IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
WMA lossy to other lossy format - clear answer desperately needed
steve317
post Feb 4 2013, 20:58
Post #1





Group: Members
Posts: 1
Joined: 4-February 13
Member No.: 106414



I hope this post isn't redundant, wrongly categorized, or otherwise irritating, but I'm new to this forum today and I'm desperate: I've searched for this answer for weeks and the closest I get to an answer is an answer to the opposite question. We can't even figure out anyone to pay to get an answer to this question. I need a more-or-less black-and-white answer to the following question:


We have several hundred audio recordings of lectures (so all just spoken word), using a 5-year-old Olympus digital voice recorder (the kind of thing you buy in Staples or Best Buy) with a medium quality flat-mic. Unfortunately, the person doing these recordings used the WMA settings instead of MP3, because she had been convinced that WMA was a much better sound quality.

We have a giant archive (thousands more lectures) of these mission critical lecture recordings stored on a Mac Pro using iTunes -- they were all recorded in AAC or MP3 in the first place. We need to add the hundreds of WMAs to the collection (they are some of the most important recordings!) -- they are essential and we need to start adding them to the archive in the next week.



So here's what we need to do:

We need to convert the WMAs into some more universal format. We were planning on AAC (it's universal enough as far as we can see) if the quality and space savings was better; we'd also be happy with the more universal MP3 if we don't loose quality or space.




Here's the confusion and problem:

We want to avoid any significant quality loss. Every article on this subject of Lossy to Lossy transcoding talks about dropping the bit-rate or some other maneuver intended to decrease the file size. We don't care about that -- we are happy to keep it roughly the same size. We just don't want to DOUBLE the file size, which is what happens when we transcode from WMA lossy to some lossless files such as Apple Lossless. Storage is cheap, so we aren't willing to lose much quality to save space, but doubling the space on hundreds of tracks is a problem -- it is decimating our hard drive space and posing countless problems with mobile access strategies for our teaching team through services like iTunes Music Match etc.

The recordings are not exactly audiophile in the first place, but that's all the more reason it is important that they don't lose too much additional quality in the conversion. We have experienced how a casual little effort of "simple" lossy-to-lossy transcoding using some random program can produce a noticeably dramatic quality loss...We can't have hundreds of tracks going from a little difficult to listen to, to sounding like we're using wax cylinders. That would be the opposite exponential problem from filling our hardrives with double-sized lossless files, but even less desirable. And, for space and sanity reasons, we intend to delete the WMAs when we're done, so the transcoded files will become the master files.



So, here's the answer we're looking for:

What do we do to turn lossy WMA files into the AAC or MP3 files that are as close to the same quality as the WMAs as possible, while still remaining about the same file size or smaller, or at least not too much bigger? What software options are there, what settings are critical, etc.? (We would prefer to do it on Mac OS X 10.8, but we do have access to Windows 7 or 8 if necessary just for the conversion process, but would have to get another HDD formatted for Windows)



As I said, we need to do this soon, but need to do it properly from the start, or our mistake will multiply 100's of times. Thank you so much in advance for any help you can provide!
Go to the top of the page
+Quote Post
darkbyte
post Feb 4 2013, 21:06
Post #2





Group: Members
Posts: 144
Joined: 14-June 11
Member No.: 91517



If you don't want to lose quality, lossless is your only choice i'm affraid.
Although it's not what you want i would give LossyWAV a chance and encode the audio with FLAC. You can achieve very good quality this way without encoding the file again with a heavy weight perceptual coder (LossyWAV only adds shaped noise and because you mentioned that the Mic was not the best quality i think you'll not notice any difference in the audio quality). My estimated bitrate would be around ~192kbps with 44.1Khz Mono audio processed through LossyWAV and encoded with FLAC.

This post has been edited by darkbyte: Feb 4 2013, 21:09


--------------------
Wavpack -b450x1c
Go to the top of the page
+Quote Post
pdq
post Feb 4 2013, 21:13
Post #3





Group: Members
Posts: 3371
Joined: 1-September 05
From: SE Pennsylvania
Member No.: 24233



If it were me then I would convert first to a lossless version, then try some post-processing to improve the sound, noise reduction, volume leveling, etc.

Only when you are completely satisfied with the results should you convert back to lossy, even if it is LossyWav.
Go to the top of the page
+Quote Post
AndyH-ha
post Feb 4 2013, 21:55
Post #4





Group: Members
Posts: 2204
Joined: 31-August 05
Member No.: 24222



There are various aspects of theory, how processing effects the data, that while true, are frequently inconsequential when one gets to the what one actually hears. I have transcoded many hundreds of hours of spoken audio from various CBR mp3 encodings to VBR mp3 at a lower bit rate, saving 50% or more of the storage space. There may be a measurable lossy of quality by some objective standard but neither I nor several other people who use the audio can hear any difference.

Possibly if we faithfully performed ABX tests on many samples we would find something that is distinguishable but we certainly don't notice any difficulties in understanding the material. Some tests to your own standards would seem the best way for you to go. Pick what you would like to use, transcode some material, listen to it. Do ABX tests on samples if you are uncertain. If you can find any audible differences, then decide if it is really a deterioration or just some slight difference in sound that does not matter.

If you do want to make changes, such as noise reduction, equalization, or volume levels, then converting to lossless, and keeping it as lossless until you are definitely done making changes, is the reasonable thing to do.
Go to the top of the page
+Quote Post
Porcus
post Feb 4 2013, 22:17
Post #5





Group: Members
Posts: 1842
Joined: 30-November 06
Member No.: 38207



QUOTE
What do we do to turn lossy WMA files into the AAC or MP3 files that are as close to the same quality as the WMAs as possible, while still remaining about the same file size or smaller, or at least not too much bigger?


You will always have a generation loss transcoding lossy-to-lossy. This is however speech, and it might not be much of an issue. I see that you are also writing as follows:

QUOTE
We just don't want to DOUBLE the file size, which is what happens when we transcode from WMA lossy to some lossless files such as Apple Lossless.


If your WMA files are half of a lossless format, then maybe the bitrate is already quite a lot? Can you open them in a media player and read off the bitrate?



QUOTE
And, for space and sanity reasons, we intend to delete the WMAs when we're done, so the transcoded files will become the master files.


I would normally say that that would be for space and not for sanity. Especially when you have only a week to go, I would keep the originals.


But your constraints are:
- You insist on using iTunes and you cannot have the wma files there?
- You cannot do any extensive testing for the moment, with this time to go?
- How many hours? How many GB's?


--------------------
One day in the Year of the Fox came a time remembered well
Go to the top of the page
+Quote Post
DonP
post Feb 4 2013, 23:56
Post #6





Group: Members (Donating)
Posts: 1471
Joined: 11-February 03
From: Vermont
Member No.: 4955



A GB of disk space these days runs about 5 cents US, so uncompressed voice (assuming the voice recorder is 1 channel, 22 kHz sampling rate) is going to be about 1 penny per hour of disk space cost. Granted, you have to spend about $100 to get a drive big enough for that kind of price efficiency, but you did drop phrases like "mission critical" and you would have enough to archive 20,000 hours of lecture (about double that as FLAC files)
Go to the top of the page
+Quote Post
Mach-X
post Feb 5 2013, 05:02
Post #7





Group: Members
Posts: 269
Joined: 29-July 12
From: Windsor, On, Ca
Member No.: 101859



I think the OP is implying that wma makes the files inaccessible to some of his users. lossywav certainly wouldnt make them more accessible. Being spoken word, if the original bitrate is high enough you could safely transcode to a similar aac bitrate without creating insanely audible issues. see if itunes will allow you to transcode the files directly, make a few test files and pick the setting that works. You dont always need a hammer to crack a nut...
Go to the top of the page
+Quote Post

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: 23rd July 2014 - 14:42