Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: lossyWAV 1.4.0 Released (Read 78596 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

lossyWAV 1.4.0 Released

Reply #50
I am not using any custom params with lossyWAV, just -q preset.


Could you please post your command line(s) for the processing / encoding?

Thanks in advance.
lossyWAV -q X -a 4 -s h -A --feedback 2 --limit 15848 --scale 0.5 | FLAC -5 -e -p -b 512 -P=4096 -S- (having set foobar to output 24-bit PCM; scaling by 0.5 gives the ANS headroom to work)

lossyWAV 1.4.0 Released

Reply #51
Could you please post your command line(s) for the processing / encoding?  Thanks in advance.


Sure thing, Nick.

 
Code: [Select]
lossyWAV.exe source.wav -q E -f

WMAencode.exe source.lossy.wav output.wma --codec lsl


Using version 0.2.9b of WMAencode.

The reason I'm using file input/output and not pipe redirect is because I'm transcoding to multiple formats, so it's faster to preprocess once to file and encode multiple times.

 


lossyWAV 1.4.0 Released

Reply #52
I suspect that it can be a bug in WMAencode. I changed it a little, so try version 0.2.9c.

lossyWAV 1.4.0 Released

Reply #53
I suspect that it can be a bug in WMAencode. I changed it a little, so try version 0.2.9c.


Thanks, I will give that a try. Any idea what change between lossyWAV 1.3.0 and 1.4.0 may have triggered it?


Update:  Yep, 0.2.9c works with 1.4.0 output! Thanks!

I'm still curious about what change had triggered the issue in the first place. 


lossyWAV 1.4.0 Released

Reply #54
When using 96KHz with 512 sample blocksize is like 44.1 (or 48 because it's close) with a 256 sample blocksize...

Am I right?


Indeed you are - 44.1/48 kHz use 512 samples per codec-block; 88.2/96 kHz use 1024; 176.4/192 kHz use 2048 and 352.8/384 kHz would use 4096.

I should really add this to the wiki....


Hello Nick.C,

I'm still looking for that killer sample. I've tried "herding Calls" a thousand times but it doesn't convince me, I cannot ABX the artifacts that are created by LossyWAV (did not really ABX because even without I cannot hear the difference). This offcourse is a good thing! Still I am looking for a sample that has some hearable difference agains the original in X mode. Is there a sample out there? My ears are quite allright for my age of 32, shouldn't be the problem I hope. Sometimes I think I hear it, but then I think it just my knowledge if cheating me because I know which sample I'm playing.

PS. Is there a way to tell LossyWAV to do some extreme things? Like taking out way more bits than in the X mode? I would like to see how far I can go before it gets really nasty.

Unfortunately I do not have the programming skills to create this myself.

Thanks you in advance for your answer !


I reported it in the past. CD artist Ivri Lider, track: Dear sir.  Noise 'blip' in right channel very audible at low settings and still exists at -P. This was using lossywav 1.1 with noise shaping off. It still existed in v1.3 with A=OFF though I think with the recent NS it has improved or eliminated the problem. Going up to -C with recent versions or --Quality 3 with v1.1 also fixes it.

This is lossywav v1.3 -P setting: -  more subtle but still audiable

foo_abx 1.3.4 report
foobar2000 v1.0.3
2012/07/28 19:56:53

File A: C:\stuff\music\abx\09 - Dear Sir.flac
File B: C:\stuff\music\abx\Dear Sir2.flac

19:56:53 : Test started.
19:57:10 : 01/01  50.0%
19:57:21 : 02/02  25.0%
19:57:54 : 03/03  12.5%
19:58:05 : 04/04  6.3%
19:58:37 : 05/05  3.1%
19:58:50 : 06/06  1.6%
19:59:10 : 07/07  0.8%
19:59:32 : 08/08  0.4%
19:59:45 : Test finished.

----------
Total: 8/8 (0.4%)


Theres also the 'serioustrouble' mp3 problem track that sounded bad to me below -P and also 'Mandylion_short' sample . Again recent noise shaping changes in v1.3 ~ 1.4 may have improved these.

lossyWAV 1.4.0 Released

Reply #55
When using 96KHz with 512 sample blocksize is like 44.1 (or 48 because it's close) with a 256 sample blocksize...

Am I right?


Indeed you are - 44.1/48 kHz use 512 samples per codec-block; 88.2/96 kHz use 1024; 176.4/192 kHz use 2048 and 352.8/384 kHz would use 4096.

I should really add this to the wiki....


Hello Nick.C,

I'm still looking for that killer sample. I've tried "herding Calls" a thousand times but it doesn't convince me, I cannot ABX the artifacts that are created by LossyWAV (did not really ABX because even without I cannot hear the difference). This offcourse is a good thing! Still I am looking for a sample that has some hearable difference agains the original in X mode. Is there a sample out there? My ears are quite allright for my age of 32, shouldn't be the problem I hope. Sometimes I think I hear it, but then I think it just my knowledge if cheating me because I know which sample I'm playing.

PS. Is there a way to tell LossyWAV to do some extreme things? Like taking out way more bits than in the X mode? I would like to see how far I can go before it gets really nasty.

Unfortunately I do not have the programming skills to create this myself.

Thanks you in advance for your answer !


I reported it in the past. CD artist Ivri Lider, track: Dear sir.  Noise 'blip' in right channel very audible at low settings and still exists at -P. This was using lossywav 1.1 with noise shaping off. It still existed in v1.3 with A=OFF though I think with the recent NS it has improved or eliminated the problem. Going up to -C with recent versions or --Quality 3 with v1.1 also fixes it.

This is lossywav v1.3 -P setting: -  more subtle but still audiable

foo_abx 1.3.4 report
foobar2000 v1.0.3
2012/07/28 19:56:53

File A: C:\stuff\music\abx\09 - Dear Sir.flac
File B: C:\stuff\music\abx\Dear Sir2.flac

19:56:53 : Test started.
19:57:10 : 01/01  50.0%
19:57:21 : 02/02  25.0%
19:57:54 : 03/03  12.5%
19:58:05 : 04/04  6.3%
19:58:37 : 05/05  3.1%
19:58:50 : 06/06  1.6%
19:59:10 : 07/07  0.8%
19:59:32 : 08/08  0.4%
19:59:45 : Test finished.

----------
Total: 8/8 (0.4%)


Theres also the 'serioustrouble' mp3 problem track that sounded bad to me below -P and also 'Mandylion_short' sample . Again recent noise shaping changes in v1.3 ~ 1.4 may have improved these.


Thanks,

I'll try that one asap. I'm at work now so can't do it now.

At what moment (m:s) is the artifact(s) audible? I am not familiar with the album

lossyWAV 1.4.0 Released

Reply #56
Theres also the 'serioustrouble' mp3 problem track that sounded bad to me below -P and also 'Mandylion_short' sample . Again recent noise shaping changes in v1.3 ~ 1.4 may have improved these.


If you have the inclination to re-visit these samples, I would suggest that the "--feedback <n>" parameter might be of some benefit when added to the processing command line.
lossyWAV -q X -a 4 -s h -A --feedback 2 --limit 15848 --scale 0.5 | FLAC -5 -e -p -b 512 -P=4096 -S- (having set foobar to output 24-bit PCM; scaling by 0.5 gives the ANS headroom to work)

lossyWAV 1.4.0 Released

Reply #57
Theres also the 'serioustrouble' mp3 problem track that sounded bad to me below -P and also 'Mandylion_short' sample . Again recent noise shaping changes in v1.3 ~ 1.4 may have improved these.


If you have the inclination to re-visit these samples, I would suggest that the "--feedback <n>" parameter might be of some benefit when added to the processing command line.


Totally offtopic, but your signature says:

:lossyWAV -q X -a 4 --feedback 4| FLAC -8 ~= 320kbps

Wouldn't it be better to suggest the right blocksize, because I think many people would just try -8 without further parameters:

:lossyWAV -q X -a 4 --feedback 4| FLAC -8 -b 512 ~= 320kbps


lossyWAV 1.4.0 Released

Reply #58
Wouldn't it be better to suggest the right blocksize, because I think many people would just try -8 without further parameters:

:lossyWAV -q X -a 4 --feedback 4| FLAC -8 -b 512 ~= 320kbps



It might, I suppose - done. It will be sub-optimal higher or lower sample-rates than 44.1/48kHz though.
lossyWAV -q X -a 4 -s h -A --feedback 2 --limit 15848 --scale 0.5 | FLAC -5 -e -p -b 512 -P=4096 -S- (having set foobar to output 24-bit PCM; scaling by 0.5 gives the ANS headroom to work)

lossyWAV 1.4.0 Released

Reply #59
Wouldn't it be better to suggest the right blocksize, because I think many people would just try -8 without further parameters:

:lossyWAV -q X -a 4 --feedback 4| FLAC -8 -b 512 ~= 320kbps



It might, I suppose - done. It will be sub-optimal higher or lower sample-rates than 44.1/48kHz though.


You're right, but the most music is encoded at 44,1kHz and the resulting bitrate matches. When using 96kHz you cannot maintain the ~320kbit bitrate. I'm not sure if there are a lot of people who choose for 96kHz 24 bit samples would use LossyFLAC, these people often think you really need all those bits. It makes more sense to convert that to 16bit 44,1kHz (just my opinion) and THEN convert to LossyFLAC. I have a family member who says the sound of his high definition BD's is way better than CD. Also the video and colors are way better than the same BD ripped and played through XBMC for example  . Cannot convince the guy, don't attempt to again 

This weekend I will be testing the LossyFLAC X in my car (Peugeot 208 touchscreen plays FLAC and OGG). Ogg was my first choice, because a bitrate of ~128kbit is enough in a noisy car (for me), but every song the last few seconds (10?) are cut off. Not so nice. I will try FLAC to see if that performs better.

Again: My compliments for LossyWAV. Really like it. Do you have any new versions, functionality plans in the near future?

I'd really love to see some extreme options, like 160kbit LossyFLAC files. They don't have to be transparent. I'd like to see if that can be a good alternative for portable/car use....

lossyWAV 1.4.0 Released

Reply #60
When using 96kHz you cannot maintain the ~320kbit bitrate. I'm not sure if there are a lot of people who choose for 96kHz 24 bit samples would use LossyFLAC, these people often think you really need all those bits.

I downloaded the free Peter Gabriel 24bit/96kHz tracks that were posted on 1st January - lossless FLAC was 3,000 kbit/s; lossyFLAC --standard -a 4 --feedback 4 resulted in 647 kbit/s with an average of between 12 and 14 bits removed for the tracks.

.... which got me thinking - what would the resultant bitrate have been at the "Insane" quality preset.... Answer: 901 kbit/s with an average of between 10.6 and 12 bits removed.
lossyWAV -q X -a 4 -s h -A --feedback 2 --limit 15848 --scale 0.5 | FLAC -5 -e -p -b 512 -P=4096 -S- (having set foobar to output 24-bit PCM; scaling by 0.5 gives the ANS headroom to work)

lossyWAV 1.4.0 Released

Reply #61
I'm not sure if there are a lot of people who choose for 96kHz 24 bit samples would use LossyFLAC

I downloaded the free Peter Gabriel 24bit/96kHz tracks that were posted [..]
.... which got me thinking - what would the resultant bitrate have been at the "Insane" quality preset.... Answer: 901 kbit/s with an average of between 10.6 and 12 bits removed.

Exactly, and the bits should be removed only where you can't hear them. I read on this forum the opinion that the least significant bits (5 or 6) of 24bit audio contain probably "random noise". That is because the noise floor of equipment and studios.
The savings on this type of files are huge, when using lossyWav, sometimes multiple GB per album (especially when they are in 5.1).

I totally agree that downsampling to 16/48kHz is also a valid option.

BTW did you remember to use FLAC -b 1024 ?
In theory, there is no difference between theory and practice. In practice there is.

lossyWAV 1.4.0 Released

Reply #62
The savings on this type of files are huge, when using lossyWav, sometimes multiple GB per album (especially when they are in 5.1).

I totally agree that downsampling to 16/48kHz is also a valid option.

BTW did you remember to use FLAC -b 1024 ?


The savings are as follows:

FLAC: 620MB; 3,000 kbit/s;
Insane: 187MB, 30.2%; 901 kbit/s; Saving: 433MB, 69.8%;
Standard: 130MB, 21.0%; 625 kbit/s; Saving: 490MB, 79.0%.

I had remembered to set the FLAC block size to 1024 for Insane but reprocessed the Standard set and saw a 22 kbps, 5MB, 0.8% reduction in size.
lossyWAV -q X -a 4 -s h -A --feedback 2 --limit 15848 --scale 0.5 | FLAC -5 -e -p -b 512 -P=4096 -S- (having set foobar to output 24-bit PCM; scaling by 0.5 gives the ANS headroom to work)

lossyWAV 1.4.0 Released

Reply #63
I'm not sure if there are a lot of people who choose for 96kHz 24 bit samples would use LossyFLAC
I downloaded the free Peter Gabriel 24bit/96kHz tracks that were posted [..]
.... which got me thinking - what would the resultant bitrate have been at the "Insane" quality preset.... Answer: 901 kbit/s with an average of between 10.6 and 12 bits removed.
Exactly, and the bits should be removed only where you can't hear them. I read on this forum the opinion that the least significant bits (5 or 6) of 24bit audio contain probably "random noise". That is because the noise floor of equipment and studios.
The savings on this type of files are huge, when using lossyWav, sometimes multiple GB per album (especially when they are in 5.1).

I totally agree that downsampling to 16/48kHz is also a valid option.

BTW did you remember to use FLAC -b 1024 ?

As he is the author of LossyWAV I guess he is using the right command line options am I wrong

PS. in response to Nick.C. The gain in saved space is indeed quite a lot, but the people who stick to these bitdepths/samplingsfrequencies are hard to convince.

I downloaded some HD tracks also, but immediatly downsampled them to 44/16 (SoX) via Foobar. I absolutely cannot hear ANY difference. I even didn't bother using 48kHz as it is dividable by 2.

lossyWAV 1.4.0 Released

Reply #64
I wonder, what is best in terms of re-encoding afterwards.

Wave -> lossyWav -> "some lossy format"
or
Wave -> Opus(x bitrate) -> "some lossy format"

I know of course Flac is the best (or any lossless), but you probably get my question.

Also is there any special commands i need to care about, saw one saying something about 1024 block?

lossyWAV 1.4.0 Released

Reply #65
Also is there any special commands i need to care about, saw one saying something about 1024 block?

1024 blocks are only used when the sample rate is 88.2k or 96k. So in a (more) normal case you should specify a block size of 512 to the lossless codec you will use to compress a lossy.wav file. (example: flac -b 512 ).
Look for command line or script examples from Nick.C. There must be a wiki page somewhere.

About going from one lossy to another, you should see what suits you best. It all depends on what you use it for and what's acceptable to you.
In theory, there is no difference between theory and practice. In practice there is.

lossyWAV 1.4.0 Released

Reply #66
Also is there any special commands i need to care about, saw one saying something about 1024 block?

1024 blocks are only used when the sample rate is 88.2k or 96k. So in a (more) normal case you should specify a block size of 512 to the lossless codec you will use to compress a lossy.wav file. (example: flac -b 512 ).
Look for command line or script examples from Nick.C. There must be a wiki page somewhere.

About going from one lossy to another, you should see what suits you best. It all depends on what you use it for and what's acceptable to you.


Ah i see, thanks.

Well the lossy -> lossy is a hard one for me.
I i can easily spare bitrate, but lossless is a bit much.
Therefore i wondered if Opus would be able to make more transparent at higher bitrate compared to lossyWAV (i can afford around 320-512 bitrate).

lossyWAV 1.4.0 Released

Reply #67
Look for command line or script examples from Nick.C. There must be a wiki page somewhere.

About going from one lossy to another, you should see what suits you best. It all depends on what you use it for and what's acceptable to you.

There is: here (although it is a bit out of date with respect to some of the newer options).

Ah i see, thanks.

Well the lossy -> lossy is a hard one for me.
I i can easily spare bitrate, but lossless is a bit much.
Therefore i wondered if Opus would be able to make more transparent at higher bitrate compared to lossyWAV (i can afford around 320-512 bitrate).

Standard quality yields c.440 kbit/s for my music collection. Regarding lossy to lossy - it's your call.
lossyWAV -q X -a 4 -s h -A --feedback 2 --limit 15848 --scale 0.5 | FLAC -5 -e -p -b 512 -P=4096 -S- (having set foobar to output 24-bit PCM; scaling by 0.5 gives the ANS headroom to work)

lossyWAV 1.4.0 Released

Reply #68
Are you going to add mutithreading processing in order to speed up whole process? I have 8C/16T CPU and processing of movie track 5.1 takes to much time due single threaded code.

lossyWAV 1.4.0 Released

Reply #69
Are you going to add mutithreading processing in order to speed up whole process? I have 8C/16T CPU and processing of movie track 5.1 takes to much time due single threaded code.


I suggest using Foobar2000 and process from there. It is capable to start more conversions simultaneously.

With the task manager you can set the amount of processors assigned to foobar. Setting too much CPU's can result in HDD seeking (overload) slowing things down. Just try with 4 first and build up from there keeping an eye on the system resources.

lossyWAV 1.4.0 Released

Reply #70
Are you going to add mutithreading processing in order to speed up whole process? I have 8C/16T CPU and processing of movie track 5.1 takes to much time due single threaded code.


I suggest using Foobar2000 and process from there. It is capable to start more conversions simultaneously.

With the task manager you can set the amount of processors assigned to foobar. Setting too much CPU's can result in HDD seeking (overload) slowing things down. Just try with 4 first and build up from there keeping an eye on the system resources.


Your method is: Build 8 houses in the same time
But I prefer: Build 1 house 8 times faster

I was thinking about encoding multiple chunks of data (let's say 1min) in the same time and then combine all chunks into one FLAC file.
I think I can achieve that with something like this

ffmpeg.exe 00:00:00 00:00:00.999 -> lossywav -> flac encoder -> 01.flac
ffmpeg.exe 00:00:01 00:00:01.999 -> lossywav -> flac encoder -> 02.flac
ffmpeg.exe 00:00:02 00:00:02.999 -> lossywav -> flac encoder -> 03.flac

Unfortunately there is major problem. How to combine multiple flacs into one flac without RE-ENCODING?! Btw. Forget about copy /b method...

lossyWAV 1.4.0 Released

Reply #71
I think I have found cosmetic error. Instead of 00:60.84 it should show 01:00.84.

lossyWAV 1.4.0 Released

Reply #72
I think I have found cosmetic error. Instead of 00:60.84 it should show 01:00.84.


Thanks Atak - I will investigate. Apologies on the lack of multi-threaded processing - it's a fairly major rewrite that I'm not ready to commit to.
lossyWAV -q X -a 4 -s h -A --feedback 2 --limit 15848 --scale 0.5 | FLAC -5 -e -p -b 512 -P=4096 -S- (having set foobar to output 24-bit PCM; scaling by 0.5 gives the ANS headroom to work)

 

lossyWAV 1.4.0 Released

Reply #73
Last question. Processing all channels at the same will be difficult to implement? I mainly work with 6 , 8 channel tracks so 6 , 8 times faster encoding would be still great on my xeon 8c/16t.

lossyWAV 1.4.0 Released

Reply #74
Are you going to add mutithreading processing in order to speed up whole process? I have 8C/16T CPU and processing of movie track 5.1 takes to much time due single threaded code.


I suggest using Foobar2000 and process from there. It is capable to start more conversions simultaneously.

With the task manager you can set the amount of processors assigned to foobar. Setting too much CPU's can result in HDD seeking (overload) slowing things down. Just try with 4 first and build up from there keeping an eye on the system resources.


Your method is: Build 8 houses in the same time
But I prefer: Build 1 house 8 times faster

I was thinking about encoding multiple chunks of data (let's say 1min) in the same time and then combine all chunks into one FLAC file.
I think I can achieve that with something like this

ffmpeg.exe 00:00:00 00:00:00.999 -> lossywav -> flac encoder -> 01.flac
ffmpeg.exe 00:00:01 00:00:01.999 -> lossywav -> flac encoder -> 02.flac
ffmpeg.exe 00:00:02 00:00:02.999 -> lossywav -> flac encoder -> 03.flac

Unfortunately there is major problem. How to combine multiple flacs into one flac without RE-ENCODING?! Btw. Forget about copy /b method...


I get the point, but multi-cpu processing one file is a major change in code, just like the developer says. Many file compression programs have problems with multicore cpu's. You can use multi cpu's, but many times at the cost of compression ratio, because the data becomes uncomparable when it is sliced in separate chunks. For FLAC's this is less of a problem, but general data compression relies on big dictionaries...

I see no problem in the 8 houses method. Normally you have so much FLAC files that there is no gain in speed by processing 1 file by 8 cpu's.