Please provide the following information when requesting support.
Hardware - GPU T4
Hardware - CPU
Operating System Ubuntu 20.04
Riva Version
TLT Version (if relevant)
Hello, I’m currently synthesizing audio from long texts with SSML tags in offline mode, however, the longer the duration of the audio, the worse the quality of the audio it generates. I’m thinking of using the online/streaming mode to split the text into smaller requests however the chunk sizes are very small, how can I change the chunk size to 10 seconds?
These are the parameters with which a new model can be built, however the documentation does not indicate what values are currently being used and I do not fully understand how to increase the size of the chunks:
–chunk_length CHUNK_LENGTH
Chunk length in mel frames to synthesize at one time
–overlap_length OVERLAP_LENGTH
Chunk length in mel frames to overlap neighboring
chunks
chunkerFastPitch:
–chunkerFastPitch.max_sequence_idle_microseconds CHUNKERFASTPITCH.MAX_SEQUENCE_IDLE_MICROSECONDS
Global timeout, in ms
–chunkerFastPitch.max_batch_size CHUNKERFASTPITCH.MAX_BATCH_SIZE
Default maximum parallel requests in a single forward
pass
–chunkerFastPitch.min_batch_size CHUNKERFASTPITCH.MIN_BATCH_SIZE
–chunkerFastPitch.opt_batch_size CHUNKERFASTPITCH.OPT_BATCH_SIZE
–chunkerFastPitch.preferred_batch_size CHUNKERFASTPITCH.PREFERRED_BATCH_SIZE
Preferred batch size, must be smaller than Max batch
size
–chunkerFastPitch.batching_type CHUNKERFASTPITCH.BATCHING_TYPE
–chunkerFastPitch.preserve_ordering CHUNKERFASTPITCH.PRESERVE_ORDERING
Preserve ordering
–chunkerFastPitch.instance_group_count CHUNKERFASTPITCH.INSTANCE_GROUP_COUNT
How many instances in a group
–chunkerFastPitch.max_queue_delay_microseconds CHUNKERFASTPITCH.MAX_QUEUE_DELAY_MICROSECONDS
Maximum amount of time to allow requests to queue to
form a batch in microseconds
–chunkerFastPitch.optimization_graph_level CHUNKERFASTPITCH.OPTIMIZATION_GRAPH_LEVEL
The Graph optimization level to use in Triton model
configuration
Is there another way to avoid this loss of audio quality with long texts?