It is obvious that YT starting to use strange and false compression
scheme (and may be EQ as well) sometime in June 2008. Because
everything I uploaded up till that point sounded fine compared to
original sound I uploaded but not any more. The main problem is too
much compression and peak level is set to exceed max value of 0db in
their current settings probably by an error by inexperienced engineer
in audio field, in turn it results clipping sound in loud passages and
produces unacceptable breathing sound/noise in quiet passages which I
checked the wave forms on encoded audio track by YT. In other words,
loudest passage of the given track is encoded above 0db by some reason
and all those parts would sound distorted being clipped. Peak of
clipped wave form resembles to Trapezoid since any signal exceeded 0db
would be pushed back down to 0db instead of smooth mountain top shape/
curve how natural dynamics should be reproduced. To work around with
this problem after several experiments, 1) The peak of uploading audio
track should be kept under -3db (or normalize entire track to -3db or
even less) and this seems to help avoiding loud passages being clipped
by YT compression. 2) However this would not help quiet passages to
remain in original level, they will be boosted up nearly 12db or even
more along with hiss and noise thus resulting typical audible
breathing effect. If you have cooledit pro or similar professional
audio tools, you can mix inaudible signal (20,000hz since wave) over
those quiet passages to boost overall level to around -4 to -6db. For
example, if your quiet passage is -16db, you can overlap mixing -6db
20,000hz sine wave will results around -4db signal power, thus YT
compression will not boost this passage’s signal that much and it
remains in original volume or reverb decay etc. It’s a bit tricky, but
it works both in theory and practice. Or, to make it simpler,
normalize original track to -6db and mix -8db 20,000hz since wave over
entire track and this will result around max peak around -3db and
quiet passages will be raised dramatically without making any audible
difference, but it will not be boosted by YT compressor during
encoding, though I have not tested this method yet. 3) As for EQ, it
requires a bit more extensive test to see how and how much EQ being
applied by YT encoder, however I have some idea how to test this when
I find time to do so, but it would be much faster if YT audio
department fixes all those glitches on their end soon, basically just
use whatever they were doing in the past few years and that was
absolutely fine considering bitrate limit.
No, current YT audio encoding is limited to 128kbps with 44.1khz max
16,000 hz frequency, thus 20,000hz frequency would not be encoded by
YT audio scheme, but u can still upload audio track mixed with
20,000hz using 48khz audio track :) Try it, it works :) Or if not,
pls, let me know.
I've just been experimenting a bit, and getting encouraging results.
My software won't allow me to normalize sound to a specific volume
(you'd think it would, but no), so I have to do that by hand, but
doable. It definitely helps, though.
Procedure: Mix down audio, adjust volume so that it peaks at around
-6dB. Then add audio track consisting of 20kHz sine wave at -8dB, and
export and upload.
> No, current YT audio encoding is limited to 128kbps with 44.1khz max
> 16,000 hz frequency, thus 20,000hz frequency would not be encoded by
> YT audio scheme, but u can still upload audio track mixed with
> 20,000hz using 48khz audio track :) Try it, it works :) Or if not,
> pls, let me know.
> On Jul 25, 11:55 pm, rewboss wrote:
> > Just a thought though: Won't this drive dogs crazy?
Iff your software doesn't allow normalizing, reduce entire level down
to 75%, that's -3db, down to 50% would be the same as normalizing
level to - 6db, down to 25% would be the same as normalizing at -12db.
> I've just been experimenting a bit, and getting encouraging results.
> My software won't allow me to normalize sound to a specific volume
> (you'd think it would, but no), so I have to do that by hand, but
> doable. It definitely helps, though.
> Procedure: Mix down audio, adjust volume so that it peaks at around
> -6dB. Then add audio track consisting of 20kHz sine wave at -8dB, and
> export and upload.
> If you could give the technical details of the incorrect compression/
> EQ problem, that might help YouTube fix the problem.
> On Jul 25, 11:08 pm, Sopranoguitar wrote:
> > No, current YT audio encoding is limited to 128kbps with 44.1khz max
> > 16,000 hz frequency, thus 20,000hz frequency would not be encoded by
> > YT audio scheme, but u can still upload audio track mixed with
> > 20,000hz using 48khz audio track :) Try it, it works :) Or if not,
> > pls, let me know.
> > On Jul 25, 11:55 pm, rewboss wrote:
> > > Just a thought though: Won't this drive dogs crazy?
Thank you, Sopranoguitar. I've been going crazy with this same audio
problem too. I've had to reduce my audio levels by about 70% to get an
acceptable upload - that's a huge difference! I hope YT becomes aware
of this issue. I've been wondering if it's only a problem in certain
areas (I'm in New York City). Thanks again, Sopranoguitar.
> Iff your software doesn't allow normalizing, reduce entire level down
> to 75%, that's -3db, down to 50% would be the same as normalizing
> level to - 6db, down to 25% would be the same as normalizing at -12db.
It works with standard quality, but when you switch to high quality it
messes up the volume yet again -_-
The effect of hard limiting is less noticeable on pop, dance or hip-
hop music since they’re mixed on relatively consistent strong rhythm
track and dynamic range is so small compare to acoustic jazz, classic
or solo performances or spoken word which requires much larger dynamic
range to express their music or statement properly and naturally, so
those are the ones suffers the most with this hard limiting scheme. I
don’t know why they even bother to apply this new false audio scheme
recently, may be for some political reasons or pressures to present
presidential campaigns to be heard more loud and clear?
As for HQ video, what I gathered is that they have at least three
different encoding processes once you upload video larger than 320x240
such as 480x360 or 640x480 etc with different bitrate allowances both
for audio and video for three versions and they encode this clip to:
1) Default 320x240 with 64kbps mono audio 2) 128kbps stereo audio with
&fmt=18 tag to view it. 3) HQ video with 128kbps mono audio with
&fmt=6 tag to view it. However, I have no clue what they’re exactly
doing or playing with these new formats, all I can tell you is that
obviously they don’t have professional audio engineer in their team or
company, probably they’re all kids listening to hip-hop or dance music
day and night so that they don’t realize how hard limiting is harmful
to real artistic natural audio expressions. Only thing I know is that
I can upload stereo audio clip using larger than 320x240 video clip if
I view it using &fmt=18 which worked so far up to around 512kbps total
bitrate, but audio quality was not as good as original clip though :(
Well, I must tell you that YT is giving a lot of works on users ends
especially in audio department recently! Yes, it indeed sucks for
their professionalism ignoring that people are more sensitive to audio
than visuals to enjoy life! Simply because, we can close our eyes or
look at different direction at anytime while we can’t close our ears
so easily. so bad sounds simply makes us unhappy for sure, aren’t
they?
> > Iff your software doesn't allow normalizing, reduce entire level down
> > to 75%, that's -3db, down to 50% would be the same as normalizing
> > level to - 6db, down to 25% would be the same as normalizing at -12db.
> It works with standard quality, but when you switch to high quality it
> messes up the volume yet again -_-
Now that this thread is front page on Slashdot (http://
tech.slashdot.org/tech/08/07/27/1810212.shtml) maybe it'll get noticed
and we'll get some relief.
I had the same problem and, in addition, colors are altered. I define
a shape using only 1 tone of a color and after upload I get 5 o 6
actually different tones. I want to get exactly what I put on my
video, do not try to improve nothing, cause you mess it up, please no
modifications!
I've noticed this issue as well and it affects my videos in a negative
way. I use background music a lot, and the background music gets
turned up a lot when I'm not talking. It's distracting and sounds very
odd.
I think even 10:1 might be an underestimate. This is some of the worst
compression abuse I've ever heard. It's like YT just discovered
compressors and cranked the knobs to max. I'm trying to post classical
recordings and the crunched audio makes them unlistenable.
My question is... once YT fixes this (and please, YT, please fix
this), will we have to re-upload the affected videos? I assume so.
> I think even 10:1 might be an underestimate. This is some of the worst
> compression abuse I've ever heard. It's like YT just discovered
> compressors and cranked the knobs to max. I'm trying to post classical
> recordings and the crunched audio makes them unlistenable.
> My question is... once YT fixes this (and please, YT, please fix
> this), will we have to re-upload the affected videos? I assume so.
This problem is eerily similar to macrovision protection on DVDs...
maybe youtube is purposely screwing with certain videos to detract
people from uploading music videos (especially copyrighted ones) ...
guess we will have to stick to watching videos of people getting hit
in the balls.. I doubt this will go away anytime soon
This problem occurs in every video format I have tried, Mpeg 1 Mpeg 2,
Mpeg 4, Mpeg 4(H.264) and Pre-encoded in FLV formats in any screen
size, bitrate or audio sampling rate, or length of clip. Perhaps, it
is not the problem in encoding process its self, it is a problem with
limiter or some sort of pre-amp settings chained in front or back of
encoder (Pre-encoding or Post-encoding process, most likely the former
though), since all standard FLV or mp4 encoders do not produce this
type of additional audio problem with dynamics or amplitudes. It is
also important for YT to make announcement when this problem being
acknowledged and resolved since it is a waste of time for users
uploading new clips until this issue being resolved, especially for
the ones care about decent audio quality for their clips. I don't know
if it is a global problem through out all servers/encoding process in
YT, or certain numbers of servers/encoding process in YT, since users
cannot trace where their files being encoded and how many encoders or
servers being employed by YT.
> Been having this happen since June 27, 2008 too. Uploading WMV format
> -- have tried all sorts of audio mixes, none work. Audio goes high,
> then low.
> Please admit the problem and fix Youtbe!
> I can't upload anything to Youtube until this is fixed. Vimeo is
> handling the same videos perfectly.
> On Jul 28, 7:46 am, kbub6f wrote:
> > I think even 10:1 might be an underestimate. This is some of the worst
> > compression abuse I've ever heard. It's like YT just discovered
> > compressors and cranked the knobs to max. I'm trying to post classical
> > recordings and the crunched audio makes them unlistenable.
> > My question is... once YT fixes this (and please, YT, please fix
> > this), will we have to re-upload the affected videos? I assume so.
I had thought some recently uploaded clips sounded lousy, but I didn't
really think much about it until I stubled into this thread.
If YouTube is indeed now altering the audio this is a serious problem
-- at least for anybody who knows what they are doing, and cares about
audio quality.
Some here seem to recommend altering the audio within a clip before
uploading it -- but that strikes me as a not-so-great idea -- because
YouTube keeps a copy of the originally uploaded clip, and might use it
to recompress new videos later. Meaning, if you alter your audio to
compensate for a bug now, then later they might fix the bug, and your
video might then start to sound wrong.
Has YouTube anywhere indicated that they know about this?
As for us, we CANNOT upload our clips if they are going to be altered
in this way -- I'd actually ecen have to find another way to serve our
videos if this proves to be the case, but I'd really much prefer to
use YouTube!
I'm sure it's intentional and part of the conversion chain on all
servers. You can sort of understand YT wanting to normalize audio to
make levels more consistent from video-to-video. Radio stations do the
exact same thing. But YT has done a really bad job of it. The
distortion and pumping/breathing are off the charts. Even a kid in the
music store getting his hands on a compressor for the first time
doesn't abuse it this badly.
Sopranoguitar and Scott, you're absolutely right: there's no point in
uploading new videos until YT fixes this. And it definitely makes
sense to start shopping for a new video host....
> I had thought some recently uploaded clips sounded lousy, but I didn't
> really think much about it until I stubled into this thread.
> If YouTube is indeed now altering the audio this is a serious problem
> -- at least for anybody who knows what they are doing, and cares about
> audio quality.
Hey all, just wanted to drop in and add myself to this unfortunate
thread. I uploaded my debut music video (OK, I'm just an amateur), and
I was dumbfounded when I gave it a listen. I felt like pulling it
down, except for the fact that my original version sounds reasonable.
Thanks for this thread, and even though I don't make a living off
music, I'm crossing my fingers that YouTube acknowledges and corrects
all this.
This post may be a bit wordy, so pls skip it if you don’t care to
read, but I had to address it for the ones may care.
I can clearly see now that youtube may not have proper support teams
and representatives or men powers to deal with different issues
concerning current problems or malfunctioning issues (bugs, so to
speak) within their site recently by attending only 3days to this
forum, ones may be more technical oriented and others may involve more
social, moral, political or civil infringements, or threatening/
offensive issues, and the bottom line for this is that they cannot be
handled by the same representatives cannot speak all these different
criteria of languages.
For example, how much “decibels”, “compressions”, “limiters” mean to
dear Anna or Liz who have been responding in this forum, or do they
really know what copyright infringement stands and fought for in court
and what is the difference how statue of limitations would be applied
in NY and CA courts in what circuit concerning copyright infringement?
They’re very different indeed! Or what kind of programming codes
(mainly database management schemes and codes) can cause those
technical problems, servers, encoders (including all audio plug-ins
currently used), current play-list problems, glitches and all that
stuff we’re talking about in this forum? Some knows better and some
knows less, but it makes no difference if YT ignores them.
There’re completely at least two different natures of problems here,
one is plain technical glitches which need to be simply corrected,
fixed or resolved from plain engineering/programming professionalism
no matter what! To maintain it’s functionality and quality just to
keep YT ship fly or run properly without crashing or sinking while
other topics are more debatable subjects based on law, policies,
opinions, preferences, beliefs or whatever, which can wait to be
resolved for a matter of week or so, but for technical stuff, we can’t
wait, simply because we’d crash or sink if we don’t work altogether as
part of YT community as users and service provider in harmony,
basically it would become useless tool for all users, if all those
technical glitches cannot be fixed in very short time :( There will
be other service providers soon or later works better than YT today
and every YT user would use them, it is just a matter of transferring
digital files while we sleep to new provider(s), no big deal! May be
some gets lost and some survives during this transition, but who
cares? That’s how life goes, just to see how this new providers work
out of curiosity. If it doesn’t work, we try another one whatever pops
up in time to time, and that’s how any industry works; changes and
evolves in time always, old ones without proper maintenance just die,
and new ones take it over to fill that gap.
What I see as main problem in YT is that they don’t have experienced/
proper/professional sound engineer in their company and it is very
obvious fact, and they can not get away with this short coming in
their operation in the long run as broadcasting system as they claim
it to be as their slogan and propaganda. Broadcast means audio
transmission; other media such as video are just decoration. Can you
imagine BBC with terrible/poor audio to survive as BBC? They care and
maintain it for the entire century so far, with decent audio engineers
in their team to take care of it. I’ve enjoyed so many historical
recordings by BBC, and now YT has a chance to preserve such historical
recordings by unknown artists if they care how to process and preserve
their recordings properly. Wake up! And face this problem today!
Audio transmission means broadcasting! So please improve audio quality
today, if you can.
> > I had thought some recently uploaded clips sounded lousy, but I didn't
> > really think much about it until I stubled into this thread.
> > If YouTube is indeed now altering the audio this is a serious problem
> > -- at least for anybody who knows what they are doing, and cares about
> > audio quality.
> Hey all, just wanted to drop in and add myself to this unfortunate
> thread. I uploaded my debut music video (OK, I'm just an amateur), and
> I was dumbfounded when I gave it a listen. I felt like pulling it
> down, except for the fact that my original version sounds reasonable.
> Thanks for this thread, and even though I don't make a living off
> music, I'm crossing my fingers that YouTube acknowledges and corrects
> all this.
> This post may be a bit wordy, so pls skip it if you don’t care to
> read, but I had to address it for the ones may care.
> I can clearly see now that youtube may not have proper support teams
> and representatives or men powers to deal with different issues
> concerning current problems or malfunctioning issues (bugs, so to
> speak) within their site recently by attending only 3days to this
> forum, ones may be more technical oriented and others may involve more
> social, moral, political or civil infringements, or threatening/
> offensive issues, and the bottom line for this is that they cannot be
> handled by the same representatives cannot speak all these different
> criteria of languages.
> For example, how much “decibels”, “compressions”, “limiters” mean to
> dear Anna or Liz who have been responding in this forum, or do they
> really know what copyright infringement stands and fought for in court
> and what is the difference how statue of limitations would be applied
> in NY and CA courts in what circuit concerning copyright infringement?
> They’re very different indeed! Or what kind of programming codes
> (mainly database management schemes and codes) can cause those
> technical problems, servers, encoders (including all audio plug-ins
> currently used), current play-list problems, glitches and all that
> stuff we’re talking about in this forum? Some knows better and some
> knows less, but it makes no difference if YT ignores them.
> There’re completely at least two different natures of problems here,
> one is plain technical glitches which need to be simply corrected,
> fixed or resolved from plain engineering/programming professionalism
> no matter what! To maintain it’s functionality and quality just to
> keep YT ship fly or run properly without crashing or sinking while
> other topics are more debatable subjects based on law, policies,
> opinions, preferences, beliefs or whatever, which can wait to be
> resolved for a matter of week or so, but for technical stuff, we can’t
> wait, simply because we’d crash or sink if we don’t work altogether as
> part of YT community as users and service provider in harmony,
> basically it would become useless tool for all users, if all those
> technical glitches cannot be fixed in very short time :( There will
> be other service providers soon or later works better than YT today
> and every YT user would use them, it is just a matter of transferring
> digital files while we sleep to new provider(s), no big deal! May be
> some gets lost and some survives during this transition, but who
> cares? That’s how life goes, just to see how this new providers work
> out of curiosity. If it doesn’t work, we try another one whatever pops
> up in time to time, and that’s how any industry works; changes and
> evolves in time always, old ones without proper maintenance just die,
> and new ones take it over to fill that gap.
> What I see as main problem in YT is that they don’t have experienced/
> proper/professional sound engineer in their company and it is very
> obvious fact, and they can not get away with this short coming in
> their operation in the long run as broadcasting system as they claim
> it to be as their slogan and propaganda. Broadcast means audio
> transmission; other media such as video are just decoration. Can you
> imagine BBC with terrible/poor audio to survive as BBC? They care and
> maintain it for the entire century so far, with decent audio engineers
> in their team to take care of it. I’ve enjoyed so many historical
> recordings by BBC, and now YT has a chance to preserve such historical
> recordings by unknown artists if they care how to process and preserve
> their recordings properly. Wake up! And face this problem today!
> Audio transmission means broadcasting! So please improve audio quality
> today, if you can.
> On Jul 29, 10:08 pm, rickumali wrote:
> > On Jul 29, 6:34 am, Scott (ehx) wrote:
> > > I had thought some recently uploaded clips sounded lousy, but I didn't
> > > really think much about it until I stubled into this thread.
> > > If YouTube is indeed now altering the audio this is a serious problem
> > > -- at least for anybody who knows what they are doing, and cares about
> > > audio quality.
> > Hey all, just wanted to drop in and add myself to this unfortunate
> > thread. I uploaded my debut music video (OK, I'm just an amateur), and
> > I was dumbfounded when I gave it a listen. I felt like pulling it
> > down, except for the fact that my original version sounds reasonable.
> > Thanks for this thread, and even though I don't make a living off
> > music, I'm crossing my fingers that YouTube acknowledges and corrects
> > all this.
Hello folks! I just want to put the end to this recent audio
compression problem by YT, so I have run and created some simple tests
something easy for anyone to see and hear how screwed up it is, it is
simply out of line and totally ridiculous how bad it is, once you can
see it! Simply it is unacceptable what YT is doing for their audio
recently. Here I put my two cents and a day of work to help, if
ignored, I have no regret. However, if you care at all for the sound
quality of your video clip posted on YT, just check out below what I
found out through my tests today.
1) Test Tones: For this purpose, I have created special test tones to
reveal recent YT compression behaviors. Very simple, I’ve chained
50-100-250-500-1000-5000-8000 and 10,000 Hz sine waves at -3db for
5sec blast each with gap of 1sec complete silence in between them and
each blast are faded out using linear fade out within 5sec from -3db
to -240db so that I can easily observe the differences between
original waveforms and waveforms after encoded by YT from attack to
decay, to speed and amount of compression, how they’re applied and
behaves as a result.
2) Well, the results are astounding to the point that I’ve never-ever
seen such an abuse of compression so badly by any professionals in
audio management in any media from the inception of radio, TV or
record industries, or even by telephone industries, and I’m 61 years
old professional sound engineer and recording artist as my background
lived through all of audio technology changes and developments since
1950’s. Worst audio in history by most popular internet media site in
2008, ironical indeed! How did it happen? I have no clue. They simply
applied Jimi Hendrix guitar distortion technique to all of their audio
tracks on their newly uploaded video clips, that was all to it. I
can’t believe that it is intentional, I believe that someone in their
company screwed it up really bad and there’s no quality control
department or supervisor oversees audio portion of contents being
uploaded since majority of audio are bad to begin with being captured
by mobile phone, camcorders etc. Probably no one in their company
watches high quality audio track video being uploaded to hear the
differences, just too many clips there to see those differences clip
by clip, therefore only users know how bad it is who uploaded quality
audio along with their clips, and that’s the problem I see here. But
it was fine until middle of June 2008, so that’s my point here to seek
improvement promptly to eliminate that glitch in their system, or to
restore something was working fine until 40days ago or so. It
shouldn’t be a hard work on their part once the source of problem can
be traced back correctly within this timeline. So, sooner is better to
resolve this!
3) Now, please click this link: http://casadelagos.com/YT_TestTone/ This is a result of my work and proof today to establish firm ground
that how current YT compression scheme is harmful to anyone cares
about sound quality of their clip, since they’re completely altering
this original sound how it was meant to be by users. It is not even
degrading it, it is altering it that you can easily witness by
comparing Fig1 & Fig2, how different they are, and I also posted
standard Mpeg to FLV conversion done by Riva converter at Fig3 locally
on my PC, the result is a bit degraded, but basically maintaining the
same shape except a bit low at 10,000hz and overall amplitude is a bit
less, but we can live with that kind of difference, since original
sound was not altered, just degraded a bit for overall level and
frequency responses, but that’s normal trade off converting it to
lower resolution for smaller file, however it is basically the same
sound in this conversion in this sample.
The next focal point is how attacking point (initial point) of each
sound being treated by YT encoding. You can compare Fig4 and Fig5 for
this test. They are the same waveforms as Fig1 & Fig2, but magnified
at attacking point of 100Hz for comparison. While original sine wave
has normal smooth waveforms constantly at -3db in Fig4, YT conversion
made it jump up to the ceiling of 0db or even beyond and hammered back
down to around -0.2db or so being normalized, however during this
process, original sine wave was completely distorted as you can
clearly see in Fig5 that it no longer represents smooth shape of sine
wave how it should be, YT conversion clearly clipped and distorted
original sound and made it almost look like a pulse wave flat at top
of every attacking portion of wave forms, it is no longer the
reproduction of since waves, just plain distortion here! It is the
same for any frequency above 100Hz, so I did not illustrate it for
each frequency since it is redundant, but they’re all more or less the
same observing Fig2 YT waveforms compare to Fig1. This basically
indicates that any sound presented at -3db or above after silence or
quiet passages will be guaranteed to be clipped and distorted after YT
conversion through current compression/limiting scheme.
Next thing is to observe decaying behavior before and after the
conversion by YT, and for this reason I intentionally faded out each
test tone linearly within 5sec of each blast of test tone. Comparison
can be viewed between Fig6 & Fig7. As you can see YT compression is
trying to boost every bits of your decaying sound until there’s no
sound exist to be boosted, completely against the law of nature how
things should decay naturally, such as echo, reverberation or even
life itself. What’s the point of amplifying decaying process? It is
just an extreme artificial attempt to preserve something doesn’t need
to be preserved or heard. Let everything decay naturally, why not?
Anyway, this will produce heavy breathing/pumping effect in existing
audio track unnecessarily including background noises along with
boosting up natural decay of instruments or reverberation effects. And
it is indeed, very BAD!
Finally, I present overlay of both waveforms, before and after YT
conversion at Fig8 to see how different they are comparing original
waveforms and converted waveforms by YT. Do you really want to sound
so different after uploading your clip being distorted and compressed
by YT without your permission or agreement in that matter? And that’s
my final point in this issue.
If you want to hear those test tone clips for comparison, be my guest!
But please make sure to turn your playback volume way down to minimum
before playing any of those clips, because they’re just tones and I
don’t want anyone to hurt their ears by accident listening to those
clips in high volume especially above 1,000Hz test tones, they can
indeed hurt your ears without care! So please turn your playback
volume down and be careful!
The same clip encoded from Mpeg to FLV (the same format as youtube
uses as default) locally without additional audio manipulation in
320x240 with 128kbps/44.1Hz audio bit/sampling rates:
http://casadelagos.com/YT_TestTone/ tones2.flv
Though, probably you need local FLV player to view the last clip.
Please feel free to comment or raise questions on this issue for
seeking prompt resolution in this matter.
> I've just been experimenting a bit, and getting encouraging results.
> My software won't allow me to normalize sound to a specific volume
> (you'd think it would, but no), so I have to do that by hand, but
> doable. It definitely helps, though.
> Procedure: Mix down audio, adjust volume so that it peaks at around
> -6dB. Then add audio track consisting of 20kHz sine wave at -8dB, and
> export and upload.
> If you could give the technical details of the incorrect compression/
> EQ problem, that might help YouTube fix the problem.
> On Jul 25, 11:08 pm, Sopranoguitar wrote:
> > No, current YT audio encoding is limited to 128kbps with 44.1khz max
> > 16,000 hz frequency, thus 20,000hz frequency would not be encoded by
> > YT audio scheme, but u can still upload audio track mixed with
> > 20,000hz using 48khz audio track :) Try it, it works :) Or if not,
> > pls, let me know.
> > On Jul 25, 11:55 pm, rewboss wrote:
> > > Just a thought though: Won't this drive dogs crazy?
This test and analysis above by viewer76x is very clear and thorough
to illustrate this issue and how serious problem it is. So, please
keep this thread alive until YT acknoledges and respond to it before
burried under other issues since this forum is not divided properly
based on nature of topics while some are more serious than others.
View his test result here, by your own eyes!: http://casadelagos.com/YT_TestTone/
> Hello folks! I just want to put the end to this recent audio
> compression problem by YT, so I have run and created some simple tests
> something easy for anyone to see and hear how screwed up it is, it is
> simply out of line and totally ridiculous how bad it is, once you can
> see it! Simply it is unacceptable what YT is doing for their audio
> recently. Here I put my two cents and a day of work to help, if
> ignored, I have no regret. However, if you care at all for the sound
> quality of your video clip posted on YT, just check out below what I
> found out through my tests today.
> 1) Test Tones: For this purpose, I have created special test tones to
> reveal recent YT compression behaviors. Very simple, I’ve chained
> 50-100-250-500-1000-5000-8000 and 10,000 Hz sine waves at -3db for
> 5sec blast each with gap of 1sec complete silence in between them and
> each blast are faded out using linear fade out within 5sec from -3db
> to -240db so that I can easily observe the differences between
> original waveforms and waveforms after encoded by YT from attack to
> decay, to speed and amount of compression, how they’re applied and
> behaves as a result.
> 2) Well, the results are astounding to the point that I’ve never-ever
> seen such an abuse of compression so badly by any professionals in
> audio management in any media from the inception of radio, TV or
> record industries, or even by telephone industries, and I’m 61 years
> old professional sound engineer and recording artist as my background
> lived through all of audio technology changes and developments since
> 1950’s. Worst audio in history by most popular internet media site in
> 2008, ironical indeed! How did it happen? I have no clue. They simply
> applied Jimi Hendrix guitar distortion technique to all of their audio
> tracks on their newly uploaded video clips, that was all to it. I
> can’t believe that it is intentional, I believe that someone in their
> company screwed it up really bad and there’s no quality control
> department or supervisor oversees audio portion of contents being
> uploaded since majority of audio are bad to begin with being captured
> by mobile phone, camcorders etc. Probably no one in their company
> watches high quality audio track video being uploaded to hear the
> differences, just too many clips there to see those differences clip
> by clip, therefore only users know how bad it is who uploaded quality
> audio along with their clips, and that’s the problem I see here. But
> it was fine until middle of June 2008, so that’s my point here to seek
> improvement promptly to eliminate that glitch in their system, or to
> restore something was working fine until 40days ago or so. It
> shouldn’t be a hard work on their part once the source of problem can
> be traced back correctly within this timeline. So, sooner is better to
> resolve this!
> 3) Now, please click this link:http://casadelagos.com/YT_TestTone/ > This is a result of my work and proof today to establish firm ground
> that how current YT compression scheme is harmful to anyone cares
> about sound quality of their clip, since they’re completely altering
> this original sound how it was meant to be by users. It is not even
> degrading it, it is altering it that you can easily witness by
> comparing Fig1 & Fig2, how different they are, and I also posted
> standard Mpeg to FLV conversion done by Riva converter at Fig3 locally
> on my PC, the result is a bit degraded, but basically maintaining the
> same shape except a bit low at 10,000hz and overall amplitude is a bit
> less, but we can live with that kind of difference, since original
> sound was not altered, just degraded a bit for overall level and
> frequency responses, but that’s normal trade off converting it to
> lower resolution for smaller file, however it is basically the same
> sound in this conversion in this sample.
> The next focal point is how attacking point (initial point) of each
> sound being treated by YT encoding. You can compare Fig4 and Fig5 for
> this test. They are the same waveforms as Fig1 & Fig2, but magnified
> at attacking point of 100Hz for comparison. While original sine wave
> has normal smooth waveforms constantly at -3db in Fig4, YT conversion
> made it jump up to the ceiling of 0db or even beyond and hammered back
> down to around -0.2db or so being normalized, however during this
> process, original sine wave was completely distorted as you can
> clearly see in Fig5 that it no longer represents smooth shape of sine
> wave how it should be, YT conversion clearly clipped and distorted
> original sound and made it almost look like a pulse wave flat at top
> of every attacking portion of wave forms, it is no longer the
> reproduction of since waves, just plain distortion here! It is the
> same for any frequency above 100Hz, so I did not illustrate it for
> each frequency since it is redundant, but they’re all more or less the
> same observing Fig2 YT waveforms compare to Fig1. This basically
> indicates that any sound presented at -3db or above after silence or
> quiet passages will be guaranteed to be clipped and distorted after YT
> conversion through current compression/limiting scheme.
> Next thing is to observe decaying behavior before and after the
> conversion by YT, and for this reason I intentionally faded out each
> test tone linearly within 5sec of each blast of test tone. Comparison
> can be viewed between Fig6 & Fig7. As you can see YT compression is
> trying to boost every bits of your decaying sound until there’s no
> sound exist to be boosted, completely against the law of nature how
> things should decay naturally, such as echo, reverberation or even
> life itself. What’s the point of amplifying decaying process? It is
> just an extreme artificial attempt to preserve something doesn’t need
> to be preserved or heard. Let everything decay naturally, why not?
> Anyway, this will produce heavy breathing/pumping effect in existing
> audio track unnecessarily including background noises along with
> boosting up natural decay of instruments or reverberation effects. And
> it is indeed, very BAD!
> Finally, I present overlay of both waveforms, before and after YT
> conversion at Fig8 to see how different they are comparing original
> waveforms and converted waveforms by YT. Do you really want to sound
> so different after uploading your clip being distorted and compressed
> by YT without your permission or agreement in that matter? And that’s
> my final point in this issue.
> If you want to hear those test tone clips for comparison, be my guest!
> But please make sure to turn your playback volume way down to minimum
> before playing any of those clips, because they’re just tones and I
> don’t want anyone to hurt their ears by accident listening to those
> clips in high volume especially above 1,000Hz test tones, they can
> indeed hurt your ears without care! So please turn your playback
> volume down and be careful!
> The same clip encoded from Mpeg to FLV (the same format as youtube
> uses as default) locally without additional audio manipulation in
> 320x240 with 128kbps/44.1Hz audio bit/sampling rates:http://casadelagos.com/YT_TestTone/tones2.flv
> Though, probably you need local FLV player to view the last clip.
> Please feel free to comment or raise questions on this issue for
> seeking prompt resolution in this matter.
> All the best.
> On Jul 26, 12:56 am, rewboss wrote:
> > I've just been experimenting a bit, and getting encouraging results.
> > My software won't allow me to normalize sound to a specific volume
> > (you'd think it would, but no), so I have to do that by hand, but
> > doable. It definitely helps, though.
> > Procedure: Mix down audio, adjust volume so that it peaks at around
> > -6dB. Then add audio track consisting of 20kHz sine wave at -8dB, and
> > export and upload.
> > If you could give the technical details of the incorrect compression/
> > EQ problem, that might help YouTube fix the problem.
> > On Jul 25, 11:08 pm, Sopranoguitar wrote:
> > > No, current YT audio encoding is limited to 128kbps with 44.1khz max
> > > 16,000 hz frequency, thus 20,000hz frequency would not be encoded by
> > > YT audio scheme, but u can still upload audio track mixed with
> > > 20,000hz using 48khz audio track :) Try it, it works :) Or if not,
> > > pls, let me know.
> > > On Jul 25, 11:55 pm, rewboss wrote:
> > > > Just a thought though: Won't this drive dogs crazy?
