IPB

Welcome Guest ( Log In | Register )

2 Pages V   1 2 >  
Reply to this topicStart new topic
Quicktime pro resampler quality?
Mix3dmessagez
post Mar 28 2012, 20:23
Post #1





Group: Members
Posts: 87
Joined: 22-March 09
Member No.: 68252



I use quicktime to downsample needledrops to 48hz tvbr *for ipod use*
I've also come across foobar sox resampler which I've read is the best
I've tried to do comparisons *downsample with sox to 48000hz then convert with quicktime or convert using quicktime*
Usually the bitrates will be the same *on average* however sometimes sox downsamples will be one bit higher *on average*
I've tried to look at the individual bitrates for the first few seconds of each song *Just to see if there's a conclusive difference* and they seem to handle things quit differently, I notice the bitrate changes faster when downsampled with sox, I know that probably isn't a determinant as to quality, but it's all I know
With sox I use very high, 48000hz, 95% passband, allow aliasing, 25% phase response *someone at another forum said those are best settings*
On quicktime, best quality, variable bit rate, and recommended sample rate
I also want to know as if Sox is better I can resample all needledrops to 48hz flac *from 96 to save space*
Go to the top of the page
+Quote Post
DVDdoug
post Mar 28 2012, 20:48
Post #2





Group: Members
Posts: 2542
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



QUOTE
*someone at another forum said those are best settings*
I think you worry too much about stuff you can't hear...

Whenever I've needed to resample (usually between 44.1 an 48kHz), I've never had a problem or noticed ANY sound quality difference between the original and the resampled file. I just used whatever audio editor I was using at the time. Mostly, I've used GoldWave, which is reputed to have a terrible resampler. Still, I don't hear anything... It's apparantly been upgraded since the tests were done, but I've been using GoldWave for a long time so I've probably use the old re-sampler.

Of course, if downsample to 12kHz 8-bits, you're going to hear quality loss. wink.gif

QUOTE
...to downsample needledrops...
Compared to the limitations of analog vinyl, any "damage" done by downsampling is not likely to be noticeable.

P.S.
If you think there might be a difference with different resamplers, you can do some ABX Tests. The goal would be to choose a resampler (and settings) that makes it impossible for you to hear a difference between the original and the re-sampled file.

My guess is that you won't hear a difference with any resampler... You didn't claim to hear a difference. anyway... You said someone else told you there was a difference...

This post has been edited by DVDdoug: Mar 28 2012, 20:59
Go to the top of the page
+Quote Post
IgorC
post Mar 28 2012, 21:20
Post #3





Group: Members
Posts: 1540
Joined: 3-January 05
From: ARG/RUS
Member No.: 18803



http://src.infinitewave.ca/ has Apple resampler in its DB.
Go to the top of the page
+Quote Post
Mix3dmessagez
post Mar 28 2012, 22:27
Post #4





Group: Members
Posts: 87
Joined: 22-March 09
Member No.: 68252



QUOTE (DVDdoug @ Mar 28 2012, 15:48) *
[giant quote removed]

What are the limitations of analog vinyl?I choose vinyl because it is better to me than listening to digital cd's with -10DB gains on some of my favorite music, not to mention in instances such as with radiohead *their 45rpm* vs the brickwalled cd's is better to me and has allot more dynamics going on *once again in my opinion*.If there's any benefit to digital, people who master cd's seem to try and kill that.

I will abx it, it's just all i have are ipod earbuds so i dont think it'll do any good, if i upload samples would you mind abxing?

QUOTE (IgorC @ Mar 28 2012, 16:20) *
http://src.infinitewave.ca/ has Apple resampler in its DB.

Hello, which one is the quicktime one?I see tiger and lepoerd, but I don't know what that means, and which one shows me what I want to know?Is it the passband test?

This post has been edited by db1989: Apr 12 2012, 09:26
Go to the top of the page
+Quote Post
saratoga
post Mar 28 2012, 23:19
Post #5





Group: Members
Posts: 4869
Joined: 2-September 02
Member No.: 3264



QUOTE (Mix3dmessagez @ Mar 28 2012, 17:27) *
What are the limitations of analog vinyl?


Low SNR, high distortion.
Go to the top of the page
+Quote Post
Mix3dmessagez
post Mar 28 2012, 23:29
Post #6





Group: Members
Posts: 87
Joined: 22-March 09
Member No.: 68252



But what if the source material is digital and pressed on vinyl, couldn't that then be eradicated?

This post has been edited by db1989: Apr 12 2012, 09:27
Reason for edit: deleting unnecessary full quote of above post
Go to the top of the page
+Quote Post
saratoga
post Mar 28 2012, 23:31
Post #7





Group: Members
Posts: 4869
Joined: 2-September 02
Member No.: 3264



To be clear, I'm saying that the process of creating and then playing a vinyl introduces noise and distortion.

This post has been edited by db1989: Apr 12 2012, 09:27
Reason for edit: as above
Go to the top of the page
+Quote Post
nu774
post Mar 29 2012, 05:05
Post #8





Group: Developer
Posts: 514
Joined: 22-November 10
From: Japan
Member No.: 85902



If you are using MacOS X, you can try afconvert (CLI application), which lets you to control detailed parameters of CoreAudio SRC from command line options.
QuickTime is working on top of CoreAudio; AFAIK you can only configure "rendering quality" from QuickTime exporter. I don't know how this is mapped to CoreAudio parameter.

AAC codec component of CoreAudio has it's own SRC, which is not configurable in terms of quality.
This SRC is used when you are encoding to AAC and sampling rates of source/target are different;
From what I read at the Apple site, "mastered for iTunes" droplet works in two pass;
First it runs afconvert with the highest SRC setting and convert to 44.1kHz LPCM, then encodes into AAC. This way it avoids rate conversion by AAC codec component.

In the past (at the time of QuickTime 7.6.5 or so), I have heard about audible aliasing problem of QT AAC sample rate converter from a HA user.
I suppose it doesn't have such a serious one in the recent versions, but I cannot say for sure.
Go to the top of the page
+Quote Post
SebastianG
post Mar 29 2012, 10:35
Post #9





Group: Developer
Posts: 1317
Joined: 20-March 04
From: Göttingen (DE)
Member No.: 12875



I don't know about how well Quicktime does resampling. I just wanted to comment on some things here.

Mix3dmessagez, it seems you're lumping two processes together into one. Downsampling is the process of reducing the sampling rate. Encoding is -- in this context -- the process of turning a PCM signal into some other compressed format.

I guessing that you're trying to convert a 96 kHz signal into a compressed 48 kHz signal. So far so good. Just pick a decent resampler for downsampling (at least sox qualifies as such) and a decent encoder for compressing the downsampled signal.

QUOTE (Mix3dmessagez @ Mar 28 2012, 20:23) *
With sox I use very high, 48000hz, 95% passband, allow aliasing, 25% phase response *someone at another forum said those are best settings*

Sounds like overkill to me. Though, I don't know what "25% phase response" is supposed to mean. A 20 kHz pass band (84% for fs=48000 Hz) is probably sufficient. Allowing a little bit of aliasing above 20 kHz is ok, I guess. The content above 20 kHz is likely to be removed during compression anyways.

This post has been edited by SebastianG: Mar 29 2012, 10:36
Go to the top of the page
+Quote Post
Wombat
post Mar 29 2012, 16:58
Post #10





Group: Members
Posts: 988
Joined: 7-October 01
Member No.: 235



QUOTE (SebastianG @ Mar 29 2012, 10:35) *
Though, I don't know what "25% phase response" is supposed to mean.

It is an audiophile attempt to solve a problem we canīt hear. With telling sox to use 25% phase response you try to get rid of the "pre" ringing around fs/2 and trade it against lots more "post" ringing. This is meant to sound much better. The little detail that you alter the signal completely into the audible signal of cause doesnīt count smile.gif

Go to the top of the page
+Quote Post
Wombat
post Mar 31 2012, 03:15
Post #11





Group: Members
Posts: 988
Joined: 7-October 01
Member No.: 235



I hope it is of some help. I did some simple examples on how these resampling filters act when using sox with a full scale impulse.
I think these spectral responses tell more as these waveforms often shown for pre- and post ringing because here you can clearly see what frequencies we talk about.
I am with Sebastian here that choosing parameters that preserve frequencies above 20kHz should be enough for us humans and this can be reached without violating the phase response. So even IF ringing was audible it shouldnīt be of an issue with some gentle chosen setting as you can see.
sox only allows a minimum BW of 85% with allowed aliasing so i did use that. I added a high BW, steep setting without aliasing as reference.
Also like mentioned elsewhere it all depends what your DAC makes out of all this. There are DACs that donīt let anythging pass above 20kHz which makes all this pretty theoretical.
Still i am waiting for some serious ABX results since my own told me i canīt ABX garbage from ringing smile.gif

Go to the top of the page
+Quote Post
Mix3dmessagez
post Apr 12 2012, 01:40
Post #12





Group: Members
Posts: 87
Joined: 22-March 09
Member No.: 68252



QUOTE (nu774 @ Mar 29 2012, 00:05) *
If you are using MacOS X, you can try afconvert (CLI application), which lets you to control detailed parameters of CoreAudio SRC from command line options.
QuickTime is working on top of CoreAudio
[…]
In the past (at the time of QuickTime 7.6.5 or so), I have heard about audible aliasing problem of QT AAC sample rate converter from a HA user.
I suppose it doesn't have such a serious one in the recent versions, but I cannot say for sure.

I use windows sadly, but I tested this by converting to apple lossless in qucktime and on itunes.I then converted both to quicktime, and compared with the directly downsampled from aac quicktime.They all had the same exact bitrate, this leads me to believe that the downsampling is exactly in the same in quicktime, itunes, and aac codec within quicktime.

The latest windows version is 7.7.1, so I assume whatever issues present would either be fixed or atleast tweaked from that far back *hopefully*

QUOTE (SebastianG @ Mar 29 2012, 05:35) *
I don't know about how well Quicktime does resampling. I just wanted to comment on some things here.

Mix3dmessagez, it seems you're lumping two processes together into one. Downsampling is the process of reducing the sampling rate. Encoding is -- in this context -- the process of turning a PCM signal into some other compressed format.

I guessing that you're trying to convert a 96 kHz signal into a compressed 48 kHz signal. So far so good. Just pick a decent resampler for downsampling (at least sox qualifies as such) and a decent encoder for compressing the downsampled signal.
[…]

Sox is what I'm comparing to quicktime, I am trying to decide between which of the two is better, I already have converted all my rips and downsampled in quicktime, and am trying to see which is the better choice.I've heard great things about sox, and made this thread to find out more about quicktime to see how it stacks.

QUOTE (Wombat @ Mar 29 2012, 11:58) *
QUOTE (SebastianG @ Mar 29 2012, 10:35) *
Though, I don't know what "25% phase response" is supposed to mean.

It is an audiophile attempt to solve a problem we canīt hear. With telling sox to use 25% phase response you try to get rid of the "pre" ringing around fs/2 and trade it against lots more "post" ringing. This is meant to sound much better. The little detail that you alter the signal completely into the audible signal of cause doesnīt count smile.gif

Hmmm, well if that's the case I assume it's better that way than normal, just to be on the safe side to have the best options enabled?

QUOTE (Wombat @ Mar 30 2012, 22:15) *
I am with Sebastian here that choosing parameters that preserve frequencies above 20kHz should be enough for us humans and this can be reached without violating the phase response. So even IF ringing was audible it shouldnīt be of an issue with some gentle chosen setting as you can see. […] Still i am waiting for some serious ABX results since my own told me i canīt ABX garbage from ringing smile.gif
[image]

I have uploaded two samples, both 25 seconds, one downsampled and dithered in foobar with sox resampling to 48000 with 95% passband, very high quality, aliasing allowed 25% phase response.
The quicktime one, no dithering or downsampling bits in foobar *and instead did all in quicktime*, and downsampled from 96000 to 48000 *in quicktime*, allowing quicktime to completely handle the dithering and downsampling, and bit depth *To fully access its functionality*

Quicktime
http://www.mediafire.com/?hgjv2mfu8b095w2
Sox
http://www.mediafire.com/?xcdj0hj7yawio0g

I would like any of the amazing audio enthusiasts here to see if you can hear a difference, as I know many of you are listening on much better equipment with ears for detail and everything more than I do.Please post your results!

This post has been edited by db1989: Apr 12 2012, 09:32
Reason for edit: Please do not unnecessarily quote entire posts; edit them to the relevant sections.
Go to the top of the page
+Quote Post
Wombat
post Apr 12 2012, 02:14
Post #13





Group: Members
Posts: 988
Joined: 7-October 01
Member No.: 235



If there really someone was interested in listening these files you should offer the original source so if there is an audible difference one has to check what is nearer to the original.
Go to the top of the page
+Quote Post
nu774
post Apr 12 2012, 03:26
Post #14





Group: Developer
Posts: 514
Joined: 22-November 10
From: Japan
Member No.: 85902



If you are interested, now qaac lets you play around with CoreAudio sample rate converter. If you want to try sample rate converter only, use -D option (WAV output).
https://sites.google.com/site/qaacpage/
https://github.com/nu774/qaac/wiki/Sample-rate-conversion
Go to the top of the page
+Quote Post
Mix3dmessagez
post Apr 12 2012, 04:23
Post #15





Group: Members
Posts: 87
Joined: 22-March 09
Member No.: 68252



QUOTE (Wombat @ Apr 11 2012, 21:14) *
If there really someone was interested in listening these files you should offer the original source so if there is an audible difference one has to check what is nearer to the original.


I don't know how to cut flac files


QUOTE (nu774 @ Apr 11 2012, 22:26) *
If you are interested, now qaac lets you play around with CoreAudio sample rate converter. If you want to try sample rate converter only, use -D option (WAV output).
https://sites.google.com/site/qaacpage/
https://github.com/nu774/qaac/wiki/Sample-rate-conversion


Hi, I would like to know if you could tell me how to make tvbr like in quicktime pro with highest sample rate conversion, and how to create it in "custom" foobar encoder option please?

I tried myslef and got this

1 out of 1 tracks converted with major problems.

Source: "C:\Users\************\Downloads\Kid Cudi - Man on the Moon II- The Legend of Mr Rager [Vinyl][FLAC]\11 MANIAC.flac"
An error occurred while writing to file (The encoder has terminated prematurely with code 1 (0x00000001); please re-check parameters) : "C:\Users\Patricia Shustin\Desktop\alac\11. MANIAC (feat. Cage & St. Vincent).mp4"
Additional information:
Encoder stream format: 96000Hz / 2ch / 24bps
Command line: "C:\Users\******\Desktop\desktop\qaac_1.31\x86\qaac.exe" --native-resampler=bats,127 "11. MANIAC (feat. Cage & St. Vincent).mp4"
Working folder: C:\Users\*******\Desktop\alac\

Conversion failed: The encoder has terminated prematurely with code 1 (0x00000001); please re-check parameters

This post has been edited by Mix3dmessagez: Apr 12 2012, 04:43
Go to the top of the page
+Quote Post
nu774
post Apr 12 2012, 05:08
Post #16





Group: Developer
Posts: 514
Joined: 22-November 10
From: Japan
Member No.: 85902



Try something like the following
QUOTE
-V 127 --native-resampler=bats,127 - -o %d

-V means TVBR AAC encoding. If you want WAV output, change "-V 127" to "-D".
"-" (a hyphen before -o) means receiving audio from stdin/pipe, which foobar2000 will stream audio to.
-o %d means to set output filename to %d (placeholder of foobar2000).
You will always need "- -o %d" part. Others are up to your own taste.
Go to the top of the page
+Quote Post
Mix3dmessagez
post Apr 12 2012, 05:35
Post #17





Group: Members
Posts: 87
Joined: 22-March 09
Member No.: 68252



Hi, thank you so much!I see you are the creator of this amazing app.

I still am having trouble though, i used qaac as the encodder, extension as mp4, perimeters as -V 127 --native-resampler=bats,127 - -o %d *thanks to you smile.gif* and format as lossy, 16 bit maximum, encoder name TVBR, settings TVBR.

I am providing all this info as at this point I'm sure the error now is not on the permiters part but something with my configuration.

I already use your great program to convert flac to alac using qaac, which keeps the 24 bits and high sample rate, then convert using quicktime pro, sadly, all the tags become lost though, does this preserve tags?

This post has been edited by db1989: Apr 12 2012, 09:32
Reason for edit: as in post #6
Go to the top of the page
+Quote Post
nu774
post Apr 12 2012, 06:41
Post #18





Group: Developer
Posts: 514
Joined: 22-November 10
From: Japan
Member No.: 85902



Still getting errors? See what foobar2000 says on console. If it says like:
QUOTE
(The encoder has terminated prematurely with code 1 (0x00000001); please re-check parameters)

The exit code 1 means parameter is invalid, so CLI encoder configuration on foobar2000 have to be fixed (although setting on your last post seems OK).
If the code is 2, it means parameter at least was successfully parsed, but some problem occurred after that.
If you are getting code 2, adding "--log %d.txt" to command line might help you. With --log option specified, qaac will output logging/error messages to the log file (in this case, on the same folder with the resulting file).

Also, you can simply try directly running from command prompt, so that you will able to see all the messages from qaac.
Go to the top of the page
+Quote Post
nu774
post Apr 12 2012, 06:49
Post #19





Group: Developer
Posts: 514
Joined: 22-November 10
From: Japan
Member No.: 85902



Ah, I've forgotten one thing:
You will need "--rate 44100" or something to specify target sample rate, if you want to test sample rate conversion!
Go to the top of the page
+Quote Post
Mix3dmessagez
post Apr 12 2012, 06:56
Post #20





Group: Members
Posts: 87
Joined: 22-March 09
Member No.: 68252



QUOTE (nu774 @ Apr 12 2012, 01:49) *
Ah, I've forgotten one thing:
You will need "--rate 44100" or something to specify target sample rate, if you want to test sample rate conversion!


1 out of 1 tracks converted with major problems.

Source: "C:\Users\******\Downloads\Kid Cudi - Man on the Moon II- The Legend of Mr Rager [Vinyl][FLAC]\11 MANIAC.flac"
An error occurred while writing to file (The encoder has terminated prematurely with code 1 (0x00000001); please re-check parameters) : "C:\Users\*******\Desktop\alac\11. MANIAC (feat. Cage & St. Vincent).mp4"
Additional information:
Encoder stream format: 96000Hz / 2ch / 16bps
Command line: "C:\Users\P*******\Desktop\desktop\qaac_1.31\x86\qaac.exe" -V 127 --rate 48000 --native-resampler=bats,127 - -o "11. MANIAC (feat. Cage & St. Vincent).mp4"
Working folder: C:\Users\******\alac\

Conversion failed: The encoder has terminated prematurely with code 1 (0x00000001); please re-check parameters

Am I doing something wrong?

heres a pic



This post has been edited by Mix3dmessagez: Apr 12 2012, 07:22
Go to the top of the page
+Quote Post
nu774
post Apr 12 2012, 07:25
Post #21





Group: Developer
Posts: 514
Joined: 22-November 10
From: Japan
Member No.: 85902



QUOTE (Mix3dmessagez @ Apr 12 2012, 14:56) *
Command line: "C:\Users\P*******\Desktop\desktop\qaac_1.31\x86\qaac.exe" -V 127 --rate 48000 --

Please upgrade to 1.35 if you are running 1.31.
--native-resampler was updated very recently.
Go to the top of the page
+Quote Post
Mix3dmessagez
post Apr 12 2012, 07:39
Post #22





Group: Members
Posts: 87
Joined: 22-March 09
Member No.: 68252



thanks, solved my issue!last problem though, I'm gettign bitrates of 200-300's
here's what it says in tools
qaac 1.35, CoreAudioToolbox 7.9.7.9, AAC-LC Encoder, TVBR q127, Quality 96

i want quicktime best vbr but in its quicktime present mode

Ok, i checked some other threads and found that 64 produces similar results as quicktime, but the file is slightly smaller and the bitrates are different *in certain places although it gives the same avg bitrate in foobar*, does that mean its not the exact setting quicktime uses?

Also, I know I might be stretching it, but can I have my needledrops in 24 bit now too *or is that redundant with quicktime* *nevermind read in another thread it serves no purpose for aac*

This post has been edited by db1989: Apr 12 2012, 09:33
Reason for edit: as above
Go to the top of the page
+Quote Post
nu774
post Apr 12 2012, 08:09
Post #23





Group: Developer
Posts: 514
Joined: 22-November 10
From: Japan
Member No.: 85902



QUOTE
want quicktime best vbr but in its quicktime present mode

I don't get it; What do you mean by this?
As far as I know, QuickTime pro for Windows allows you to choose not-constrained true VBR mode, but doesn't let you configure VBR quality. Therefore, practically you had better be using constrained VBR on QT pro.
(Actually, you can always stick to CVBR... which has been used by iTunes, and showed slightly better result on HA listening test)

You can configure "quality" from QuickTime GUI, but actually it's a different parameter, which corresponds to -q option of qaac.
This parameter is not for controlling size/quality trade off. It just controls complexity of encoding process, and simply higher is better... as long as encoding speed is acceptable for you.
Go to the top of the page
+Quote Post
Mix3dmessagez
post Apr 12 2012, 08:37
Post #24





Group: Members
Posts: 87
Joined: 22-March 09
Member No.: 68252



I'm sorry for being unspecific.I'll try to better translate my gibberish LOL

In QT, i choose best quality, variable bit rate encoding strategy and sound quality recommended *only choice*
I wanted settings that reflected this choice in qaac
From what i understand it tries to be as transparent as possible around 128 bit aac.
While qt didn't let you configure quality, it gave quality control of like 4 different options.These from what I remember, affected sound quality and changed the bitrates around differently.

I see about the -q option, but from what i read in your faq by default the best is chosen *-q2*
I tried to reproduce the closest settings to quicktimes files, but am getting slighty different bitrates *although same average as whole*, and filesizes
I appreciate your program and don't want to come off as a jack a i just want to input settings that will give me exactly what quicktime will, with the best resampler

This post has been edited by db1989: Apr 12 2012, 09:33
Reason for edit: as above
Go to the top of the page
+Quote Post
nu774
post Apr 12 2012, 09:45
Post #25





Group: Developer
Posts: 514
Joined: 22-November 10
From: Japan
Member No.: 85902



Well, it's interesting. Quickly tried now QT pro 7.7.1 for Win with a few configuration parameters, and I also couldn't get identical result between QT pro and qaac. Same for CVBR.
On the other hand, I could confirm iTunes plus is still bit-identical with qaac --cvbr 256 -q2, and iTunes custom (CVBR) is identical with qaac --cvbr -q1;

qaac is now built directly upon CoreAudioToolbox, and is not using QuickTime API.
QT pro might be doing something different... I don't know for sure.
Go to the top of the page
+Quote Post

2 Pages V   1 2 >
Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 2nd August 2014 - 08:52