Go Back   Cockos Incorporated Forums > REAPER Forums > REAPER General Discussion Forum

Reply
 
Thread Tools Display Modes
Old 11-09-2017, 07:53 AM   #1
Tsjanith
Human being with feelings
 
Join Date: Dec 2014
Posts: 30
Default Discrepancy in what people are hearing in output MP3

Sorry if the title is a bit vague, but I really don't know how else to describe it.

So I have a client who I am editing podcasts for. I did the editing, normalized the LUFS to -16, exported at 96kbps constant bit rate, mono, 44100 sample rate, in MP3 format via the LAME encoder.

What I hear is absolutely fluid sound, identical to the original files he has given me. He tells me that it is "metallic and barely audible." He sent a "test" file with my edited portion at the beginning, with his original file at the latter section of the file. I am unable to attach it as it comes up as "invalid file" here in the forums...

I tried it on a total of 6 other devices, and I am unable to hear even the slightest difference with my expensive mixing headphones. He says he sent it to 2 friends and they heard the exact same distortion...so I am stumped. He's an honest, straightforward guy. He's not some sociopath just trying to set me up for frustration, but everyone on my end hears consistency, and everyone on his end hears a metallic mess.

I have never encountered this before and I could really, REALLY use some insight into what might be happening

Thanks in advance!
Tsjanith is offline   Reply With Quote
Old 11-09-2017, 07:56 AM   #2
rhythmtech
Human being with feelings
 
Join Date: Dec 2008
Posts: 505
Default

96kbps - That's your problem.

Even at 320, it's a huge compromise on the integrity of the audio.

96 is like listening to audio in a running washing machine.
__________________
Producer | Engineer
Previous credits include:
BP Fallon, Emma Lou & The Agenda, The Hot Sprockets, Leaders Of Men, Travis Oaks
rhythmtech is offline   Reply With Quote
Old 11-09-2017, 07:58 AM   #3
EvilDragon
Human being with feelings
 
EvilDragon's Avatar
 
Join Date: Jun 2009
Location: Croatia
Posts: 19,557
Default

I wouldn't call 320 kbps a "huge compromise", at all - it works well for majority of cases. 96k IS a pretty huge blow to integrity of audio, though.
EvilDragon is offline   Reply With Quote
Old 11-09-2017, 08:00 AM   #4
rhythmtech
Human being with feelings
 
Join Date: Dec 2008
Posts: 505
Default

Quote:
Originally Posted by EvilDragon View Post
I wouldn't call 320 kbps a "huge compromise", at all - it works well for majority of cases. 96k IS a pretty huge blow to integrity of audio, though.
From an engineering standpoint, it is a huge compromise.

From a consumer standpoint, not so much.
__________________
Producer | Engineer
Previous credits include:
BP Fallon, Emma Lou & The Agenda, The Hot Sprockets, Leaders Of Men, Travis Oaks
rhythmtech is offline   Reply With Quote
Old 11-09-2017, 08:11 AM   #5
EvilDragon
Human being with feelings
 
EvilDragon's Avatar
 
Join Date: Jun 2009
Location: Croatia
Posts: 19,557
Default

Well, lossy compression is always a compromise, but 320 kbps is very listenable, I'd say. Of course, you'd never want to use MP3 for processing things (even though you could, if you're looking for some special effects or whatnot)...
EvilDragon is offline   Reply With Quote
Old 11-09-2017, 08:15 AM   #6
rhythmtech
Human being with feelings
 
Join Date: Dec 2008
Posts: 505
Default

Quote:
Originally Posted by EvilDragon View Post
Well, lossy compression is always a compromise, but 320 kbps is very listenable, I'd say. Of course, you'd never want to use MP3 for processing things (even though you could, if you're looking for some special effects or whatnot)...
Totally agreed, I have no problem with listening to a 320 on a consumer system.

However if i have to work with one inside a project, it is the last resort after other options have been exhausted.

Like i said, a compromise.
__________________
Producer | Engineer
Previous credits include:
BP Fallon, Emma Lou & The Agenda, The Hot Sprockets, Leaders Of Men, Travis Oaks
rhythmtech is offline   Reply With Quote
Old 11-09-2017, 08:15 AM   #7
EvilDragon
Human being with feelings
 
EvilDragon's Avatar
 
Join Date: Jun 2009
Location: Croatia
Posts: 19,557
Default

Yeah - no disagreement there.
EvilDragon is offline   Reply With Quote
Old 11-09-2017, 08:18 AM   #8
rhythmtech
Human being with feelings
 
Join Date: Dec 2008
Posts: 505
Default

Consensus reached! Hopefully OP finds his answer within it.
__________________
Producer | Engineer
Previous credits include:
BP Fallon, Emma Lou & The Agenda, The Hot Sprockets, Leaders Of Men, Travis Oaks
rhythmtech is offline   Reply With Quote
Old 11-09-2017, 08:28 AM   #9
Xenakios
Human being with feelings
 
Xenakios's Avatar
 
Join Date: Feb 2007
Location: Oulu, Finland
Posts: 6,738
Default

Quote:
Originally Posted by Tsjanith View Post
some insight into what might be happening
Is there a chance the mp3 you sent was somehow re-encoded/processed at the other end? Some sound uploading services might do that. The results would probably be pretty abysmal if the service did that to a 96kbps mp3 file...
__________________
For info on SWS Reaper extension plugin (including Xenakios' previous extension/actions) :
http://www.sws-extension.org/
https://github.com/Jeff0S/sws
--
Xenakios blog (about HourGlass and λ) :
http://xenakios.wordpress.com/
Xenakios is online now   Reply With Quote
Old 11-09-2017, 09:36 AM   #10
Tsjanith
Human being with feelings
 
Join Date: Dec 2014
Posts: 30
Default

Thanks all for the replies

The 96kbps is per his instructions, and it's worked fine for him in the past.

The issue here is, when I and everyone on my end (from a multitude of different devices) listens to it, it sounds perfectly fine. When he hears it, see OP.

I'll try and upload it to soundcloud so you all can hear.

Quote:
Originally Posted by Xenakios View Post
Is there a chance the mp3 you sent was somehow re-encoded/processed at the other end? Some sound uploading services might do that. The results would probably be pretty abysmal if the service did that to a 96kbps mp3 file...
It's been uploaded to both Google Drive and Libsyn. Again, sounds perfectly fine over here, but it comes through unacceptable from his end.
Tsjanith is offline   Reply With Quote
Old 11-09-2017, 09:41 AM   #11
Tsjanith
Human being with feelings
 
Join Date: Dec 2014
Posts: 30
Default

https://soundcloud.com/tobias-thunderjoo/test-ep75

You can skip to :33. The first 2/3 is the part that's been edited by me and processed by Reaper. At :59 is the unprocessed part which he claims sounds perfectly fine, as opposed to before :59. I genuinely can't hear any difference from here, nor can anyone else I have showed it to.

If you would, could you please give it a listen and tell me what you hear?

Thanks again
Tsjanith is offline   Reply With Quote
Old 11-09-2017, 09:52 AM   #12
rhythmtech
Human being with feelings
 
Join Date: Dec 2008
Posts: 505
Default

Quote:
Originally Posted by Tsjanith View Post
https://soundcloud.com/tobias-thunderjoo/test-ep75

You can skip to :33. The first 2/3 is the part that's been edited by me and processed by Reaper. At :59 is the unprocessed part which he claims sounds perfectly fine, as opposed to before :59. I genuinely can't hear any difference from here, nor can anyone else I have showed it to.

If you would, could you please give it a listen and tell me what you hear?

Thanks again
33->59 I hear a horrible washed out 96kps mp3. Exactly what i expected to hear.

59 onwards sounds better, much more clarity and presence.
__________________
Producer | Engineer
Previous credits include:
BP Fallon, Emma Lou & The Agenda, The Hot Sprockets, Leaders Of Men, Travis Oaks
rhythmtech is offline   Reply With Quote
Old 11-09-2017, 09:58 AM   #13
dug dog
Human being with feelings
 
Join Date: Jan 2009
Location: With Elvis Watching The Deflectives
Posts: 466
Default

I can hear a significant difference, too.
dug dog is offline   Reply With Quote
Old 11-09-2017, 09:59 AM   #14
Xenakios
Human being with feelings
 
Xenakios's Avatar
 
Join Date: Feb 2007
Location: Oulu, Finland
Posts: 6,738
Default

There shouldn't be much surprise the music bit sounds how it sounds like at that bit rate...(Is the music itself sourced from a lossy mp3 maybe?) Also the speech at the beginning sounds different to the later speech, has it been recorded with a different set up?
__________________
For info on SWS Reaper extension plugin (including Xenakios' previous extension/actions) :
http://www.sws-extension.org/
https://github.com/Jeff0S/sws
--
Xenakios blog (about HourGlass and λ) :
http://xenakios.wordpress.com/
Xenakios is online now   Reply With Quote
Old 11-09-2017, 10:05 AM   #15
RDBOIS
Human being with feelings
 
RDBOIS's Avatar
 
Join Date: Feb 2016
Location: It changes
Posts: 452
Default

Ok. I did listen to the audio.

Yes I hear a difference between the first and second audio clip.

This is more evident in this phrase:


"...in our series on cycling science..." @ 0:48

VS

"...in our series on cycling science..." @ 1:12

Listen carefully with headphones!

The part at 0:48 has some audio sound artifacts; could be characterized as "metallic", but I think there's a better word for that kind of artifact. It's more like a 'sizzling' or 'fluttering' sound.

This happens to me, at times, when I use plugins like ReaTune or even some De-Essers with improper settings, or placed in the wrong place in the chain of FXs. I also have this problem sometimes at the recording stage, especially if I'm too close to a cheap condenser microphone. Sometimes I think the Pop Filter is creating reflections; not sure if it's the cheap mic or reflections?

It could have to do with rendering and audio bit quality, but in my experience it often happens elsewhere in the tracking and/or mixing (i.e. due to plugins).

Sorry, but he is correct. Something happened when you processed his file?!

Good luck
RDBOIS is offline   Reply With Quote
Old 11-09-2017, 11:20 AM   #16
serr
Human being with feelings
 
Join Date: Sep 2010
Posts: 6,438
Default

Quote:
Originally Posted by Tsjanith View Post
I did the editing, normalized the LUFS to -16, exported at 96kbps constant bit rate, mono, 44100 sample rate, in MP3 format via the LAME encoder.

What I hear is absolutely fluid sound, identical to the original files he has given me.
If I read that literally, that means he gave you 96k mp3s for source files.
If you did work on them and exported back to mp3 again, they were reencoded to mp3. The audio went mp3 -> wav (behind the scenes) -> mp3. So the new mp3 would be only a subset of the data from the original. Treated as though it was uncompressed audio and run through the mp3 grinder again.

Those should not have sounded the same. Something is up on your end with monitoring if that scenario is correct.

If that's not what happened...
Perhaps he gave you uncompressed wav files for the source and you don't realize mp3 is lossy (and at different severity levels as the bitrate goes down)? And again then something is up with your monitor system where you didn't hear this?

Or maybe if this is a voice only recording, the degradation was a moot point to you but he is sensitive to it more like the perspective of a music listener?

Quote:
Originally Posted by Tsjanith View Post
He tells me that it is "metallic and barely audible." He sent a "test" file with my edited portion at the beginning, with his original file at the latter section of the file. I am unable to attach it as it comes up as "invalid file" here in the forums...

I tried it on a total of 6 other devices, and I am unable to hear even the slightest difference with my expensive mixing headphones. He says he sent it to 2 friends and they heard the exact same distortion...so I am stumped.
That's still a strong complaint suggesting something beyond mp3 destruction. You'll have to get him to tell you what media player he used. Ask what OS too. Is there something that can be set wrong in Windows to do that even with a proper media player? (Mac user so I assume such things but have no idea.)


Aside:
There are mp3 editor apps that let you edit and re-save mp3 with no further loss. If the source files are really already stepped down to 96k mp3, such a thing might be in order. Or does Reaper have a mode for that for all the apparent mp3 lovers out there?

Something doesn't add up. 96k mp3 sounds shitty for music but not at the level of a malfunctioning cassette deck or anything described as "unlistenable". Someone working with dynamic music in 24 bit HD formats might be "appalled" at the grainy compressed sound but still wouldn't usually call that literally "unlistenable".

So something must be up with someone's media player or OS.
__________________
Mac Pro 8x3.33GHz i7, 16GB, 256GB SSD(OS, apps), 3x2TB 7200 HD(data); MacBook Pro 2.8GHz, 6GB, 128GB SSD HD(OS, apps), 750GB 7200 HD(data); 2xTrue Precision 8; Apogee AD-16; 2xMOTU 828mk3, Evolution UC-33e; Faderport; WiRanger, iPad & the analog mixer has retired

Last edited by serr; 11-09-2017 at 11:27 AM.
serr is offline   Reply With Quote
Old 11-09-2017, 11:24 AM   #17
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 19,699
Default

Quote:
Originally Posted by rhythmtech View Post
96kbps - That's your problem.
agreed.

Quote:
Even at 320, it's a huge compromise on the integrity of the audio.
Are you willing to blind test this?
__________________
Choosing to play the simpler part instead of the fancier one, takes a lot of courage
karbomusic is online now   Reply With Quote
Old 11-09-2017, 11:47 AM   #18
Tsjanith
Human being with feelings
 
Join Date: Dec 2014
Posts: 30
Default

Okay so the absolute most confusing part is, I genuinely can't hear it. I've always had a keen ear for detail (no I haven't been to any loud concerts/had an ear infection recently) and it sounds identical over here, both ends.

I have never experienced anything like this before.

What in the world could be causing this??

Edit:

https://www.thepodcasthost.com/editi...for-a-podcast/

From what I've read, 96kbps is the standard for podcasts. Is there some special setting I need to incorporate for it to sound good for a podcast?

Last edited by Tsjanith; 11-09-2017 at 12:08 PM.
Tsjanith is offline   Reply With Quote
Old 11-09-2017, 12:03 PM   #19
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 19,699
Default

Quote:
Originally Posted by Tsjanith View Post
Okay so the absolute most confusing part is, I genuinely can't hear it. I've always had a keen ear for detail (no I haven't been to any loud concerts/had an ear infection recently) and it sounds identical over here, both ends.

I have never experienced anything like this before.

What in the world could be causing this??
I think 96k could be 'generally' audible if one knows what to listen for but I didn't listen to these to be clear, why don't you just export at something like 256kbs/44k and call it a day?
__________________
Choosing to play the simpler part instead of the fancier one, takes a lot of courage
karbomusic is online now   Reply With Quote
Old 11-09-2017, 12:15 PM   #20
serr
Human being with feelings
 
Join Date: Sep 2010
Posts: 6,438
Default

Maybe your client is hyper sensitive to generation loss? (Which doesn't add up with him delivering the sources to you as 96k mp3s!)

Someone have a link for that mp3 editor? Try that.

Again, if your client delivered lossless files, lose the mp3 conversions! Deliver FLAC files at the original audio resolution and get back to sanity. Then investigate what's going wrong with the mp3 conversions. It sounds like something is way beyond the expected level of destruction by what you say.

That you can't hear it but you feel that you are in familiar territory with your system and work really points to some media player or OS shenanigans on his end. I'm suggesting to figure that out or find a way to make it compatible in the long run, not just assign blame.
__________________
Mac Pro 8x3.33GHz i7, 16GB, 256GB SSD(OS, apps), 3x2TB 7200 HD(data); MacBook Pro 2.8GHz, 6GB, 128GB SSD HD(OS, apps), 750GB 7200 HD(data); 2xTrue Precision 8; Apogee AD-16; 2xMOTU 828mk3, Evolution UC-33e; Faderport; WiRanger, iPad & the analog mixer has retired
serr is offline   Reply With Quote
Old 11-09-2017, 12:21 PM   #21
Tsjanith
Human being with feelings
 
Join Date: Dec 2014
Posts: 30
Default

Thanks a lot guys.

I suppose I have a lot of trial and error ahead of me.

Your input has been very valuable
Tsjanith is offline   Reply With Quote
Old 11-09-2017, 12:22 PM   #22
rhythmtech
Human being with feelings
 
Join Date: Dec 2008
Posts: 505
Default

Quote:
Originally Posted by karbomusic View Post
Are you willing to blind test this?
As i said, i happily listen to 320 on my consumer system but saying its a massive compromise in a production environment, is in no way false.

Am i willing to blind test mp3 320 on my work monitoring system against 96k (the format i work in)? Yes, anyday. Have done at GSlutz many times through the years and had no issues.

Go down to 44.1k and its a totally different conversation we'd be having.
__________________
Producer | Engineer
Previous credits include:
BP Fallon, Emma Lou & The Agenda, The Hot Sprockets, Leaders Of Men, Travis Oaks
rhythmtech is offline   Reply With Quote
Old 11-09-2017, 12:24 PM   #23
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 19,699
Default

Quote:
Originally Posted by rhythmtech View Post
As i said, i happily listen to 320 on my consumer system but saying its a massive compromise in a production environment, is in no way false.

Am i willing to blind test mp3 320 on my work monitoring system against 96k (the format i work in)? Yes, anyday. Have done at GSlutz many times through the years and had no issues.

Go down to 44.1k and its a totally different conversation we'd be having.
Cool, I'll send you link for testing later, you can take the test and screen shot the results back here.

Oh wait, this isn't the right conversation or claim, this is about 320kbps MP3 vs 44.1k and you'd be right since I've never found anyone who could reliably tell the difference between CD and 320kbps IIRC. In that respect, I'd take issue with massive reduction in quality beacuse CD is the standard for this purpose (public consumption) anyway.
__________________
Choosing to play the simpler part instead of the fancier one, takes a lot of courage

Last edited by karbomusic; 11-09-2017 at 12:30 PM.
karbomusic is online now   Reply With Quote
Old 11-09-2017, 01:52 PM   #24
rhythmtech
Human being with feelings
 
Join Date: Dec 2008
Posts: 505
Default

Quote:
Originally Posted by karbomusic View Post
Cool, I'll send you link for testing later, you can take the test and screen shot the results back here.

Oh wait, this isn't the right conversation or claim, this is about 320kbps MP3 vs 44.1k and you'd be right since I've never found anyone who could reliably tell the difference between CD and 320kbps IIRC.
Thats why i said we'd be having a different conversation


Quote:
Originally Posted by karbomusic View Post
In that respect, I'd take issue with massive reduction in quality beacuse CD is the standard for this purpose (public consumption) anyway.
So why dont we just record in mp3 320kps format?
__________________
Producer | Engineer
Previous credits include:
BP Fallon, Emma Lou & The Agenda, The Hot Sprockets, Leaders Of Men, Travis Oaks
rhythmtech is offline   Reply With Quote
Old 11-09-2017, 01:57 PM   #25
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 19,699
Default

Quote:
Originally Posted by rhythmtech View Post

So why dont we just record in mp3 320kps format?
As far as "recording".... loses from reencoding which is irrelevant here.

However, I'm saying that 44.1k is public consumption (CD) and that I've not found anyone who can tell the difference (blind ABX) between it and 320kps. On a side note, I'm betting the vast majority would have a difficult time realiably identifying beyond chance the difference between 128kps (the new public consumption) and CD but 128 does tend to rear it's head from time to time depending on content so I have no argument with that one.
__________________
Choosing to play the simpler part instead of the fancier one, takes a lot of courage
karbomusic is online now   Reply With Quote
Old 11-09-2017, 02:06 PM   #26
rhythmtech
Human being with feelings
 
Join Date: Dec 2008
Posts: 505
Default

Quote:
Originally Posted by karbomusic View Post
On a side note, I'm betting the vast majority would have a difficult time realiably identifying beyond chance the difference between 128kps and CD but 128 does tend to rear it's head from time to time depending on content.
No doubt.

Ive often walked in on Mrs. Tech listening to some horrendous mess on youtube. When i point out to her to always search HD, she says it makes no difference.

That is what we are up against.
__________________
Producer | Engineer
Previous credits include:
BP Fallon, Emma Lou & The Agenda, The Hot Sprockets, Leaders Of Men, Travis Oaks
rhythmtech is offline   Reply With Quote
Old 11-09-2017, 02:19 PM   #27
RDBOIS
Human being with feelings
 
RDBOIS's Avatar
 
Join Date: Feb 2016
Location: It changes
Posts: 452
Default

Quote:
Originally Posted by Tsjanith View Post
Okay so the absolute most confusing part is, I genuinely can't hear it. What in the world could be causing this??
If you can't hear it, and assuming your hearing is not impaired, then what are you listening to the audio with?

Headphones, right?

What kind? Should not matter that much, but maybe your headphones are not working properly?

I have entry-level DT770 80ohm headphones and can hear the difference.

So... What did you do to the signal in the DAW when mixing? What was your role in this project? Did you compress, De-ess, gate, autotune, add a music track, limit?

Really, you need to figure out a way to 'hear' what we are hearing, because until you do there is no point trying to find a solution to fix the problem, because you won't be able to hear it!

You can't fix what you can't hear. Right?

If I was you I would try using another pair of headphones. If you stil can't hear it, then either it's your ears or some ear-to-brain processing (i.e. a mind perception issue). Sort of like when I was younger I never really heard the bass riff in songs, but now that I play bass I can hear all the riffs, all the time.
RDBOIS is offline   Reply With Quote
Old 11-10-2017, 05:18 PM   #28
Tsjanith
Human being with feelings
 
Join Date: Dec 2014
Posts: 30
Default

Quote:
Originally Posted by RDBOIS View Post
If you can't hear it, and assuming your hearing is not impaired, then what are you listening to the audio with?

Headphones, right?

What kind? Should not matter that much, but maybe your headphones are not working properly?

I have entry-level DT770 80ohm headphones and can hear the difference.

So... What did you do to the signal in the DAW when mixing? What was your role in this project? Did you compress, De-ess, gate, autotune, add a music track, limit?

Really, you need to figure out a way to 'hear' what we are hearing, because until you do there is no point trying to find a solution to fix the problem, because you won't be able to hear it!

You can't fix what you can't hear. Right?

If I was you I would try using another pair of headphones. If you stil can't hear it, then either it's your ears or some ear-to-brain processing (i.e. a mind perception issue). Sort of like when I was younger I never really heard the bass riff in songs, but now that I play bass I can hear all the riffs, all the time.
I have Shure SRH440. If it's malfunctioning, could have fooled me! I can still hear flaws in music recordings with it that I cannot detect (at all) with normal speakers.

In the DAW I used Reafir to remove background noise, but then removed it and rerendered when he told me about the issue. It was still there.... I don't think it could have been Reafir, unless it did something destructive. I'm sending him some test files to see if I can get to the bottom of it.

When I bounced back and fourth between the time stamps you listed, I was able to detect the absolute most superficial, minuscule difference...but I REALLY had to concentrate.

Is it really that overt? If so, I might actually have to get my hearing checked...strange though. I have had 0 issues until now.

I don't think the ear-to-brain processing theory makes sense. I mean, you can always hear a bass riff if you try, no?
Tsjanith is offline   Reply With Quote
Old 11-11-2017, 02:45 AM   #29
JamesPeters
Human being with feelings
 
JamesPeters's Avatar
 
Join Date: Aug 2011
Location: Winnipeg
Posts: 1,010
Default

Quote:
Originally Posted by Tsjanith View Post
...exported at 96kbps constant bit rate
For music, I've considered 128kbps "barely acceptable". For voice-only content, maybe 96kbps can be fine...but that's not the only setting to consider.

When rendering, check the further option of encoding quality. Set it to "maximum (slow)". See if that helps.

I realized years ago if I set the encoder to that setting (not just with LAME but also Fraunhofer) I could get away with encoding at a somewhat lower bitrate than usual since the artifacts were reduced. Also, the difference in time spent encoding was hardly noticeable once CPUs became more powerful, so there's no significant tradeoff in terms of quality versus speed anymore. Just encode on the "slowest" setting for maximum quality.

I just listened to the podcast sample you posted at Soundcloud. I can barely tell a difference in the content at the two time indexes you mentioned, and I'm listening with Steinberg MR816X (and Adam A7 monitors, or Sennheiser HD280 Pro). Perhaps before it was uploaded to Soundcloud there was a more noticeable difference. (It gets re-encoded when it's uploaded to Soundcloud, I think.)
__________________
http://petersamplification.com

Last edited by JamesPeters; 11-11-2017 at 02:55 AM.
JamesPeters is offline   Reply With Quote
Old 11-11-2017, 09:09 AM   #30
Tsjanith
Human being with feelings
 
Join Date: Dec 2014
Posts: 30
Default

Update: I've rendered the file in 320 kbps and this is what he had to say "It's still there in ep75, although it sounds quite different. Neither better nor worse, just a different tone to it."

I also tried retracing my steps and adding what I added to that episode to a subsequent episode (reafir, EQ) to try and recreate it in order to know what to avoid doing. It didn't show up.

So, bitrate can be ruled out in contrast to the consensus here.

Reafir and EQ weren't the cause either.

Still massively, massively stumped

Quote:
Originally Posted by JamesPeters View Post
For music, I've considered 128kbps "barely acceptable". For voice-only content, maybe 96kbps can be fine...but that's not the only setting to consider.

When rendering, check the further option of encoding quality. Set it to "maximum (slow)". See if that helps.

I realized years ago if I set the encoder to that setting (not just with LAME but also Fraunhofer) I could get away with encoding at a somewhat lower bitrate than usual since the artifacts were reduced. Also, the difference in time spent encoding was hardly noticeable once CPUs became more powerful, so there's no significant tradeoff in terms of quality versus speed anymore. Just encode on the "slowest" setting for maximum quality.

I just listened to the podcast sample you posted at Soundcloud. I can barely tell a difference in the content at the two time indexes you mentioned, and I'm listening with Steinberg MR816X (and Adam A7 monitors, or Sennheiser HD280 Pro). Perhaps before it was uploaded to Soundcloud there was a more noticeable difference. (It gets re-encoded when it's uploaded to Soundcloud, I think.)
No I checked with my client and it sounds exactly as it does in all its other forms. I already tried rendering at maximum (slow) and it's a no go.

Thanks for your suggestions though
Tsjanith is offline   Reply With Quote
Old 11-11-2017, 11:12 AM   #31
DVDdoug
Human being with feelings
 
Join Date: Jul 2010
Location: Silicon Valley, CA
Posts: 1,505
Default

I think we should just accept the fact that MP3 is lossy compression. The lower the bitrate, the more dats is being thrown-away. And the "damage" is cumulative, even if you re-encode to the same bitrate.

Accumulated damage is the main reason to avoid MP3 during production. If you want a lossy format, you should compress ONCE as the final step. (AAC is supposed to be more-immune to multiple generations of compression.)

If you want perfection, don't use lossy compression!

But high-quality, high-bitrate, first-generation MP3 is usually transparent in a blind ABX test (depending on the program material and the listener's ability to hear compression artifacts). A lot of people claim that MP3 compression is "obvious" but most of those people think it's so obvious that they have not bothered to do a proper, scientific, level-matched, blind, ABX Test. The truth is with a high-quality MP3 (or AAC or AC3) it's very-hard to hear a difference (with most program material) and you have to listen very carefully... if you can hear a difference at all.

And, (with a high-bitrate MP3) it's unlikely for anyone to hear/identify compression artifacts without the original to compare to. (Every time I thought I was hearing a compression artifact on my 'V0' MP3s made from CDs, it's turned-out that the same "defect" was on the CD.)
DVDdoug is offline   Reply With Quote
Old 11-11-2017, 11:55 AM   #32
Lunar Ladder
Human being with feelings
 
Join Date: Jan 2016
Posts: 168
Default

Quote:
Originally Posted by Tsjanith View Post
He tells me that it is "metallic and barely audible."
If he means the whole material is "metallic and barely audible", you can discount all the hifi musings and try to figure out what makes it sound like that at the customer's end. Even though you can hear the artefacts, and there are different opinions on how horrible they are , the signal is nowhere near "metallic and barely audible" as a whole.

If that's indeed what the client meant and I didn't misinterpret what you're saying, ask them to record, with a microphone, what it sounds like at their end. If it's that horrible, they can just use their smartphone or whatever, and send you a video. Sure that degrades the audio even further, hah, but if there's something THAT MASSIVELY wrong with the signal, you'll be able to hear if there's indeed something out of the ordinary going on over there.
Lunar Ladder is offline   Reply With Quote
Old 11-11-2017, 01:31 PM   #33
lolilol1975
Human being with feelings
 
Join Date: Dec 2015
Posts: 1,189
Default

Quote:
Originally Posted by rhythmtech View Post
Consensus reached! Hopefully OP finds his answer within it.
Personally, I have no problem using org Opus or Musepack at high settings (at least 240 kbps). My feeling is they sound a tad better than mp3. Of course, it's still a lossy compression, but quite decent. I would NEVER export at 96 kbps unless it's only for talking like a podcast, and not music.
lolilol1975 is online now   Reply With Quote
Old 11-11-2017, 01:42 PM   #34
lolilol1975
Human being with feelings
 
Join Date: Dec 2015
Posts: 1,189
Default

Quote:
Originally Posted by karbomusic View Post
As far as "recording".... loses from reencoding which is irrelevant here.

However, I'm saying that 44.1k is public consumption (CD) and that I've not found anyone who can tell the difference (blind ABX) between it and 320kps. On a side note, I'm betting the vast majority would have a difficult time realiably identifying beyond chance the difference between 128kps (the new public consumption) and CD but 128 does tend to rear it's head from time to time depending on content so I have no argument with that one.
At 128 kbps or lower, everything above 16 kHz is simply filtered out and it's pretty noticeable on material that is very well recorded and has frequencies all the way up to 20 kHz and above. Above 128 kbps, mp3 adds the most important missing frequencies when needed. You can verify that with a simple spectrogram.
BTW, with a spectrogram, you can also verify that vinyles reproduce frequencies up to 30 kHz.
lolilol1975 is online now   Reply With Quote
Old 11-11-2017, 02:12 PM   #35
grinder
Human being with feelings
 
grinder's Avatar
 
Join Date: Jan 2009
Location: New Zealand
Posts: 1,554
Default

Sorry I have said it here I have said it anywhere there are people who can tell the difference between Mp3 and 16it 44 and upward and I am one of them.
Interestingly tape has much more information than digital it would be common knowledge by now if it made as much money.

Grinder
grinder is offline   Reply With Quote
Old 11-11-2017, 02:26 PM   #36
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 4,484
Default

I can't hear a dramatic difference. Mainly the already sibilant voice is made worse in the 96k version, and the music has that nasty streamed audio quality.

What's the headroom on the mp3? Maybe he's got a cheap DAC that is distorting it if there's ISP's?
Judders is offline   Reply With Quote
Old 11-11-2017, 02:32 PM   #37
lolilol1975
Human being with feelings
 
Join Date: Dec 2015
Posts: 1,189
Default

Quote:
Originally Posted by rhythmtech View Post
33->59 I hear a horrible washed out 96kps mp3. Exactly what i expected to hear.

59 onwards sounds better, much more clarity and presence.
I think I can hear a difference, mainly the first version is more aggressive, probably what your client calls metallic.

Also I noticed from looking at the waveforms that the two takes aren't playing at the same speed. Did you time stretch the audio ? It can introduce artefacts. Try different algorithms and post them here. Also maybe play with some dither and noise shapig during reder.
In any case, I got much better results by adding ReaXcomp and toning down the band 200-1000Hz with a threshold of -14dB and a ratio of 3:1. This removes some nasal frequencies. And the result is much better with the second version than with the first.

Last edited by lolilol1975; 11-11-2017 at 02:47 PM.
lolilol1975 is online now   Reply With Quote
Old 11-11-2017, 02:41 PM   #38
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 19,699
Default

Quote:
Originally Posted by lolilol1975 View Post
You can verify that with a simple spectrogram.
I'm interested in how many people can reliably hear 128.

http://www.noiseaddicts.com/2009/03/...-test-128-320/
__________________
Choosing to play the simpler part instead of the fancier one, takes a lot of courage
karbomusic is online now   Reply With Quote
Old 11-11-2017, 02:45 PM   #39
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 4,484
Default

Quote:
Originally Posted by karbomusic View Post
Oh wait, this isn't the right conversation or claim, this is about 320kbps MP3 vs 44.1k and you'd be right since I've never found anyone who could reliably tell the difference between CD and 320kbps IIRC. In that respect, I'd take issue with massive reduction in quality beacuse CD is the standard for this purpose (public consumption) anyway.
I've found it depends on the source. Some things you can (particularly when concentrating on the stereo image) and some you can't. Sometimes older recordings sound higher quality to people as mp3 because the encoding distortion adds some high end excitement! They are certainly very close and often indistinguishable though.

I recently did an online test, and shared the results with a few people. About 1/3 of the clips were reliably identified, and about 1/5 reliably misidentified, but if you took the total scores of each test it would appear to be random chance. Not in any way a big enough sample to be statistically significant, but once you hear something you can't un-hear it, if you know what I mean.
Judders is offline   Reply With Quote
Old 11-11-2017, 02:55 PM   #40
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 4,484
Default

Quote:
Originally Posted by karbomusic View Post
I'm interested in how many people can reliably hear 128.

http://www.noiseaddicts.com/2009/03/...-test-128-320/
Again, I wouldn't be so quick to conclude that misidentification means identical. It may well do, but it could be more complicated than that. I don't think that was a good source for the test either.

I completely agree that mp3 is fine, especially since the encoding became a lot better than when they first appeared, but the fact that people think they hear a difference in critical listening situations, even if they misidentify which is which, is not logically conclusive proof that there is no difference. It could mean that, but it could mean something else is at work...

Edit: here's another test: https://www.npr.org/sections/thereco...-audio-quality
Judders is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 12:13 AM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2017, vBulletin Solutions Inc.