similar to: Ogg/Vorbisreport, FFT optimizations

We've implemented a vorbis decoder based on Tremor and as part of the documentation we're also writing about psycho acoustic models and encoding. We're quite up to date with the decoding process and psycho acoustics in general but unfortunately not on the psycho acoustic encoding used in Vorbis. We have a few questions that would we would be very thankful to have answer to: Which

Hello all, I have recently ported Vorbis to a Texas instruments DSP (TMS320C5510), The problem I have is that it is not decoding the audio in real time( about 50% of real time). Could any one give me a clue in which area I could start to optimize the code. Regards Kevin <p><p><p>--- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage:

As part of the documentation of a Vorbis decoding project, a quick explanation of the encoding procedure is required. I quite up to date with the decoding process but not with the encoding part. Is there any documentation explaining the different steps during encoding? What I'm looking for is a list of encoding steps, very briefly explained and more details on a few fundamental steps, like

Is there a archive with a collection of different .ogg files encoded with different settings that can be used to verify compatibility? * Different window size (e.g. != 256,2048) * Much larger codebooks and other extreme cases that are not generated using a standard encoder. * floor0 Regards -- / Johannes Sandvall --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project

Hello !! While reading I am facing lot of problems . Doing reserve enginnering without any document is really putting in trouble . ( that is not the case in ac3 , mp3 ) . The problems are ----> 1. How physcoacoustic model is generated . I was reading noise masking . I have understood that FFT of 2048 samples is used for finding tone masking . And MDCT of same samples is used for

Hi Sam, Thanks for your answer! Sorry I didn't really know how to process, but this issue block me to make some legacy C code works without having some glue with a bit static lib ;( (The change have really low risk) The changes I used locally to make dlopen/dlclose et dlsym mechanism works :) : Index: wasm/Writer.cpp =================================================================== ---

Is someone working on the infamous high frequency boost effect ? S.Gesemann published an interesting paper three months ago: http://www.xiph.org/archives/vorbis-dev/200403/att-0005/01-sq.pdf <p> -- Giuliano. --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to

Hi, i just wonder if there's a cheap way of getting the number of oggVorbis pcm frames in an ogg_page, without actually decoding it? Can i use the granulepos of the page, substract the granulepos of the previous page? is this always guaranteed to work? thanks, Chris

I'm trying to install packages on windows XP and I have trouble with command Rcmd build (R version 1.8.1) : In the Windows console for package maanova for example, answer is : C:\Documents and Settings\dillies\Mes documents\ghis\packages>Rcmd build maanova * checking for file 'maanova/DESCRIPTION' ... OK * preparing 'maanova': * cleaning src *

Hello, The table generated by lld (Wasm target) have a max size. So this make impossible to add some others functions in the table . As for the moment only one Table is available in WAsm this is the only way to add function in a dynamic way. Having a none static size of the table is useful for a kind dynamic loader/inker at run-time: by adding some exported functions coming from another

Hi , I was trying to guess the channel order or Tremor decoder for 5.1 channel file. Though the standard state it to be FL,C,FR,RL,RR,LFE but in my opinion its FL,C,FR, LFE RL,RR. I am not sure if my guess is correct . I came to this conclusion after downmixing from five to two channels. If i assume the channel order to be as per the standard then my two channel output is not correct but if i

Hi vorbis experts! I'd like ot write application under Symbian that use theora as video codec & tremor as an audio one. I played at first with libogg & theora on PC, then successfully ported them both. Next I decided to port Tremor but I found that it use its own files from Ogg library. The API is almost identical, so I tried to bind the Tremor with upstream Ogg library (it actually

Hi vorbis experts! I'd like ot write application under Symbian that use theora as video codec & tremor as an audio one. I played at first with libogg & theora on PC, then successfully ported them both. Next I decided to port Tremor but I found that it use its own files from Ogg library. The API is almost identical, so I tried to bind the Tremor with upstream Ogg library (it actually

Hi, While following the developer mailing list, I came across "tremor-lowmem" version. I did some search of it in Xiph.org and "tremor" source code but wasn't able to find it. I would like to know if there is any separate thread for the "tremor-lowmem" version as we are working on the fixed point implementation of Vorbis code and we are looking forward for

Jean-Marc Valin wrote: > There is definitely some use for a Neon fixed-point FFT. How much > exactly I'm not sure. Fixed-point is a bit more than just a fall-back Well, we use fixed-point mode by default in Firefox for both Firefox OS and Fennec (Firefox on Android). The reason is that, although there is some NEON-class hardware where float does finally appear to be a little bit

Thanks Timothy and Jean-Marc, I will start NEON optimizing fixed-point FFT. Is it int32 good enough? Benchmark data shows that FFT using int16 is much faster than FFT using int32. > -----Original Message----- > From: Timothy B. Terriberry [mailto:tterribe at xiph.org] > Sent: Friday, December 26, 2014 6:52 AM > To: Phil Wang; opus at xiph.org > Cc: Zhongwei Yao; Yang Zhang; Zhou

There is definitely some use for a Neon fixed-point FFT. How much exactly I'm not sure. Fixed-point is a bit more than just a fall-back for CPUs with no FPU. There are CPUs for which fixed-point is still faster. It depends on the exact model but also on what you run. For example, even on x86 I believe that SILK encoding is slightly faster in fixed-point, even though CELT is faster in float.

Hello! thank you for your efforts creating R! While preparing the next lessons I found that I had to multiply with 1/length(x) by hand if I want to get the inverse Fouriertransform. The help states it will do it by itself. Version R-0.60.1, Linux Seems to be pretty easy to correct either the fft or :-) the help(fft)! Cu Detlef Detlef Steuer Universitaet Dortmund ///////

Hi everyone, I have profiled Opus on AArch64. I just run opus_demo with some pcm files. Following is time proportion of FFT with different bitrate. Bitrate | Time cost by FFT/iFFT 24kb/s | 15% 48kb/s | 15% 96kb/s | 13% Any comment? I want some data close to real application, any suggestion? Thanks, Phil Wang -------------- next part -------------- An HTML attachment was scrubbed... URL:

> I am working on DSP module of Ne10. I see there are fixed-point and > floating-point FFT inside Opus. Is fixed-point FFT only a fall back for CPU > without VFP? On ARMv7-A and ARMv8-A, benchmark result shows that fixed-point > (int32) and floating-point (float32) FFT have similar performance. I guess > fixed-point version is not often used on these platforms. Is it worth the >