no computer is or ever will be capable of. Like how the members of the Fraunhofer committee back in 1984 thought consumer CPUs will never reach the stage to decode MP3s in real-time? |
#14
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Industrial One" wrote in
message How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You might be able do obtain a perceived improvement with filtering and noise gating. The clips in question are 48 KHz 128 kbps CBR/VBR and a couple at 96. I don't understand why those assholes can't downsample to 32 KHz if they're gonna use such low bitrates. Downsampling to 32 KHz can actually improve the results when you use such low bitrates. 32 KHz isn't that ugly of a sample rate - it allows some kind of frequency response up to about 16 KHz. Please remember that FM stereo pretty well tops out at 15 KHz. Some general rules for coding to low bitrates are to forget stereo and go to mono, and decrease the bandwidth as much as you can without losing too much intelligibility. Spoken word in mono with a 5 to 8 Hz bandwidth isn't usually all that bad, and classical music with 11 KHz bandwidth can often be quite satisfying. |
#15
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
Arny Krueger wrote:
"Industrial One" wrote in message How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You might be able do obtain a perceived improvement with filtering and noise gating. Such as? The clips in question are 48 KHz 128 kbps CBR/VBR and a couple at 96. I don't understand why those assholes can't downsample to 32 KHz if they're gonna use such low bitrates. Downsampling to 32 KHz can actually improve the results when you use such low bitrates. Duh, all of my MP3s in the past were 32 KHz 96-128 kbps. No artifacts at all. Unfortunetaly, I can't control how other retards encode their material, and the **** I downloaded was some old anime ripped from a Laserdisc. I really doubt there is a higher quality copy available on the net beside the one I snatched which already took forever to download. 32 KHz isn't that ugly of a sample rate - it allows some kind of frequency response up to about 16 KHz. Please remember that FM stereo pretty well tops out at 15 KHz. I doubt there is any significant difference at all, as most can't hear over 16 khz anyway. I'm 18 and can hear up to 17, which is probably why I sometimes notice a difference if I concentrate really hard. For all intents and purposes, even 22 KHz is allright -- you lose some cymbals but meh. Some general rules for coding to low bitrates are to forget stereo and go to mono, and decrease the bandwidth as much as you can without losing too much intelligibility. Spoken word in mono with a 5 to 8 Hz bandwidth isn't usually all that bad, and classical music with 11 KHz bandwidth can often be quite satisfying. **** on that ****! With the advent of spectral band replication and parametric stereo there's no need for downsampling or downmixing anymore. |
#16
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
On Oct 16, 10:39 am, Industrial One
wrote: I'm 18 And that explains everything. and can hear up to 17, have the technical skills of 15, social skills of 12, and most of the time act like a 2 year old. How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? Perhaps your best choice is to not use the computer without your Mom and Dad's permission. |
#17
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Industrial One" wrote in
message Arny Krueger wrote: "Industrial One" wrote in message How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You might be able do obtain a perceived improvement with filtering and noise gating. Such as? At really low bit rates there is often background noise and echos. Low pass filtering can mitigate some of the irritation due to the noise, and a noise gate can help with the some of the background noise and some of the echoes. |
#18
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
On Oct 16, 3:12 pm, wrote:
On Oct 16, 10:39 am, Industrial One wrote: I'm 18 And that explains everything. and can hear up to 17, have the technical skills of 15, social skills of 12, and most of the time act like a 2 year old. How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? Perhaps your best choice is to not use the computer without your Mom and Dad's permission. You got a problem? P.S. I own this computer and apartment. My mom probably OD'ed and my dad is in a nuthouse. On Oct 16, 3:21 pm, "Arny Krueger" wrote: "Industrial One" wrote in Arny Krueger wrote: "Industrial One" wrote in message How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You might be able do obtain a perceived improvement with filtering and noise gating. Such as? At really low bit rates there is often background noise and echos. Low pass filtering can mitigate some of the irritation due to the noise, and a noise gate can help with the some of the background noise and some of the echoes. Oh **** it, the audio stays. The problem is not the noise and removing any echo would probably remove legitimate reverb effects in the audio. I don't even know why I'm bitching. It doesn't sound bad, it's just not not up to par to the quality it could've had. Oh well, I doubt the mental lonely ****s on eBay would care after they buy my "remastered" copies. The picture is fine and that's all I care about. |
#19
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
On Oct 16, 11:45 am, Industrial One
wrote: Perhaps your best choice is to not use the computer without your Mom and Dad's permission. You got a problem? Nope, but it seems you're willing to share yours with the world. P.S. I own this computer and apartment. My mom probably OD'ed and my dad is in a nuthouse. No, they're probably hiding under a rock, regretting the day they didn't pay attention to the "birth control" chapter in sex ed. I don't even know why I'm bitching Because you're an unsocialized annoying little ass with poor impulse control whose best skill is attention seeking behavior. Once in a great while, you're a source of mild entertainment. |
#20
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
Industrial One writes:
Arny Krueger wrote: "Industrial One" wrote in message How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You might be able do obtain a perceived improvement with filtering and noise gating. Such as? The clips in question are 48 KHz 128 kbps CBR/VBR and a couple at 96. I don't understand why those assholes can't downsample to 32 KHz if they're gonna use such low bitrates. Downsampling to 32 KHz can actually improve the results when you use such low bitrates. Duh, all of my MP3s in the past were 32 KHz 96-128 kbps. No artifacts at all. Unfortunetaly, I can't control how other retards encode their material, and the **** I downloaded was some old anime ripped from a Laserdisc. I really doubt there is a higher quality copy available on the net beside the one I snatched which already took forever to download. 32 KHz isn't that ugly of a sample rate - it allows some kind of frequency response up to about 16 KHz. Please remember that FM stereo pretty well tops out at 15 KHz. I doubt there is any significant difference at all, as most can't hear over 16 khz anyway. I'm 18 and can hear up to 17, which is probably why I sometimes notice a difference if I concentrate really hard. For all intents and purposes, even 22 KHz is allright -- you lose some cymbals but meh. I agree. Some general rules for coding to low bitrates are to forget stereo and go to mono, and decrease the bandwidth as much as you can without losing too much intelligibility. Spoken word in mono with a 5 to 8 Hz bandwidth isn't usually all that bad, and classical music with 11 KHz bandwidth can often be quite satisfying. **** on that ****! With the advent of spectral band replication and parametric stereo there's no need for downsampling or downmixing anymore. So you've read http://en.wikipedia.org/wiki/Parametric_Stereo? These are some impressive new developments in audio ENCODING - won't help you too much with DECODING your files. -- % Randy Yates % "She's sweet on Wagner-I think she'd die for Beethoven. %% Fuquay-Varina, NC % She love the way Puccini lays down a tune, and %%% 919-577-9882 % Verdi's always creepin' from her room." %%%% % "Rockaria", *A New World Record*, ELO http://www.digitalsignallabs.com |
#21
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
On Oct 16, 4:39 pm, wrote:
On Oct 16, 11:45 am, Industrial One [SNIP] **** on that ****! With the advent of spectral band replication and parametric stereo there's no need for downsampling or downmixing anymore. So you've readhttp://en.wikipedia.org/wiki/Parametric_Stereo? Yeah, wrote part of it too. These are some impressive new developments in audio ENCODING - won't help you too much with DECODING your files. No ****... I was replying to Arny who said audio should be downmixed and downsampled if one aims to save space. So I explained why it ain't necessary. I really wish them dickbrains would ditch MP3 and start using MP4. I recently downloaded this 100-meg 9-hour trance collection track. I didn't even realize it was so long until I noticed the duration tab displayed an extra digit for this song. It's at 24kbps mono, 22 KHz. A ****LOAD of quality could've been preserved if the retard ripper used AAC-HEv2. |
#22
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
On Oct 16, 4:39 pm, wrote:
On Oct 16, 11:45 am, Industrial One wrote: Perhaps your best choice is to not use the computer without your Mom and Dad's permission. You got a problem? Nope, but it seems you're willing to share yours with the world. P.S. I own this computer and apartment. My mom probably OD'ed and my dad is in a nuthouse. No, they're probably hiding under a rock, regretting the day they didn't pay attention to the "birth control" chapter in sex ed. I don't even know why I'm bitching Because you're an unsocialized annoying little ass with poor impulse control whose best skill is attention seeking behavior. Once in a great while, you're a source of mild entertainment. Aww, poor Dickpierce. I got someone to cheer you up: www.goatse.cz FEEL THE STRETCH! |
#23
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Industrial One" wrote in message ... On Oct 15, 3:39 pm, "Richard Crowley" wrote: "Randy Yates" wrote ... Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. Hence the term "lossy compression". Does I1 have a list of troll questions that he posts regularly? "I1..." I kinda like it, despite how robotic it sounds. On Oct 15, 9:56 pm, Randy Yates wrote: Which is a ridiculous request. Sorta like "improving the quality" of the output of an 8-bit A/D. The noise is (or in your case, artifacts are) there to stay. Bull****, the noise can be removed via noise-removal techniques and re- saved as 16-bit. If that were true, we'd just save everything as 8-bits, and do the noise removal. Noise removal techniques are iffy at best, and obnoxious at worst, even when meticulously tuned and applied by hand. |
#24
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Randy Yates" wrote in message ... Industrial One writes: [...] So you've read http://en.wikipedia.org/wiki/Parametric_Stereo? These are some impressive new developments in audio ENCODING - won't help you too much with DECODING your files. Interesting article. I thought it was quite telling that the effect doesn't work particularly well at higher bitrates. Of course, other encoding schemes use sum and difference, taking advantage of the fact that the difference between the two channels channels is usually much smaller than the mono sum. That goes all the way back to stereo encoding on vinyl as well as FM and TV stereo, and later, FLAC, et. al. |
#25
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Chronic Philharmonic" wrote in
message "Randy Yates" wrote in message ... Industrial One writes: [...] So you've read http://en.wikipedia.org/wiki/Parametric_Stereo? These are some impressive new developments in audio ENCODING - won't help you too much with DECODING your (existing) files. Interesting article. I thought it was quite telling that the effect doesn't work particularly well at higher bitrates. I suspect that it works no better or worse at higher bitrates in an absolute sense, but it is not as acceptable because listener expectations are so much higher at higher bitrates. The mention of Satellite radio in one of the Wiki articles is telling, because the audio quality standards for the best known satellite radio network in the U.S. are abysmal. They might be good enough for Howard Stern or a NASCAR race, but they are not for what most people here would call quality audio. Of course, other encoding schemes use sum and difference, taking advantage of the fact that the difference between the two channels channels is usually much smaller than the mono sum. IOW, you don't need a high quality, full-bandpass difference channel to create the perception of space and directionality. That goes all the way back to stereo encoding on vinyl as well as FM and TV stereo, and later, For most of the life of FM stereo, real world FM stereo receivers characteristically lost lots of separation at high frequencies. FLAC, et. al. AFAIK FLAC is lossless, and makes no compromises at all. |
#26
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Arny Krueger" wrote in message ... "Chronic Philharmonic" wrote in message "Randy Yates" wrote in message ... Industrial One writes: [...] [...] FLAC, et. al. AFAIK FLAC is lossless, and makes no compromises at all. Right, but they store the sum and difference, rather than essentially duplicating the majority of both channels. It is lossless, but not wasteful. |
#27
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
In article ,
"Arny Krueger" wrote: -- snip -- For most of the life of FM stereo, real world FM stereo receivers characteristically lost lots of separation at high frequencies. Having worked on some of the earliest FM stereo encoders which actually *met* all the FCC specifications, I would say that a lot of the problem was with *encoders*, not decoders. The degree of matching (both amplitude and phase) required between the L and R low-pass filters necessary for good HF separation was not widely understood -- and even less often realized. With the advent of digital filtering techniques, things got a *lot* easier. Isaac |
#28
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
isw writes:
In article , "Arny Krueger" wrote: -- snip -- For most of the life of FM stereo, real world FM stereo receivers characteristically lost lots of separation at high frequencies. Having worked on some of the earliest FM stereo encoders which actually *met* all the FCC specifications, I would say that a lot of the problem was with *encoders*, not decoders. The degree of matching (both amplitude and phase) required between the L and R low-pass filters necessary for good HF separation was not widely understood -- and even less often realized. With the advent of digital filtering techniques, things got a *lot* easier. Having implemented a full BTSC decoder in the digital domain about a year and a half ago, I can say from personal experience that it's not all that easy. The time I spent on the various filters involved - trying to get them designed to the required accuracy - was very painful. Not to detract, however, from your correct point, isw, that implementing a good encoder/decoder (for FM or analog TV broadcast) is no mean feat. -- % Randy Yates % "Watching all the days go by... %% Fuquay-Varina, NC % Who are you and who am I?" %%% 919-577-9882 % 'Mission (A World Record)', %%%% % *A New World Record*, ELO http://www.digitalsignallabs.com |
#29
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
On Oct 17, 4:02 am, "Chronic Philharmonic"
wrote: "Industrial One" wrote in message ... On Oct 15, 3:39 pm, "Richard Crowley" wrote: "Randy Yates" wrote ... Industrial One writes: How can I improve the quality of a clip that suffers from smearing/ ringing artifacts due to low-medium bitrates? You cannot. You cannot replace information that has been lost. Hence the term "lossy compression". Does I1 have a list of troll questions that he posts regularly? "I1..." I kinda like it, despite how robotic it sounds. On Oct 15, 9:56 pm, Randy Yates wrote: Which is a ridiculous request. Sorta like "improving the quality" of the output of an 8-bit A/D. The noise is (or in your case, artifacts are) there to stay. Bull****, the noise can be removed via noise-removal techniques and re- saved as 16-bit. If that were true, we'd just save everything as 8-bits, and do the noise removal. Noise removal techniques are iffy at best, and obnoxious at worst, even when meticulously tuned and applied by hand. Because it's useless if I'm gonna compress to MP3 since it'll smear and **** up the noise, making it harder to detect and remove. But as long as the noise dB are significantly lower than the signal, it can be easily removed, especially by hand. |
#30
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Chronic Philharmonic" wrote in message ... "Randy Yates" wrote in message ... Industrial One writes: [...] So you've read http://en.wikipedia.org/wiki/Parametric_Stereo? These are some impressive new developments in audio ENCODING - won't help you too much with DECODING your files. Interesting article. I thought it was quite telling that the effect doesn't work particularly well at higher bitrates. Of course, other encoding schemes use sum and difference, taking advantage of the fact that the difference between the two channels channels is usually much smaller than the mono sum. That goes all the way back to stereo encoding on vinyl as well as FM and TV stereo, and later, FLAC, et. al. I know this is getting off-topic, but I thought it might be interesting to point out the there wasn't really any "encoding" of stereo as such on vinyl. The two channels independently moved the stylus, each at 45° (thus at 90° to each other). Today it's called "discrete" channels. The result was that if there was no LR difference, then the stylus moved only laterally, which means that a mono record would play properly on a stereo system. That's also why stereo records would not play properly on a mono cartridge, because it probably wasn't designed to allow much vertical movement, and would cause damage to the extent that there was LR difference. In the worst case of LR difference, such as where one channel was the same stuff as the other, but of inverse polarity, the stylus moved only vertically. -- Earl |
#31
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Earl Kiosterud" wrote in message news "Chronic Philharmonic" wrote in message ... "Randy Yates" wrote in message ... Industrial One writes: [...] So you've read http://en.wikipedia.org/wiki/Parametric_Stereo? These are some impressive new developments in audio ENCODING - won't help you too much with DECODING your files. Interesting article. I thought it was quite telling that the effect doesn't work particularly well at higher bitrates. Of course, other encoding schemes use sum and difference, taking advantage of the fact that the difference between the two channels channels is usually much smaller than the mono sum. That goes all the way back to stereo encoding on vinyl as well as FM and TV stereo, and later, FLAC, et. al. I know this is getting off-topic, but I thought it might be interesting to point out the there wasn't really any "encoding" of stereo as such on vinyl. The two channels independently moved the stylus, each at 45° (thus at 90° to each other). Today it's called "discrete" channels. The result was that if there was no LR difference, then the stylus moved only laterally, which means that a mono record would play properly on a stereo system. That's also why stereo records would not play properly on a mono cartridge, because it probably wasn't designed to allow much vertical movement, and would cause damage to the extent that there was LR difference. In the worst case of LR difference, such as where one channel was the same stuff as the other, but of inverse polarity, the stylus moved only vertically. I would respectfully argue that "encoding" is whatever you do to get audio onto the disc -- mono or stereo. The implementation with some (perhaps all) cutting heads and playback pickups might have been 45/45, but it is mathematically identical to L+R (lateral) and L-R (vertical). A 45/45 pickup can play back a record made with a L+R (lateral)/L-R (vertical) cutter, without modification. Statistically, L+R is more closely correlated than L-R, so there is less vertical activity on average. Not only that, but this encoding allows the L+R amplitude to be limited separately from L-R (L-R limiting would reduce channel separation momentarily). Not only that, but the signals could be equalized separately, so less bass is sent to the L-R channel. This reduces the risk of the cutter losing contact with the surface, avoiding excessive distortion and skips on playback. I suppose we could start a new topic if there is any further interest in this. |
#32
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Chronic Philharmonic" wrote in message ... "Earl Kiosterud" wrote in message news "Chronic Philharmonic" wrote in message ... "Randy Yates" wrote in message ... Industrial One writes: [...] So you've read http://en.wikipedia.org/wiki/Parametric_Stereo? These are some impressive new developments in audio ENCODING - won't help you too much with DECODING your files. Interesting article. I thought it was quite telling that the effect doesn't work particularly well at higher bitrates. Of course, other encoding schemes use sum and difference, taking advantage of the fact that the difference between the two channels channels is usually much smaller than the mono sum. That goes all the way back to stereo encoding on vinyl as well as FM and TV stereo, and later, FLAC, et. al. I know this is getting off-topic, but I thought it might be interesting to point out the there wasn't really any "encoding" of stereo as such on vinyl. The two channels independently moved the stylus, each at 45° (thus at 90° to each other). Today it's called "discrete" channels. The result was that if there was no LR difference, then the stylus moved only laterally, which means that a mono record would play properly on a stereo system. That's also why stereo records would not play properly on a mono cartridge, because it probably wasn't designed to allow much vertical movement, and would cause damage to the extent that there was LR difference. In the worst case of LR difference, such as where one channel was the same stuff as the other, but of inverse polarity, the stylus moved only vertically. I would respectfully argue that "encoding" is whatever you do to get audio onto the disc -- mono or stereo. The implementation with some (perhaps all) cutting heads and playback pickups might have been 45/45, but it is mathematically identical to L+R (lateral) and L-R (vertical). A 45/45 pickup can play back a record made with a L+R (lateral)/L-R (vertical) cutter, without modification. Statistically, L+R is more closely correlated than L-R, so there is less vertical activity on average. Not only that, but this encoding allows the L+R amplitude to be limited separately from L-R (L-R limiting would reduce channel separation momentarily). Not only that, but the signals could be equalized separately, so less bass is sent to the L-R channel. This reduces the risk of the cutter losing contact with the surface, avoiding excessive distortion and skips on playback. I suppose we could start a new topic if there is any further interest in this. But as far as I am aware, there never has been any separate processing of the L+R and L-R, only L and R separately, but using the same EQ and compressor settings. On rock recordings, the L-R was necessarily minimised by mixing kick drums and sometimes bass to centre, with the vocalist almost always dead centre. Classical and Jazz tended to have more L-R, but as the music wasn't so heavily compressed, the levels were lower anyway. It is essential that any L,R processing be done with identical settings as otherwise the central image will wander depending on frequency content and level. There were some mono/stereo compatible records (Synchro Stereo was one I recall) which I understand mixed low frequencies to mono and thus kept the L-R signal small whilst still offering a noticeable stereo effect. I have several such records of classical music, and they sound adequate in stereo, but have sufficiently small L-R levels that they can be played with a mono pickup without damage. S. -- http://audiopages.googlepages.com |
#33
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Serge Auckland" wrote in
message "Chronic Philharmonic" wrote in message ... "Earl Kiosterud" wrote in message news "Chronic Philharmonic" wrote in message ... "Randy Yates" wrote in message ... I know this is getting off-topic, but I thought it might be interesting to point out the there wasn't really any "encoding" of stereo as such on vinyl. The two channels independently moved the stylus, each at 45° (thus at 90° to each other). Today it's called "discrete" channels. The result was that if there was no LR difference, then the stylus moved only laterally, which means that a mono record would play properly on a stereo system. That's also why stereo records would not play properly on a mono cartridge, because it probably wasn't designed to allow much vertical movement, and would cause damage to the extent that there was LR difference. In the worst case of LR difference, such as where one channel was the same stuff as the other, but of inverse polarity, the stylus moved only vertically. I would respectfully argue that "encoding" is whatever you do to get audio onto the disc -- mono or stereo. Agreed - the LP was an example of encoding an electrical signal into a mechanical signal. It was descended from an earlier process that encoded an acoustical signal as a mechanical signal. The implementation with some (perhaps all) cutting heads and playback pickups might have been 45/45, but it is mathematically identical to L+R (lateral) and L-R (vertical). A 45/45 pickup can play back a record made with a L+R (lateral)/L-R (vertical) cutter, without modification. At some point in the processing of audio recorded on LPs, the signal was turned every which way but lose, and sum/difference processing was very common because of its impact on trackability on very modest playback equipment. Statistically, L+R is more closely correlated than L-R, so there is less vertical activity on average. Not only that, but vertical (L-R) dynamic range is far more limited than horizontal (L+R) dynamic range. You run out of vertical dynamic range when the cutting stylus digs a hole or becomes airborne. Both can happen and did happen in the real world. You run out of horizontal dynamic range when the cutting stylus loops through an adjacent groove or creates a radius that can't be tracked by the probable playback stylus. The adjacent groove problem can be managed by increasing the pitch (space between adjacent tracks) of the grooves. Increasing pitch cuts the amount of time that you can record. The problem of creating radii that can't be tracked can be managed by using smaller radii, which was really what elliptical styli were all about. It's also possible within limits to modify the trajectory of the stylus so that the intended stylus has the desired mechanical trajectory despite obvious geometric limits. The real problem with mainstream vinyl was that it had to be cut for the lowest common denominator playback system or else the recording will sound extraordinarily crappy to way too many people, and have a short life. Not only that, but this encoding allows the L+R amplitude to be limited separately from L-R (L-R limiting would reduce channel separation momentarily). Not only that, but the signals could be equalized separately, so less bass is sent to the L-R channel. This reduces the risk of the cutter losing contact with the surface, avoiding excessive distortion and skips on playback. This was all done routinely, particularly in the latter days of vinyl, just before the CD came out. But as far as I am aware, there never has been any separate processing of the L+R and L-R, only L and R separately, Then with all due respect, you weren't aware of the LP SOTA in the latter days. but using the same EQ and compressor settings. Ditto. On rock recordings, the L-R was necessarily minimised by mixing kick drums and sometimes bass to centre, with the vocalist almost always dead centre. Well that too. The advance of doing this is that the best people make better artistic choices than electronics, particularly the limited electronics of the late 1970s and early 80s. Classical and Jazz tended to have more L-R, but as the music wasn't so heavily compressed, the levels were lower anyway. Except that it isn't allowable to dig a hole or send the stylus into the air or loop an adjacent track ever, even during crescendos. Just for fun they adopted the convention of recording LPs from the outside edge in, so the crescendos always ended up in the inner grooves where available dynamic range was minimized. It is essential that any L,R processing be done with identical settings as otherwise the central image will wander depending on frequency content and level. As that doesn't happen anyway. There were some mono/stereo compatible records (Synchro Stereo was one I recall) which I understand mixed low frequencies to mono and thus kept the L-R signal small whilst still offering a noticeable stereo effect. In fact that was happening to a certain degree very often on mainstream releases with no special labeling. I have several such records of classical music, and they sound adequate in stereo, but have sufficiently small L-R levels that they can be played with a mono pickup without damage. Eventually a lot of pop came to be that way for any number of reasons, and persists to this day even though most digital media has as much power bandwidth as anything. |
#34
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Arny Krueger" wrote in message . .. "Serge Auckland" wrote in message "Chronic Philharmonic" wrote in message ... "Earl Kiosterud" wrote in message news "Chronic Philharmonic" wrote in message ... "Randy Yates" wrote in message ... snipped Not only that, but this encoding allows the L+R amplitude to be limited separately from L-R (L-R limiting would reduce channel separation momentarily). Not only that, but the signals could be equalized separately, so less bass is sent to the L-R channel. This reduces the risk of the cutter losing contact with the surface, avoiding excessive distortion and skips on playback. This was all done routinely, particularly in the latter days of vinyl, just before the CD came out. But as far as I am aware, there never has been any separate processing of the L+R and L-R, only L and R separately, Then with all due respect, you weren't aware of the LP SOTA in the latter days. Possibly not, so thanks for the update. but using the same EQ and compressor settings. Ditto. I don't understand this part: If the L&R have different compressor and EQ settings, then the image will wander depending on level and frequencies. In Broadcast at least, it's normal that the L&R settings are linked in a stereo compressor/limiter and/or equaliser to avoid any image drift. Is this not also done on LP mastering? If not, how is image drift avoided? On rock recordings, the L-R was necessarily minimised by mixing kick drums and sometimes bass to centre, with the vocalist almost always dead centre. Well that too. The advance of doing this is that the best people make better artistic choices than electronics, particularly the limited electronics of the late 1970s and early 80s. Classical and Jazz tended to have more L-R, but as the music wasn't so heavily compressed, the levels were lower anyway. Except that it isn't allowable to dig a hole or send the stylus into the air or loop an adjacent track ever, even during crescendos. Just for fun they adopted the convention of recording LPs from the outside edge in, so the crescendos always ended up in the inner grooves where available dynamic range was minimized. Agreed. It always seemed odd to me that records played outside-in, when it would be more logical to play inside-out. It is essential that any L,R processing be done with identical settings as otherwise the central image will wander depending on frequency content and level. As that doesn't happen anyway. Why not? If a stereo signal has L&R processed independantly, then the image will drift with level and frequency. That's why most stereo compressor/limiters and EQs have a "link" button that provides the same control signal to both channels. S. -- http://audiopages.googlepages.com |
#35
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Serge Auckland" wrote in
message "Arny Krueger" wrote in message . .. "Serge Auckland" wrote in message "Chronic Philharmonic" wrote in message ... "Earl Kiosterud" wrote in message news "Chronic Philharmonic" wrote in message ... "Randy Yates" wrote in message ... snipped Not only that, but this encoding allows the L+R amplitude to be limited separately from L-R (L-R limiting would reduce channel separation momentarily). Not only that, but the signals could be equalized separately, so less bass is sent to the L-R channel. This reduces the risk of the cutter losing contact with the surface, avoiding excessive distortion and skips on playback. This was all done routinely, particularly in the latter days of vinyl, just before the CD came out. But as far as I am aware, there never has been any separate processing of the L+R and L-R, only L and R separately, Then with all due respect, you weren't aware of the LP SOTA in the latter days. Possibly not, so thanks for the update. but using the same EQ and compressor settings. Ditto. I don't understand this part: If the L&R have different compressor and EQ settings, then the image will wander depending on level and frequencies. Even if you the compressors are identical, there will still be wandering channels. For example, I compress both channels 2:1 above -10 dB. One channel is 10 dB below the other, and they both steadily increase their volume. The channel that hits 10 dB first starts increasing more slowly and thus starts sliding towards the center. That's why they link control signals. In Broadcast at least, it's normal that the L&R settings are linked in a stereo compressor/limiter and/or equaliser to avoid any image drift. Yes, using the same control signal on both compressors helps. Is this not also done on LP mastering? Yes, if L & R are compressed then the control signals are tied together and any effects on imaging are second order. If L-R is compressed, then both channels slide to the center. If L+R is compressed, then diffuse sound becomes more diffuse. Synchronizing their compresson without affecting imaging would be a neat trick. Maybe that is one reason why highly-compressed recordings tend to sound like mud. |
#36
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Serge Auckland" wrote ...
Agreed. It always seemed odd to me that records played outside-in, when it would be more logical to play inside-out. OTOH, note that for professional use (i.e. transcription, etc.) the practice was often to record from the center out. Note further that optical discs (CD, DVD play from the center to the outside edge. |
#37
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Serge Auckland" wrote in message ... Agreed. It always seemed odd to me that records played outside-in, when it would be more logical to play inside-out. But the consequences of the stylus jumping the run out groove and falling off the record onto the platter would be devastating. More likely, but less expensive with the cheap players used by many of course, but still not something the public would be happy with, especially since the limitations (and therefore possible benefits) were unknown to the masses in any case. MrT. |
#38
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Richard Crowley" wrote in message ... "Serge Auckland" wrote ... Agreed. It always seemed odd to me that records played outside-in, when it would be more logical to play inside-out. OTOH, note that for professional use (i.e. transcription, etc.) the practice was often to record from the center out. Note further that optical discs (CD, DVD play from the center to the outside edge. Which of course provides no audible benefit though. The reason is simply that any size disk can be used without special size detection, since the TOC always starts in the same place. MrT. |
#39
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
Agreed. It always seemed odd to me that records played outside-in, when it
would be more logical to play inside-out. But the consequences of the stylus jumping the run out groove and falling off the record onto the platter would be devastating. My recollection is that records suffer from a higher level of high-frequency tracing distortion when playing the inner grooves, due in part to the fact that records are cut using constant angular velocity, The wavelength of the signals (in the vinyl) becomes smaller in the inner grooves, making it more difficult for the diamond stylus to track the groove accurately. The fact that the stylus isn't exactly tangent to the groove (at most points), makes life even more complicated. Mastering an LP involves a set of tradeoffs involving recording time, level, and distortion. As the side length becomes greater, you need correspondingly more spirals in the groove. You can go further in towards the center, and suffer increasing levels of distortion in the inner grooves. Or, you can decrease the pitch (the distance between the grooves) so that you don't go so far in. If you do this, you end up having to reduce the audio level (turn down the volume) - otherwise, crosstalk between adjacent grooves becomes more obvious (pre- and post-echo) and in severe cases you end up accidentally cutting from one groove to the next and ruining the master. Reducing the cutting amplitude will tend to reduce playback tracing distortion, but it can result in the record's surface noise being more obvious. This all gets *really* complicated if the LP is being recorded "direct to disk" rather than via a master tape - the mastering engineer has to set the lathe's pitch adjustment "on the fly" based on his/her knowledge of what the musicians are going to be playing in the next few seconds. -- Dave Platt AE6EO Friends of Jade Warrior home page: http://www.radagast.org/jade-warrior I do _not_ wish to receive unsolicited commercial email, and I will boycott any company which has the gall to send me such ads! |
#40
Posted to rec.audio.tech
|
|||
|
|||
Restoring quality from low-bitrate MP3
"Dave Platt" wrote ...
This all gets *really* complicated if the LP is being recorded "direct to disk" rather than via a master tape - the mastering engineer has to set the lathe's pitch adjustment "on the fly" based on his/her knowledge of what the musicians are going to be playing in the next few seconds. That's why we have rehearsals and run-throughs. |
Reply |
|
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
AUDITION MP3 BITRATE | Pro Audio | |||
restoring cd quality audio to FM recordings | General | |||
MP3 bitrate for CD quality: my observations | Tech | |||
mpg bitrate for voice? | Tech |