<div dir="ltr"><div>Thanks for the info. I will look at <a href="http://encode.ru">encode.ru</a> .<br><br>In the meantime I discovered "nanozip" : <a href="http://www.nanozip.net/">http://www.nanozip.net/</a><br>
<br>On <a href="http://compressionratings.com/sort.cgi?aud1.brief+4nf_pf">http://compressionratings.com/sort.cgi?aud1.brief+4nf_pf</a>, the result is amazing : 100 MB compressed in ... 0.44 sec !<br></div><div>with compressed size = 60% of the original size ! <br>
i.e. ratio = similar to FLAC, and very very fast...<br></div><div><br>nanozip 0.09a -m.5g -cf ***** 63 710 155 bytes 60.2 % 0.44 sec 0.44 sec<br><br></div><div>Unfortunately, "nanozip" is closed source. Do you have any idea on the algorithm behind that makes this possible ?<br>
<br></div><div class="gmail_extra"><br><br><div class="gmail_quote">2014-03-17 21:22 GMT+01:00 <span dir="ltr"><<a href="mailto:neheb@hushmail.com" target="_blank">neheb@hushmail.com</a>></span>:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
flac -0 would work. I was under the impression that it just packed and did not compress but that was wrong. For FLACCL as well.<br>
<br>
FLACCL is faster and compresses better. My test file was encoded in 10 seconds using FLACCL and 26 using regular FLAC. If GPU is not possible, you're probably SOL.<br>
<br>
FLACCL -0 --fast-gpu --no-md5: 805893523 bytes - 10.687 seconds<br>
FLAC -0: 838208659 bytes - 27.170 seconds //no idea how to remove the MD5 checksum from the encode. The MD5 function is not SSE-accelerated and is pretty slow.<br>
<br>
One other variant is to use a general purpose compressor(maybe LZ4 or even Zhuff) and apply a filter to get the size down. I tried this and got miserable results, mainly because the filter that i used was slow. You may want to post on <a href="http://encode.ru" target="_blank">http://encode.ru</a> if you wish to take this path. The site is not in Russian despite the last two letters.<br>
<div class="HOEnZb"><div class="h5"><br>
On 3/17/2014 at 10:08 AM, "Nouvelle Collection" <<a href="mailto:nouvellecollection@gmail.com">nouvellecollection@gmail.com</a>> wrote:<br>
><br>
>Hello neheb,<br>
><br>
>Thanks for your answer.<br>
>Unfortunately, I won't be able to use something that needs a GPU,<br>
>because<br>
>some users may not have a GPU on their computer.<br>
>What's the fastest option when running "standard FLAC" in order to<br>
>have<br>
>fastest compression time ?<br>
><br>
>> As an aside: the author of LZ4 has also created Zhuff which is<br>
>basically<br>
>LZ4 with an entropy coder. Encoding<br>
>> speeds are slightly slower but decompression is much faster.<br>
>Compresses<br>
>better as well.<br>
>> See: <a href="http://fastcompression.blogspot.com/p/zhuff.html" target="_blank">http://fastcompression.blogspot.com/p/zhuff.html</a><br>
>Very cool ! I tried it and I think it will probably fit my needs.<br>
>I just need to find a Python-Zhuff binding :)<br>
><br>
>Best regards<br>
><br>
><br>
><br>
><br>
><br>
>2014-03-17 1:11 GMT+01:00 <<a href="mailto:neheb@hushmail.com">neheb@hushmail.com</a>>:<br>
><br>
>> FLACCL might work: <a href="http://www.cuetools.net/wiki/FLACCL" target="_blank">http://www.cuetools.net/wiki/FLACCL</a><br>
>><br>
>> The problems with it are: slow initialization time. The OpenCL<br>
>kernel must<br>
>> be compiled during the first run. Which leads me to the second<br>
>drawback.<br>
>><br>
>> It requires a GPU. Core i5s include an integrated GPU which can<br>
>be<br>
>> utilized.<br>
>><br>
>> The latest version of FLACCL is at:<br>
>> <a href="http://www.cuetools.net/install/CUETools_2.1.5.zip" target="_blank">http://www.cuetools.net/install/CUETools_2.1.5.zip</a><br>
>><br>
>> Recommended command line would probably be: -1 --lax --no-md5 --<br>
>fast-gpu<br>
>><br>
>> --slow-gpu may produce faster results on your end. Needs to be<br>
>tested. -1<br>
>> could also be increased if needed. --no-md5 could also be<br>
>removed but the<br>
>> code that calculates the MD5 hash is slow. I also think that an<br>
>MD5 hash of<br>
>> the audio data is pointless since FLAC includes CRCs for every<br>
>block.<br>
>><br>
>> Also, there is currently a bug in the <a href="http://flac.cl" target="_blank">flac.cl</a> kernel which<br>
>sometimes<br>
>> produces wrong encodes at higher compression levels. The fixed<br>
>kernel can<br>
>> be downloaded from:<br>
>><br>
><a href="http://sourceforge.net/p/cuetoolsnet/code/ci/default/tree/CUETools" target="_blank">http://sourceforge.net/p/cuetoolsnet/code/ci/default/tree/CUETools</a>.<br>
>Codecs.FLACCL/<a href="http://flac.cl?format=raw" target="_blank">flac.cl?format=raw</a><br>
>><br>
>> For a slightly smaller size, you can change the line "cbits =<br>
>min(cbits,<br>
>> clz(order + 1) + 1 - shared.task.obits);" to "cbits = min(cbits,<br>
>clz(order)<br>
>> + 1 - shared.task.obits);". I checked with the author to make<br>
>sure this is<br>
>> correct.<br>
>><br>
>> Also: BIG NOTE: The best results that I could get on my machine<br>
>was around<br>
>> 110MB/s (based on my own calculations). It may or may not be<br>
>acceptable for<br>
>> your purposes.<br>
>><br>
>> As an aside: the author of LZ4 has also created Zhuff which is<br>
>basically<br>
>> LZ4 with an entropy coder. Encoding speeds are slightly slower<br>
>but<br>
>> decompression is much faster. Compresses better as well. See:<br>
>> <a href="http://fastcompression.blogspot.com/p/zhuff.html" target="_blank">http://fastcompression.blogspot.com/p/zhuff.html</a><br>
>><br>
>> On 3/16/2014 at 10:27 AM, "Nouvelle Collection" <<br>
>> <a href="mailto:nouvellecollection@gmail.com">nouvellecollection@gmail.com</a>> wrote:<br>
>> ><br>
>> >Hello,<br>
>> ><br>
>> >Is there some version of FLAC that allows very very fast<br>
>encoding<br>
>> >(i.e.<br>
>> >able to process at least 150 MB / second of .wav input data on a<br>
>> >standard<br>
>> >computer : laptop computer, Core i5/i7, Windows 7 64 bit, 8 GB<br>
>> >RAM) ?<br>
>> >(It's ok to have a compression ratio which is a little bit lower<br>
>> >than<br>
>> >traditionnal FLAC)<br>
>> ><br>
>> >I'm looking for something which is between FLAC (very good<br>
>ratio,<br>
>> >slower<br>
>> >than LZ4) and LZ4 (very very fast compression : 400 MB / sec,<br>
>> >but lower<br>
>> >compression ratio than FLAC because it's not dedicated to<br>
>audio).<br>
>> ><br>
>> >Best regards.<br>
>><br>
>><br>
<br>
</div></div></blockquote></div><br></div></div>